Homework 2 Problems from text 4.23, 4.24, 4.30 (Do not turn in, just to help with the review) Matlab problem (Need to turn in): In this assignment, we will examine the sampling of a continuous-time sinusoidal signal xc(t) at various sampling rates. Since Matlab cannot strictly generate a continuous-time signal, you will generate a sequence {xn(nTH)} from xc(t) by sampling it at a very high rate TH such that the samples are very close to each other. A plot of xc(nTH) using the ”plot” command will then look like a continuous-time signal. 1. You can use a segment of code like the one that appears below to generate the ”continuous” signal for plotting: t = 0 : 0.005 : 10; f = 6.5; xc = cos(2 ∗ π ∗ f ∗ t); What is the frequency of the sinusoid (in Hz)? What we are doing here is actually sampling the sinusoid at a rate of 200Hz. Since this is much, much higher than the frequency of our sinusoid, the result will look like a continuous signal when plotted with the ”plot” command. You can use a statement like the one below to generate a version of the sinusoid sampled with sampling period T: n = 0 : T : 10; xs = cos(2 ∗ π ∗ f ∗ n); Use a sampling frequency of 10Hz to generate xs[n]. Use the ”subplot”, ”plot” and ”stem” commands to plot xc(t) and xs[n] on the same figure. Also, plot the frequency response of xc(t) and xs[n] using freqz (or fft and fftshift ) and specify the sinusoid frequency in terms of Hz or normalized frequency. Comment on your results. 2. Repeat the above step for four other sampling rates: 5Hz, 8Hz, 13Hz and 20Hz. Comment on your results. Feel free to changes the number of samples if that helps with obtaining better plots. 3. Using the original sampling frequency of 10Hz, repeat the first step with the sinusoids fre-quency set to 3.5Hz and then again at 13.5Hz. Is there any difference between the corre-sponding discrete-time signals and the one you generated in step 1? If not, why not? 1
Department of Statistics STATS 763: Advanced Regression Methodology Midterm Test Thursday 14 September, 16:30-18:00 Notes: • This Midterm Test consists of 3 questions on 3 pages and is marked out of 50. • Its result will count for 20% of your grade (or 0% if you do better on the Final Exam). • This is a restricted open book test. You are allowed to consult an A4 sheet of paper covered on both sides with any information. • You are not allowed any calculator, phone, digital watch or earpiece. • Identify the supplied booklet with your name and student number and write your answers in it, clearly labelled. Quantiles that may be useful: χ1(2) ;0.95 = 3.84, χ2(2) ;0.95 = 5.99, χ3(2) ;0.95 = 7.81, χ4(2) ;0.95 = 9.49, z0.975 = 1.96 (standard normal quantile). 1. (20 marks total) Data frame. ed contains data regarding the length of stay (in hours) of 2050 adult (≥ 15 years old) patients from 127 Québec hospitals in Emergency Departments (in New Zealand: Accidents & Emergencies, or A&E), collected over a period in 2002. We have data regarding the ID of the hospital, as well as the age and some comorbid (concurrent disease) conditions in the patients: respiratory condition, cardiac condition and mental condition, each encoded as Yes or No. We are interested in the effect of age on length of stay, and whether this effect depends on the comorbid conditions. We fit a GLM with a Gamma family and a log link to the length of stay, as shown below, adjusting for the Hospital ID and putting age in interaction with each comorbid conditions. Selected output is shown below. > mod1 mod0 0. [4 marks] b) A quasi-binomial model for Yi with an identity link g(μi) = μi , μi > 0. [4 marks] c) A quasi-likelihood model with variance function V (μi) = μi(2)(1 - μi)2 and a logit link g(μi) = log [μi/(1 - μi)], μi ∈ (0, 1). [4 marks] 3. (18 marks total) Answer the following questions: a) Let fac be a factor with two levels and y be some outcome of interest. We fit model1
Project #4 – Creating Wordle Step-by-Step Implement multiple helper functions to recreate the game Wordle. Your Task Your task is to implement a simple version of the game Wordle. This document will contain # steps, instructions for functions to build to help us run our Wordle game. All parts should be written in the same python file with the name Project4.py. Wordle is a type of word guessing game that became popular in 2021. You have 6 chances to guess a 5-letter word, each guess you are given feedback on your guess. If a letter is highlighted green, then that letter is in the word and in the correct spot (index). If a letter is highlighted yellow, then that letter is in the word but it is not in the correct spot. If a letter is not highlighted, then the letter is not in the word in any spot. There are 7 TODO's that will be described below, I recommend attempting them in the order they are presented. Each TODO might be described in a different way, some will be English sentences, some will be flowcharts to convert, and one is almost done but has a few parts you'll need to fill in. Make sure the name of each function matches exactly as the autograder will be testing each function directly. I would recommend reading the whole todo list before you start making any programming. Your individual solutions will all be different, but when something is output, it needs to match exactly or the autograder will yell at you. There will be no autograder for the extra credit parts, so be sure you know it works by testing it yourself before the deadline. You will still be able to submit as many times as you want before the due date, but gradescope will provide no feedback. Like project 3, here are slides breaking this down: Project 4 Instruction Breakdown You must include a comment at the top of your program containing your name, “Project #4”, and the date you turn it in. Save your program as Project4.py. This is the only file that you need to submit to gradescope. Tasks & Testing slides: Project 4 Instruction Breakdown TODO 1 get_word: This function should pick a word from a list of all lowercase options, and return the picked word. It must be called get_word. It should return the word that gets selected. This is the code we have so far, but there are 4 errors that need to be fixed before it will work correctly. import random def getwrd()->str: all_words = ['above','blame','corny','dress','EMAIL','flail','glass','horse','igloo','judge','kebab','lilac','moose','never','olive','prize','query','react','shrub','toxic','untie','vigor','wordleword','yield','zesty'] my_word = random.choice(all_words) return 'hello' Right now our word list is small, but we could grow this by adding more items to the list, importing a very large list from a file, or randomly generating 5 letters. Running the following code should print out the word 'email': if __name__ == '__main__': random.seed(1) print(get_word()) (Hopefully, but random number generation can be weird. if it's not 'email' it should at least be consistent every time you try to run the program) You should write any code you are using to test your program inside this weird if statement! Just continue writing the lines indented below the print(get_word()) line! TODO 2: get_guess() In this function, we want to get input from the user, make it lower case, and return that value. This function must be called get_guess, accepts no parameters, and returns a string. We want to make it lowercase to avoid complications later in the file. Note how if our word is 'hello' but the user types 'Hello' a simple == check would return false as 'h' does not equal 'H'. When using the input function, pass it the string "Your Guess: " so it matches exactly. HINT: To make a lowercase version of a string, use the .lower() string function [info: https://www.w3schools.com/python/ref_string_lower.asp] Example of the function running: Code: Result: #user types 'hello' when prompted guess = get_guess() print(guess) Your Guess: hello hello #user types 'PyThN' when prompted print(get_guess()) Your Guess: PyThN pythn get_guess('hello') ERROR TODO 3: print_guess In this function, we want to accept a string as a parameter to the function, and print it out (should return nothing). The print out should be all on a single line, and between each letter there should be two spaces. Your function must be called print_guess and accept one value as the parameter, and return nothing. Here is an example header to get you started: def print_guess(value:str)->None: Example of the function running: Code: Result: print_guess("hello") h e l l o #user types 'PyThN' when prompted x = print_guess(get_guess()) print(x) Your Guess: PyThN p y t h n None print_guess() ERROR TODO 4: get_diff For this function you should follow the flowchart provided below. Make sure your function is named get_diff, it should take two values as parameters (the first is the answer word and the second is the guessed word), and should return a list of strings. The goal of this function is to create a list of five elements. Each element should be a single character either 'r', 'y', or 'g'. If the first character of the guessed word matches the first character of the answer word, the first element of the list should be 'g'. If the second character of the guessed word doesn't match the second character of the answer word, but it is somewhere else in the answer word, the second element of the list should be a 'y'. If the third character of the guessed word is not in the answer word at all, the third element of the list should be a 'r' Once we finish checking the first character of ans against the first character of guess, we need to increase the index variable i by 1, this needs to be inside the while loop to avoid an infinite loop (note the block i=i+1 is inside the green section. Code: Result: diff = get_diff("hello", "hello") print(diff) ['g','g','g','g','g'] diff = get_diff("pleat", "plate") print(diff) ['g','g','y','y','y'] print(get_diff("hello", "rodeo")) ['r','y','r','y','g'] print(get_diff("corny", "email")) ['r','r','r','r','r'] TODO 5: print_diff The function print_diff will accept a list, and print out the correct series of '��' , '��' , '��' emojis, and return nothing. Corresponding to a 'r', 'y', and 'g' in the list. The print out should have a single space between each emoji. Note that to print an emoji in python, it's just like printing any other string: print(" �� ") will print �� to output. Our code will look somewhat similar to TODO 4's get_diff, and you can use that flowchart as a starting point. One approach to solving this would be to create a blank string named to_print, loop through every element in the given list, and add the appropriate emoji using an if-elif-else branch to the end of to_print. Then once the loop has ended, a simple print(to_print) will have the desired effect. Code: Result: print_diff(['r', 'g', 'y', 'y', 'r']) �� �� �� �� �� print_diff(get_diff('hello','hellh') �� �� �� �� �� print_diff('yggyg') �� �� �� �� �� TODO 6: play_turn Create a function called play_turn. The function should accept one parameter, the word that the player is trying to guess. It should return True if the player guessed correctly, and should return false if the player did not guess correctly. You can use the definition of the function below to get started. The goal of this function is to run a single turn of our wordle game. We should get the guess from the user, print that guess out to them formatted correctly, create the diff list for the guessed word and the correct word, print out that diff list with correct formatting, and finally return True if they guessed exactly correctly, or False otherwise. Remember you can use the functions you have already created by calling them. We can do this inside our play_turn function. This specific function should be rather small bool: Code: Result: play_turn('email') Your Guess: ialme i a l m e �� �� �� �� �� x = play_turn("hello") print(x) Your Guess: heros h e r o s �� �� �� �� �� False print(play_turn("bowie")) Your Guess: bowie b o w i e �� �� �� �� �� True TODO 7: play_game Create a function called play_game. It should accept no parameters, and should return none (though many things will be printed out during execution). This function will bring all the other functions together in order to play an entire wordle game. It picks a word, and keeps playing turns until the player guesses the word, or runs out of turns. You can use the following flowchart as a starting point, although there are several missing parts you will have to figure out for yourself, marked by ??? in the diagram. I'd recommend copy and pasting the prints, so they match exactly what the autograder is expecting. print(f"Out of turns, the word was: {correct_word}") print(f" TURN {turn}:") print(f"you've won! it took {turn} turns") IF YOU ARE GETTING AN EOF ERROR: MAKE SURE YOU COMMENT OUT ALL THE TESTING CODE YOU WROTE BEFORE SUBMISSION. YOUR SUBMISSION SHOULD JUST BE THE FUNCTIONS NOT ANYTHING RUNNING THE FUNCTIONS Testing slides ( ~slide 19): Project 4 Instruction Breakdown Test your program completely. Run it several times: verify that any of the extra credits you have implemented are working, and the game plays how you are expecting it to. None of the extra credits will be graded by the autograder, they will all be graded by hand. YOU MUST COMMENT OUT ANY CODE CALLING YOUR FUNCTIONS, YOUR SUBMISSION SHOULD ONLY BE THE FUNCTION DEFINITIONS/IMPLEMENTATIONS. I recommend that you test each individual function as you go, including any tests provided here as well as making your own. You should use the provided tests and outputs as an example to build your own. The autograder for this project will be testing all the functions individually, as well as how they work together. Read the descriptions/outputs for each test in gradescope. If you fail one of them they should offer a little guidance on where to start looking to fix the problem. Code: Result: random.seed(1) #email should be picked play_game() TURN 1: Your Guess: hello h e l l o �� �� �� �� �� TURN 2: Your Guess: claws c l a w s �� �� �� �� �� TURN 3: Your Guess: meale m e a l e �� �� �� �� �� TURN 4: Your Guess: email e m a i l �� �� �� �� �� you've won! it took 4 turns Extra Credit slides (at end): Project 4 Instruction Breakdown There are five places in this project where extra credit can be gained. You may do any, all, or none of these. You will not be penalized, but you won’t get any extra credit, if you attempt an extra credit and it doesn’t work correctly. I recommend at least attempting the extra credit, as some parts will be straightforward, and finishing them will lead to a better completed game in the end. Extra Credit #1 Currently the player is allowed to type in incorrect inputs and potentially crash our program. For this extra credit we will make a new function called get_guess_extra() to tell the user they tried to type in an invalid guess, and allow them to try again until a correct guess is given. For now, incorrect guesses are any guesses that aren't exactly 5 characters, and that contain non-alphabetic characters. Use the following flowchart to modify your original code for get_guess. Note: .isalpha() is a function that can be called on a string and will return True if the string is only alphabetic characters. "hello".isalpha() produces True, '7 hel ) o'.isalpha() produces False. Extra Credit #2 Currently, the player always has 6 turns to guess the word. We could make a copy of play_game() called play_game_extra2() that accepts a single parameter, the number of turns, that would allow us to make the game easier or harder. We would also have to change the if statement that checks to see if the player just used their last turn. The last thing to change would be to set the number 6 as the default value for the new parameter. If you have having trouble setting a default argument value this article should help: https://www.geeksforgeeks.org/default-arguments-in-python/ Extra Credit #3 Currently whether the player wins or loses the game simply ends. We could modify the code we use to start a game (ie the line play_game()) so that a new game is played immediately after the last game ends. I recommend writing a new function called play_infinite_game() HINT: The simplest solution would be to use an infinite loop somehow to call the play_game function until the user kills the process. Extra Credit #4 (Requires #3) Currently no score is kept keeping track of how many games the player has played/won. Once Extra Credit 3 is completed, we could make another (play_infinite_game4()) to keep track of two variables, games_played, and games_won, and print out those values after a game ends and before the next game begins. To do this, it might make sense to modify what the play_game function returns. Currently it returns None no matter what, we could have it return 1 if the player wins and return 0 if the player loses. Extra Credit #5 Currently only one person can play a game. We could create a new play_game_2_player function that allows for two players to play the game. This function will look largely similar to play_game, with the only difference being how the word to guess is picked. Instead of picking a word from a list, we will allow a user to type in the word that the other player will attempt to guess. We should ensure that the word the player types in is a valid word before continuing on with the function. The word must be 5 characters long, and contain only alphabetic characters (see extra credit #1). Submission When you are done, turn in the assignment via Gradescope, you should submit only your .py file. It must be called Project4.py exactly. You can submit to gradescope as many times as you want before the due date. IF YOU ARE GETTING AN EOF ERROR: MAKE SURE YOU COMMENT OUT ALL THE TESTING CODE YOU WROTE BEFORE SUBMISSION. SO A LINE LIKE: play_game() needs to be commented out Again, make sure you read the results from all the tests if gradescope is showing you some of the tests failed. Grading In the report of your grade, you will see a score and a set of letter codes explaining what you did wrong. If you get 10 points, there will be no associated letter codes. The grading codes A-G are defined and will be the same for all programs that you turn in. A summary of those codes is as follows (see the linked document for a full explanation): A: -10 Student’s name is missing from the top of the program. B: -100 Program cannot run due to syntax errors. C: -100 Program crashes before finishing. D: -100 Program runs to completion but does not solve the intended problem. E: -50 Program runs to completion but does not solve all of the assigned problems. F: -50 Program uses overly advanced methods not covered in class G: -50 Program works correctly, but majorly changes the assignment in order to do it. In addition, penalties for this assignment only will be incurred for the following infractions (which may supersede some of the generic codes listed above): H: -10 Solution deviates too far from the described solution J: -10 Required code is omitted (like omitting a Print statement, for example, either in part or in its entirety). K: -10 Starter code has been changed in such a way that makes the program less functional in some way L: -10 Python shortcuts and/or advanced code is used that makes other parts of the starter code redundant. This supersedes code F above. Each of the three extra credit items are worth +3.33 points added to the score only if implemented correctly. Incorrect implementation will not be penalized.
STATS 763 - 2022 - Midterm test 6 April 2022, 10:00-11:30NZST Question 1 [21 marks total] Wilms’ tumour is a rare childhood cancer of the kidney. Treatment is successful for the majority of patients, but a minority do relapse. An important risk factor for relapse is disease stage (how far it has spread). The following data on all U.S. paediatric Wilms’ tumour patients between 1980 and 1994, inclusively, were collected: • Year: Year of diagnosis, from 1980 to 1994 • Stage: Disease stage (I [least advanced], II, III, IV [most advance]) • rel5: Relapse within 5 years (0 [No], 1 [Yes]), hereafter called ”relapse” . We fit a relative risk model of rel5 on Year*Stage, to capture any secular trend in relapses by disease stage, and obtain the following results: Call: glm(formula= rel5~Year*Stage, family=binomial(link="log"), data=wilms) Coefficients: Estimate Std . Error z value Pr(>|z| ) (Intercept) 52 .59226 38 .93026 1 .351 0 .1767 Year -0 .02767 0 .01960 -1 .412 0 .1581 StageII -107 .15715 52 .12848 -2 .056 0 .0398 * StageIII 40 .62429 50 .38844 0 .806 0 .4201 StageIV -29 .17571 53.37873 -0 .547 0 .5847 Year:StageII 0 .05422 0 .02623 2 .067 0 .0387 * Year:StageIII -0 .02008 0 .02537 -0 .792 0 .4286 Year:StageIV 0 .01525 0 .02687 0 .568 0 .5703 Selected rows and columns from the estimated variance matrix of the coef- ficient estimates are given below: (Intercept) Year StageIV Year:StageIV (Intercept) 1515 .6 -0 .763 -1515 .6 0 .763 Year -0 .763 0 .000384 0 .763 -0 .000384 StageIV -1515 .6 0 .763 2849 .3 -1 .434 Year:StageIV 0 .763 -0 .000384 -1 .434 0 .000722 (a) [12 marks total] We replace Year in the model by Year1980 0. (a) [4 marks] Explain why wi = 1 when the canonical link is used. (b) [4 marks total] What is wi in the following settings? i. [2 marks] Variance function V (μ) = μ2 , μ ∈ R+ and link function g(μ) = log(μ). ii. [2 marks] Poisson(μ) family and identity link. (c) [4 marks] How do we usually estimate φ? Write an expression for the estimator. (d) [4 marks] For a certain value of β0 , you are given the observed values (this last subscript means “evaluated at β = β0 ”). Write down a test statistic that you can approximately compare to a χp(2) quantile to test H0 : β = β0 vs H1 : β ≠ β0 . Question 3 [8 marks total]+4 bonus Answer the following questions: (a) [4 marks] You have written a scientific paper containing results from a linear regression model E[YjX = x] = xβ fitted to independent count data. The data set was large and you estimated the variance of β(^) using a sandwich estimator. A reviewer writes that count data are not normally distributed, and there-fore your Wald confidence intervals are incorrect because 1)β(^) is therefore not normally distributed either and b) the variances are wrong because they are estimated under the wrong model. How do you respond? (b) [2 marks] Describe one situation in which a quasi-likelihood model may fail to produce reliable standard errors. (c) [2 marks] True or False: Data sampled according to the outcome will yield biased regression estimates unless it is appropriately weighted. (d) [4 marks] (bonus) A regression coe cient estimated from a parametric gen- eralised linear model has covariance matrix Cov(β(^)) = φ(XTX)-1 . Find two combinations of family and link that will yield this covariance.
2024-2025 COURSEWORK PART II (Second, Third and Final Year) MANAGEMENT SCIENCE MSCI 212 Statistical Methods for Business Tasks [Worth 100% of the marks] You currently work for a multinational company. The marketing department has a dataset that con- tains different budgets for marketing campaigns delivered across YouTube, Facebook, newspapers, and magazines for 170 products. You, as their business analyst, have been asked to develop a re- gression model to assess the relationship between different marketing channels and the sales. You also need to use the model to predict the sales from marketing campaigns for two other products. The dataset is already prepared (see file MarketingData.sav, adapted from kaggle.com). The data description is as follows and budget expenditures are in thousands of US dollars: • Sales – sales of each product; • YouTube – marketing budget spent on the product’s YouTube campaign; • Facebook – marketing budget spent on the product’s Facebook campaign; • Newspaper – marketing budget spent on the product’s Newspaper campaign; • Magazine – marketing budget spent on the product’s Magazine campaign. 1. Preliminary analysis [20 marks] (a) Carry out a preliminary analysis of the data using scatterplots, and comment on the findings between the dependent and the explanatory variables and between the ex- planatory variables. (b) Extend the analysis from (a) using correlations. Comment on the findings as well be- tween the dependent and the explanatory variables and between the explanatory vari- ables. 2. Modelling [30 marks] (a) How would you evaluate the quality of a multiple linear regression model? (b) Use stepwise regression starting with an “no-variable” model and identify “the best” model based on your answer in (2.a). Justify your answer. (c) Repeat (2.b) but use stepwise regression starting with an “all-in variable” model. Identify the “best” model based on your answer in (2.a). Justify your answer. (d) Based on your answers in (2.b) and (2.c), compare and discuss which model is the best. Write down your final/preferred model in an equation. 3. Residual diagnostics [20 marks] Based on your final model in (2.d), you need to produce the residual errors and assess whether they follow the assumptions of a linear regression model. (a) Plot a histogram of the residual errors. Comment on any observations about the his- togram. (b) Conduct a for normality on the residual errors. Interpret your results. (c) Conduct analysis to check whether the errors are homoscedastic. (d) Plot your errors with each explanatory variable and the predicted values. Comment on each relationship. (e) Summarise your findings on the residual diagnostics and compare them with the linear regression model assumptions. 4. Interpretation [10 marks] (a) Interpret your final model and discuss the significance of each variable in detail. (b) Assess the quality of the model. Interpret the criterion accordingly. 5. Predictions [20 marks] Your company has developed two new products, and they have suggested marketing cam- paign budgets across each marketing channel. The budgets for each channel and product are stated below: Product YouTube Facebook Newspaper Magazine Product X 100 22 50 3 Product Y 400 80 150 15 (a) Use your final model: i. to produce point predictions. Comment on their interpretation. ii. to produce the 95% confidence intervals. iii. to produce 95% prediction intervals. (b) Comment on the predictions, taking into consideration the results of residual diagnostics and the descriptive statistics of your data.
Principles of Banking - N1577 Seminar 8. Central Banking Question 1. What are the similarities and differences between European Central Bank and Federal Reserve System? Question 2. What are the advantages and disadvantages of quantitative easing as an alternative to conventional monetary policy when short-term interest rates are at the zero lower bound? Question 3. Referring to the article “The European Central Bank as a lender of last resort”, explain why ECB was not able to fully play its LOLR role during the Eurozone crisis?
Arch1102_T3 2024: Architectural Design Studio Two Project Three: Gallery Tower Ensemble + Bicycle Parking & Workshop + Kiosk in Courtyard PROGRAM Program for the Kiosk is as for Project One Program for Bicycle Parking & Workshop is as for Project Two Program for The Gallery Tower Ensemble (see site map on page 4) Gallery for the display of student artworks in progress - works on canvas or paper, sculpture, film. Staff: one person Provide: 1. Two display spaces - connected vertically. Each space is 20 meters square (min). o The vertical distance between the two spaces: 3 meters (min., not including external walls). o Include one internal staircase: it could be the existing stair with moderation or the proposed new stair. o Skylight: 1 min. 2. Storage - 20 meters square (min. not including external walls) Consider delivery of artwork from campus studios. Gallery operating hours are 10 am - 5 pm (only closed on public holidays). The building must be closed/secured after operating hours. Users: Art school students/staff and the general public (consider access from the main entrances on the campus). Think of the project as a series of architectural situations centring around verticality vs horizontality: o Spaces for viewing and studying art o A window to the sky o A stairway Rules/Limits Unless otherwise advised by your design tutor - work closely with your proposal for projects one and two. Revise/ redesign this proposal to: A. develop and refine your projects one and two. Address the comments you received during your project two reviews (W8). B. incorporate the gallery tower (consider the context and determine its height). Design on Zone C mainly with possible extensions into all zones at the upper level. • In the design of the internal stair, use risers (150mm max.) and threads (280mm min.). • Use entirely orthogonal geometry - no curves or diagonals. The roof plane may beangled. • Assemble the building using planes, folded planes, and linear elements only. See the definition included below. • Masonry and timber will be the primary materials used, and concrete slabs will be used for the floor. • Timber can also be used for joinery, doors, and windows, and lintels may be used over door and window openings. Continuous flat planes: These are elements that have two of their dimensions of relatively equivalent size compared to their substantially smaller third dimension. The planes must be flat - not curved or folded. They must maintain their surface continuity as much as possible, which means that openings such as doorways, windows, voids, etc. must be subservient to this continuity. Examples are floors, walls, ceilings, roofs, landings, benchtops, screens, and deep sills. Continuous folded planes: Each of these elements is made from a single flat plane that has been folded at least once. The planes must maintain their surface continuity as much as possible, which means that openings such as doorways, windows, voids, etc. must be subservient to this continuity. Examples of folds are extruded L- , U-, Z-, and O- shapes. The elements can be floors, walls, ceilings, roofs, stairs, awnings, balustrades, staircases, and combinations of these. Linear elements: Those in which one of the three dimensions is substantially longer than the other two. Examples are: columns, handrails, stair treads, sills, architraves, picture rails, skirtings, facias, shelves, pipes and the individual elements in screens. Associated Precedents No new precedents. You are required to work with the same set of precedents as for projects one and two. You are asked to visit /research art galleries/museums in Sydney or your location and research five galleries/museums of your selection after W8’s presentation. See Moodle for resources. Tasks + Submission Requirements WK 9 Submit Task 1: Design Elements for Project Three diagrams and sketches of the five galleries/museums selected: plans describing a space for exhibiting artwork, a section showing the connection vertically between two gallery rooms that are linked vertically by a 3m min in-between space; and skylight (how to capture, frame. and filter light?) In Studio: Develop a sketch model for Project Three as directed by your tutor. Draw inspiration from the 5 cases you've selected. Bring model making/modelling and drawing equipment/software to studio class. Have your project two model and/or modelling + ground-level plan handy. WK 10 – Submit Task 2: 1:100 Ground Plan (pdf or hard copy), including the proposed entrance/entering sequence to the gallery in relation to projects one + two. This will be marked as Pass or Failand will be considered in the 20% tutor’s mark. Developed Proposal/draft design for Project Three 1:100 plans and sections 1:100 working model/modelling WK 11 - No Teaching WK 12 - Gallery Tower Ensemble + Project One + Two Final Presentation with jury panel. (40%, Studio Review from 1 to 5 pm; open studio/collective feedback from 5 to 6 pm; and parity from 6 to 7 pm among tutors). Submission requirements: Booklet/Portfolio: Printed documentation of your Project 3 and Projects 1 + 2. Your choice of format and size. Model/Modelling: Massing model @ 1:500 (situated in context model): process models - optional Final model/modelling @ 1:100. Do your best to create a refined model in material, detail, and finish. The roof should be removable, and the interior, including the staircase, should be clearly represented. Show all three projects and adjacent buildings. Construct the selected details of the walls adjacent to the site (B14, B15, B16, & C28). Presentation drawings (minimum requirements): A site plan @ 1:500 showing surrounding context, boundaries adjoining buildings and circulation. A plan @1:100 showing pavement, entrances, and landscape, if any. A long section @1:100 (min)* A short section @1:100 (min)* *Note: one of the sections should include the concerns for comfort and environmental performance of spaces learned from ARCH1161: thermal mass, roofs, openings, building fabrics including walls and windows (CVEN students are exempted) A detail wall section @ 1:20 showing the quality of your exhibition space. This could be a sectional perspective drawing: Please consult with ARCH1162 for the construction details of the wall and floor considering material and design. CVEN students: see study material in Moodle. This is not an assessment item for you. A rendered view of your final project (min): it can be a sectional perspective, rendered exploded axonometric drawing, and so on. Format: A2 size/proportion. You can organise your preferred panel layout via InDesign. Medium: Rhino (or preferred equivalent software), SketchUp + Illustrator, & Photoshop. You can choose your preferred mode of representation. PDF: 2 files onto Moodle: 1) Submit all your drawings and photographs of your models as a single PDF file; 2) submit your booklet/portfolio as a single PDF file. Uploaded by 10 pm 26 Nov. Criteria for Review & Assessment: Project 3 assessment is worth 40% of the total course assessment. You will be assessed on the following criteria: • Compliance with the prescribed limits, including the definition of ‘Tower’; • Clarity of architectural strategies as they have been drawn from precedents and their translations into relevant design strategies and spatial tactics; • Clearly articulated and resolved the relationship between the topics of siting, enclosure and materials in reference to the program; • An elementary understanding of methods of assembly and construction demonstrated in the 1/20 detailed wall section and the environmental concerns in one of the 1/100 sectional drawings; • Precision and clarity of architectural representation. Site for Project 3: The entire courtyard. The existing staircase can be preserved and modified or removed.
F24 ECE 551 HW03, Due 11PM Thu. Sep. 26 Pr. 1. (a) Let Q denote an M × K matrix having orthonormal columns. Show that Q′Q = IK. Thus ∥Qx∥2 = ∥x∥2 for all x ∈ C K. (b) Show that the following converse is true: if A is an M × K matrix for which ∥Ax∥2 = ∥x∥2 for all x ∈ C K, then A′A = IK. Your proof should be general enough to cover the case where A has complex elements. Hint. Examine products with standard unit vectors ej and combinations like ej + ek, or consider an eigendecompo-sition of A′A − I. Pr. 2. Let A be an N × N Hermitian matrix with unitary eigendecomposition A = QΛQ′ , where λ1 ≥ λ2 ≥ . . . ≥ λk ≥ 0 and 0 ≥ λk+1 ≥ . . . ≥ λN , for some 1 < k < N. Determine an SVD of A in terms of the given components. Pr. 3. The Ch. 3 notes show the following Venn diagram of matrices and eigendecompositions: Each of the categories shown above is a strict superset of the categories nested with in it. Provide example matrices A1, . . . , A5 that belong to the each of the categories above but not the next category nested within it. Try to provide the simplest possible example in each case. For example, the matrix A1 = [0 1] is rectangular, but not square. Now you do A2, . . . , A5. Hint. All the examples can be 1 × 1 or 2 × 2, often with simple “0” and “1” elements. Pr. 4. Let (a) Determine the null space of A, denoted by N (A), and the range space or column space of A denoted by R(A). (b) Are they equal? Does your answer hold in general? If not, provide a counterexample. Pr. 5. Let A ∈ FM×N . (a) Suppose W ∈ FM×M and Q ∈ F N×N are each unitary matrices. Show that A and C ≜ W AQ have the same singular values. Consequently, A and C have the same rank, the same Frobenius norm and the same operator norm. This is why the Frobenius norm and the operator norm are called unitarily invariant norms. Their value does not change when the matrix is multiplied from the left and/or the right by a unitary matrix. Any norm that depends only on the singular values of A will, by definition, be unitarily invariant. (b) Suppose that W and Q are nonsingular but not necessarily unitary matrices. Do A and D ≜ W AQ have the same rank? Prove or give a counterexample. (c) Continuing (b), do A and D = W AQ have the same singular values? Prove or give a counterexample. Pr. 6. Prove that the orthogonal complement of the range of a matrix A equals the null space of A′ : R⊥(A) = N (A′). Pr. 7. Let A = xy′ , where neither x nor y is 0. (a) How many linearly independent columns does A have? I.e., of all the sets of columns from this matrix where the set is linearly independent, what is the maximum cardinality? (b) What is the rank of A? (c) Enter (cut and paste) this code into Julia: using LinearAlgebra: svdvals using Plots using LaTeXStrings n = 100 x = randn(n); y = randn(n); A = x*y' s = svdvals(A) scatter(s, yscale = :log10, label= "", title = "singular values: log scale", xlabel="i", ylabel=L"sigma_i") # to make σi (Ideally the ylabel should look like σi , but if it does not, it is OK. Comment out the ylabel part if needed.) • How many numerically computed singular values of A are nonzero? • What answer do you get when you type rank(A) ? • Turn in your plot of the singular values and the answers. You will have just seen that a theoretically rank-1 matrix will have multiple nonzero singular values when expressed in machine precision arithmetic; this property is due to roundoff errors (finite numerical precision). (Round-off errors should not be thought of as “wrong answers,” but just a fact of real-life computing to be understood and not feared.) (d) To help locate the code for the rank function, type this into Julia: @less rank(ones(3,2)) Examine the code for the rank function and write down the formula used to determine the threshold (aka tolerance) below which, due to roundoff errors, the code considers the singular value to be “zero.” (e) Determine the numerical value of the threshold when A = [9 9]. (f) Optional. Check your answer to the previous part by using the Julia debugger to step into the rank function and examine the tolerance. Pr. 8. When D is an M × N (rectangular) diagonal matrix, its pseudoinverse D+ is an N × M (rectangular) diagonal matrix whose nonzero entries are the reciprocals 1/dk of the nonzero diagonal entries of D. A matrix A having SVD A = UΣV ′ has A+ = V Σ+U′ . Working with just this equality, determine by hand (experimenting / checking with code is OK) the pseudoinverse of (a) A = xy′ , where neither x ∈ FM nor y ∈ F N is 0, (your answer should depend only on x and y), (b) A = xx′ , where x ≠ 0. Pr. 9. (a) Find an orthonormal basis for the range of the M × N matrix that is all ones. (b) Find an orthonormal basis for the range of (c) Find an orthonormal basis for the range of any 2 × 2 rotation matrix. (d) Use the fact that R(A) = R(Ur) to modify the compact SVD code in §4.5 to write a function that returns an orthonormal basis for the range of a given matrix, akin to how nullspace returns a null space basis. In Julia, your file should be named matrixrange.jl and should contain the following function: """ matrixrange(A::AbstractMatrix) Return orthonormal basis matrix for the range of matrix `A`. """ function matrixrange(A::AbstractMatrix) Email your solution as an attachment to [email protected]. Check your code using your answers to the preceding parts. (e) Optional. Find an orthonormal basis for the range of any 2 × 2 rotation matrix, such that the basis matrix has the smallest possible trace. Non-graded problem(s) below (Solutions will be provided for self check; do not submit to gradescope.) Pr. 10. Let A be a normal matrix of the (unnamed) type where each eigenvalue either has a different magnitude than all other eigenvalues, or has the same value as all other eigenvalues with its magnitude. In other words, having both |λj | = |λi | and λj = λi is not allowed. For example, A might have eigenvalues (3, 4ı, 4ı, −5) but cannot have eigenvalues (3, 4, 4ı, −5). Prove that every right singular vector of A is also an eigenvector of A. This problem finishes the story on relating SVD and eigendecomposition. Pr. 11. We write A ⪯ B iff B − A ⪰ 0, i.e., iff B − A is positive semidefinite. For a > 0, determine Pr. 12. Suppose that X = UxΣxVx ′ and Y = UyΣyVy ′ denote SVD s of the matrices X ∈ F m×n and Y ∈ F p×q . (a) Show that X ⊗ Y = (Ux ⊗ Uy)(Σx ⊗ Σy)(Vx ⊗ Vy) ′ is an SVD of X ⊗ Y (up to a permutation of the singular values). The block diagonal matrix Σx ⊗ Σy contains the pairwise products of the singular values of X and Y . Hint. First show that the formula is correct, then argue that it is an SVD by showing that Ux ⊗ Uy and Vx ⊗ Vy are unitary matrices. Hint: You may find the following properties of Kronecker product useful: • (X ⊗ Y ) ′ = X′ ⊗ Y ′ • (X ⊗ Y )(W ⊗ Z) = (XW) ⊗ (Y Z) (b) Now suppose that A = QaΛaQ′ a and B = QbΛbQ′ b are unitary eigendecompositions of the (normal) matrices A ∈ F N×N and B ∈ FM×M. Show that A ⊗ B = (Qa ⊗ Qb)(Λa ⊗ Λb)(Qa ⊗ Qb) ′ is a unitary eigendecomposition of A ⊗ B. The diagonal matrix Λa ⊗ Λb contains the pairwise products of the eigenvalues of A and B. Hint. First show that the formula is correct, then argue that it is an eigendecomposition by showing that Qa ⊗ Qb is a unitary matrix. (c) Continuing (b), show that A ⊕ B = (Qa ⊗ Qb)(Λa ⊕ Λb)(Qa ⊗ Qb) ′ is an eigendecomposition of the Kronecker sum A ⊕ B ≜ (A ⊗ I) + (I ⊗ B), where you must determine the size(s) of the “I” in that expression. The diagonal matrix Λa ⊕ Λb contains the pairwise sums of the eigenvalues of A and B. Hint. Start by writing IN = QaIN Q′ a and IM = QbIMQ′ b in the definition of A ⊕ B, and then apply (b). Pr. 13. Complete the “Introduction to Matrix Math” tutorial at https://pathbird.com/courses/register. (Use the registration code shown on the ECE 551 Canvas home page.) This tutorial should be helpful for anyone who wants deeper understanding of Julia for operations with matrices and vectors. Such operations are a major part of ECE 551.
MENG 4019 - Practical 2 – 2022 Task: design and simulate the operation of an electro-hydraulic curcuit. A hydraulic cylinder, stroke 2000 mm, D piston 100 mm, d rod 50 mm angled 60 degrees from horizontal is lifting a mass load of 6000 kg, with a push external force of 10 kN and pull external force of 6 kN. The cylinder needs to extend/retract as needed. Manually, then controlled by an electrical circuit. First, we build a conceptual circuit: 1. Open Automation studio, select and insert the following components from the Hydraulic set of components 2. Click on Filter to select, Right click on Filter, click on Lock Size to unlock, reduce size of Filter (drag it to a smaller size), Right Click on Filter, click on Lock Size to lock 3. Connect circuit as shown below: 4. Select the Simulation tab and run a Normal Simulation. Then, Stop Simulation and run a Slow-Motion Simulation 5. This shows that the circuit is correct, and it works. The cylinder can extend and retract, s needed. The components are all connected, there is no error reported by the system. The flows and pressures are consistent with the settings of the components. 6. We need to modify the control of the valve, to change it from a manual control to a solenoid. Double click the valve and select Technical Specifications from the left lower tab. Then, double click on the manual lever, select solenoid and confirm selection Using the arrows from the top, center the solenoid to the middle of the valve: 7. Start creating the electric control circuit. From the Electrical Control (JIC Standard), insert the power source (24 V and 0 V), the start button (push button, name it START) and a solenoid (call it SOL). Connect the components. 8. We see that the hydraulic circuit works (we can change the position of the valve manually), and the electrical circuit works – we can test by clicking on the Push Button (START). But the two circuits are not connected. 9. We need to connect the two circuits Double click on the Valve. Select Variable Assignment from the left lower tab. Connect the solenoid of the valve with the solenoid from the electric circuit (SOL) Double click on SOL on the top right panel. The two solenoids are connected (common name): NOTE: make sure the valve is not in the manual control position: If this is the case, click on the solenoid, the red tick in the valve disappears and the circuit can be controlled electrically. Now, if you press the START button, the valve switches: 10. Adjust details of the actuation based on the given data: You can rotate the cylinder to reflect the details of the problem – right click on cylinder and select Coordinates and Orientation: 11. We want to ensure that, once we press the START button, it will continue to work without having to keep it pressed. We insert a normally open contact that we connect to the solenoid and a coil: We make the connections. We check the circuit. It works: However, we want to be able to reverse when we need. We insert a Normally Closed Push Button, and we check the circuit. It works: 12. We want to insert a small automation in the circuit so that, when the cylinder reaches the end, the solenoid is de-energised, and the piston is retracting. We need to insert a Proximity Sensor at the end of the stroke. Insert a proximity sensor (call it A1), rotate it 60 degrees and place it at the max stroke of the cylinder (extend it to 100% to facilitate placement): 13. Add a Normally Closed Proximity Switch from the Electrical Control (JIC Standard) and link it to the Proximity Sensor A1 Check that the circuit works. All seems correct: 14. We want to automate the circuit so that once we start it, it will extend, retract and continue the cycle until we stop it. We need another Proximity Switch and we need to integrate it in the circuit: Due to the grid resolution, it is difficult to select the exact position of the cylinder we want (0 and 100% extension, so we revert to the normal position of the cylinder. We insert another Proximity Switch on the cylinder and a Normally Open Proximity Switch in the circuit. We link it to A0. We check that it works. All is correct, except the system only runs one cycle and stops. 15. We replace the START Push Button with a Normally Open Toggle Switch. We check that it works. All is working as expected:
Electrical Machines (EE4003) In-class questions Q1 A three-phase, four-pole, 50-Hz, 200-kVA, 1100-V, star-connected induction motor has the following design data: Specific magnetic loading = 0.6 T Specific electric loading = 25000 A-conductors/m Full load leakage impedance drop = 10% of rated voltage Slot fill factor = 0.7 Current density of conductors = 5 A/mm2 A square pole design is used. The winding is full-pitched and uniformly wound. (a) Deduce the formula for determining the output coefficient G. (b) Determine the diameter of the air gap and the axial core length. (c) If there are 48 slots in the stator and all conductors in the same phase are connected in series, estimate the number of turns per phase and the total cross-sectional area of each slot. (Note: Slot fill factor = cross-sectional area of all conductors per slot/ total cross-sectional area per slot)
PHAY0033 Formulation of Small Molecules 1) Answer BOTH parts of this question Onychomycosis is the most common nail disorder. Dr Murdan at the UCL School of Pharmacy has synthesised a new anti-onychomycotic drug molecule. Professor Basit has mixed the drug with excipients and prepared a formulation that will be directly printed onto patients’ onychomycotic nails. a) Discuss how the properties of the drug molecule might influence its ungual permeability. (50% of marks) b) Discuss how the properties of the formulation might influence its efficacy. (50% of marks) 2) You have been asked to develop an immediate relase tablet formulation containing 250mg of drug X. The drug is freely soluble across the entire pH range of the gastrointestinal tract. Propose, with appropriate justification, a suitable tablet formulation for drug X. (100% of marks) 3) Explain TWO (2) formulation factors affecting oral transmucosal bioavailability. Give ONE (1) example of a formulation approach to improve oral trasmucosal absorption. (100% of marks) 4) Answer both parts of this question: a) Describe the two main formulation principles for dry powder inhalation (DPI), including in your answer the roles of forces of adhesion and cohesion (50% of marks) b) A company formulating a DPI formulation buys milled drug from a supplier. The supplier changes its mill to a newtype; it assures that the particle size distribution of the milled drug remains the same as before, but when formulated the performance of the DPI product is reduced. Explain why this might be the case and what you might do to improve the situation. (50% of marks)
COMPUTER SCIENCE COMPSCI 753 SEMESTER TWO 2020 Algorithms for Massive Data 1 Locality-sensitive Hashing [10 marks] 1.1 Computing MinHash signatures [5 marks] Given 4 sets: S1 = {3, 4, 5}, S2 = {0, 1, 2}, S3 = {0, 5}, Q = {0, 1, 2, 3, 4, 5}. 1. Present these sets as a binary matrix where the set elements are {0, 1, 2, 3, 4, 5}. [1 mark] 2. Construct the MinHash signature matrix using 4 universal hash functions below. [1 mark] h1 (x) = x mod 6, h2 (x) = (x + 1) mod 6, h3 (x) = (x + 3) mod 6, h4 (x) = (x + 5) mod 6 3. Consider the set Q as the query set, estimate the Jaccard similarities J(S1, Q), J(S2, Q), and J(S3, Q). [1 mark] 4. Can we use the hash function in the form of h(x) = (x + a) mod 6, where a is an integer to simulate random permutations for our sets? Explain you answer. [2 marks] 1.2 Tuning parameters for LSH [2 marks] Given the number of bands b and the number of rows per band r, let p = 1 - (1 - sr )b be the probability of being a candidate pair for the pair with Jaccard similarity s. Given the following values of r and b: r = 3 and b = 10; r = 6 and b = 20; r = 5 and b = 50, we compute the value p for s in range {0.1, 0.2, . . . , 1} as follows: s (3, 10) (6, 20) (5, 50) 0.1 0.0100 0.0000 0.0005 0.2 0.0772 0.0013 0.0159 0.3 0.2394 0.0145 0.1145 0.4 0.4839 0.0788 0.4023 0.5 0.7369 0.2702 0.7956 0.6 0.9123 0.6154 0.9825 0.7 0.9850 0.9182 0.9999 0.8 0.9992 0.9977 1.0000 0.9 1.0000 1.0000 1.0000 We would like to solve the near neighbor search problem using the Jaccard similarity. In particular, given a query set Q, we want to find all sets Si such that J(Si, Q) ≥ 0.5. Which settings of b and r above should we use such that: 1. The probability that any 50%-similar pair is a candidate pair is at least 70%. Explain your solution. [1 mark] 2. The probability that any 50%-similar pair is a candidate pair is at least 70% and the number of candidate pairs is minimized. [1 mark] 1.3 Linear time of LSH on finding all similar pairs [3 marks] Assume that the average number of words in a document is constant. Without using the shingling technique, the running time of the na finding all Jaccard similarity pairs is O(n2 ) where n is the number of docu- ments. In the lecture note, we state that “With LSH, we can approximately find all similar pairs in O(n) time.” Is the statement true or false? Explain your answer. 2 Streaming Algorithms [15 marks] 2.1 Reservoir Sampling [5 marks] Fig. 1.: Illustration of how pRS1 works. In our lecture, we have studied the reservoir sampling which samples an element from a stream of size m with the same probability. If we use the reservoir sampling with the summary size s = 1, each element of a stream will be sampled with probability 1/m. We name this method as RS1 . The generalized version of reservoir sampling with the summary size s > 1 guarantees that each element in a stream will be sampled with the same probability s/m. We name this method as RSs. In the exam, we consider a new algorithm, called pRS1 , that simulates RSs for s > 1 by running s independent RS1 instances in parallel. pRS1 also uses a summary of size s, as shown in Figure 1. 1. As a function of m and s, what is the probability an element of a stream is sampled by pRS1 ? [2 marks] 2. Let fi be the number of occurrences of the element ai in a stream. Explain how we can use RS s and pRS1 for estimating fi. [3 marks] 2.2 Misra-Gries vs. Reservoir Sampling [5 marks] Run the Misra-Gries summary with k = 2 counters for the stream below: {1, 4, 5, 4, 4, 5, 4, 4} 1. Present the final summary, including the elements and their counter values, when the execution of the algorithm is finished. [2 marks] 2. What is the frequency of an element such that it is guaranteed that it would be in the summary. [1 mark] 3. If we use the generalized reservoir sampling RSs with s = 2 on this stream,what is the probability that the element 4 is in our RSs summary? [2 marks] 2.3 CountMin Sketch [5 marks] Apply CountMin Sketch to estimate the frequency of each element in the stream below: {1, 4, 5, 4, 4, 5, 4, 4, 1, 4, 5, 4, 4, 5, 4, 4, 1, 4, 5, 4, 4, 5, 4, 4} Our CountMin Sketch uses d = 3 arrays with the hash functions as follows: h1 (x) = (x + 1) mod 3, h2 (x) = (3x + 1) mod 3, h3 (x) = (5x + 2) mod 3. 1. Present the CountMin Sketch summary after processing all elements and the estimated frequency of each element. [3 marks] 2. Among Reservoir Sampling, Misra-Gries and CountMin Sketch, which algorithm we should use to find the top-1 frequent element in this stream. Explain your choice. [2 marks] 3 Algorithms for Large Graphs [15 marks] 3.1 Computing PageRank [5 marks] Given the following raw adjacency matrix of a graph: 1. Convert A into a column-stochastic matrix M. [1 mark] 2. In the lecture, we have shown that the PageRank r = M · r is the eigenvector of the column-stochastic M corresponding to eigenvalue λ = 1. Compute the PageRank of all nodes in the above graph using the eigen equation. [2 marks] 3. From the resulting PageRank scores in question 2, explain the problem of running the power iteration algorithm r(t+1) = M · r(t) on M. Describe how to solve the problem. [2 marks] 3.2 Girvan-Newman [5 marks] 1. Compute the edge betweenness for all edges in the social network in Figure 2. Which edge will be removed to partition the graph into two parts using the Girvan-Newman method? [3 marks] Fig. 2.: An example social network 2. In our lecture, we mentioned that we can use the Brandes’ algorithm to calculate the shortest path from a node to all others. Does the algorithm apply for a weighted graph? Explain your answer. [2 marks] 3.3 In uence Maximization [5 marks] 1. Compute the influence spread of the seed set S = {a} using the inde- pendent cascade model on the graph in Figure 3. Hint: Convert the stochastic graph to deterministic graphs. [3 marks] Fig. 3.: A social network with activation probabilities on edges. 2. In the lecture, we gave the definition of submodular function as f(S [ {v}) - f(S) ≥ f(T [ {v}) - f(T) for S C T ≤ U, where U is the set of all items. Another definition of submodular function is that f(A) + f(B) ≥ f(A [ B) + f(A ∩ B) for any two sets A, B ≤ U. Show the two definitions are equivalent. [2 marks] 4 Recommender Systems [10 marks] 4.1 Collaborative Filtering [6 marks] Given the following transactions in the form of (user, item, rating) tuples in a recommender system. (u1, p1, 1.5), (u1, p3, 4), (u1, p5, 0.5), (u2, p2, 4), (u2, p4, 2), (u3, p1, 4.5), (u3, p4, 2.5), (u3, p5, 5), (u4, p2, 2), (u4, p3, 3.5), (u4, p4, 4), (u4, p5, 2.5) Let the set of users be {u1, u2, u3, u4} and the set of items be {p1, p2, p3, p4, p5}. 1. Construct the user-item interaction matrix based on the above transac- tions. Use question marks to denote missing values. [1 mark] 2. Apply the basic user-based collaborative filtering with the Pearson corre- lation coe cient for user u2 without considering bias. Give the top-1 rec- ommended item to u2 and the corresponding predicted rating. [2 marks] 3. In the lecture, we discussed how to model the rating bias including the bias over all transactions, the bias of a user and the bias of an item. Give the predicted rating of user u2 to itemp5 using the collaborative filtering that incorporates the above bias information. [3 marks] Note: The predicted ratings should round to one decimal place. 4.2 Factorization Machine [4 marks] Suppose you are asked to build a system to recommend events. Users {u1, u2, u3, u4} attend events from {e1, e2, e3} in groups. Events are held in one of the two stadiums s1 and s2. Table 1 shows the transactions. Table1.: Transactions of the event recommendation system Transaction ID Group of users Event Stadium 1 u1, u2 e1 s1 2 u1, u3, u4 e2 s1 3 u2, u4 e3 s2 4 u3, u4 e1 s2 1. Construct the input feature vectors for the factorization machine using the event transactions in Table 1. [1 mark] 2. Can factorization machine predict the rating that an individual user u2 may put on e2 held in s2? Explain your answer. [1 mark] 3. If we ignore the stadium information in the transactions and only consider users and events in the above example, does the factorization machine reduce to the latent factor model? If yes, explain your answer. If no, explain in what situation the factorization machine reduces to the latent factor model. [2 marks]
IFB208TC First Part of Semester 2, AY2023-24 Quality Management and Lean Six Sigma Assessment 001 (hereafter referred to as A001) consists of an individual written report that carries 65% of the final mark for this module (see Table 1 for additional details). This written report explores several aspects of quality management that have been covered in class. Table 1. Assessment 001 - Overview Sequence Method Assessment Type (EXAM or CW) Learning Outcomes Assessed Week % of A001 Mark A001 Individual Report CW A & B Apr 21, 2024 65%* *Maximum mark for this assignment is 100 points. YOUR TASK You need to solve three quality management-related questions. Each question carries a certain number of points and has its own specifics - i.e., the questions are NOT intertwined. Given this, you are to read each question carefully and provide a detailed answer in line with the pre-set word limit. There are a number of requirements that are valid for all questions part of this assignment. These requirements areas follows: • Language: British or American English (*be consistent across the text) • Main text: Times New Roman 12; space between lines 1.5; footnotes allowed, but make sure to number them consecutively throughout the report. • Referencing: all secondary information sources (both offline and online ones) should be cited according to the Harvard (Referencing) Style (*see next page for further information). • Similarity Rate: no more than 20% (*excluding Title Page and References). • Submission Method: online, PDF or WORD format (*PDF is preferred) via Learning Mall Online • Submission Deadline: April 21st, 2024, before 18:00 Beijing time. • Penalty Policy for Late Submissions: XJTLU’s policy for late submissions will be strictly followed. QUESTIONS SECTION 1. Garvin’s Eight Dimensions of Quality (25 points) (250-350 words) In line with Garvin’s Theory on the Eight Dimensions of Quality and taking into account each dimension’s definition, you need to answer the following question: Q1. Which are the most abstract dimensions of (product) quality as per Garvin’s Theory and why? Weight: 25 points - Choice (15 points), Argumentation of Choice (10 points) SECTION 2. Product Quality & Business Strategy (25 points) (500-600 words) The quality management literature argues that product quality is a competitive advantage that many companies continue to focus on. With this in mind, you need to answer the following question: Q2. What is the relationship between (product) quality and market share? Weight: 25 points - Answer (10 points), Argumentation of Answer (15 points) *Make sure to support your argument by referring to relevant quality management theories and examples from industry. SECTION 3. Product Quality & Manufacturing Method (30 points) (500-600 words) The quality management literature argues that, in most cases, handmade products command a higher price tag than mass-manufactured products. With this in mind, you need to answer the following question: Q3. What makes handmade products more expensive than mass-manufactured ones? Weight: 30 points - Answer (15 points), Argumentation of Answer (15 points) *Make sure to support your argument by referring to relevant quality management theories and examples from industry. INSTRUCTIONS TO STUDENTS REPORT STRUCTURE COVER PAGE - your Cover Page should be the same as the Taicang Official Cover Page. Make sure, however,to fill in all pertinent details such as (i) module code and title; (ii) student name and ID; (iii) title of your coursework, (iv) word count, etc. Besides the Taicang Official Cover Page, you are also to include a separate Title Page with the following information: (i) module code and title; (ii) student name and student number; (iii) title of your coursework; (iv) module leader. Feel free to design the title page according to your preferences while ensuring it adheres to the professional standards typically followed in business and management reports. MAIN BODY - the report should embrace a question-answer structure. In other words, you must NOT combine either the questions or the answers. Also, you are advised to use sections and subsections. Make sure to identify each section and subsection using Arabic numbers. Sections should be numbered 1., 2., etc., and subsections should be numbered 1.1, 1.2, etc. Importantly, since the questions are not connected, you are NOT expected to include an Introduction and a Conclusion. REFERENCES - all secondary sources including books, journal articles, conference papers, etc., which have been used to produce this assignment must be cited accordingly. This coursework requires that you use the Harvard (Referencing) Style also known as the Harvard Referencing System. If you are not familiar with the Harvard Referencing System, please, make sure to consult the Internet for additional details. For instance, you may check the following source: https://www.mendeley.com/guides/harvard-citation-guide/. Importantly, any secondary sources you use for solving the questions must be listed right after your answer. In other words, you should have three reference sections (one for each question) instead of one. LENGTH - based on the requirements on p.2, the entire document must be between 1,250 and 1,550 words (*excluding title page, references, tables, graphs and figures). Importantly, should you decide to include an Introduction and a Conclusion, both sections will be counted towards the allowed word limit. With this in mind, reports below the minimum or above the maximum threshold will be penalised accordingly. For every 10-20 words less/ more, you will be deducted 1% of the final mark. For instance, if your report stands at 1,550 words, you will be deducted 3% of the final mark.
Econ 30041 Workshop 4 1. Poverty traps can also occur at a macro level for countries. Consider the following table from Kraay and Mckenzie (Journal of Economic Perspectives, 2014). (a) What can you deduce about the existence of poverty traps at the country level from the above table? (b) Do you think that low average per capita GDP growth for countries in the second quintile implies the existence of a poverty trap? 2. Suppose an employer need 8000 units of work (in capacity units) to be performed, and they can hire all the labourers that you want. Assume that all income earned by the labourers is paid to them by the employer and that all income is spent on nutrition. The capacity curve for each labourer is described as follows: for all payments up to £100, capacity is zero and then begins to rise by 2 units for every additional pound. This happens until an income of £500 is paid out. Thereafter, an additional pound paid out increases capacity by only 1.1 units, until total income paid is £1000. At this point additional payments have no effect on work capacity. If employer would like to get the work done in minimum cost (i) What would be the wage rate they should pay per unit of work? (ii) Draw the labour supply curve. (iii) If there are 100 employers in the society and 1200 workers, what would be the level of involuntary unemployment rate? 3. Consider the following utility function that individuals maximise U = f α - p(1 - δθ)(B - c), 0 < α < 1, 0 < δ < 1, θ Œ [0,1], subject to income y = 2c + f, where y = h(1 - θ) +P, P Æ 0 . f and c are respectively food and comfort goods, θ is the attention paid to home and p is the probability of error happening at home. h is the human capital individuals are endowed with. Individuals decide on comfort goods and then on the level of θ . [i] Find the optimal value of c. [ii] Show that individuals with high h will always choose θ=0. Assume θ ∈{0,1}. [iii] Find the optimal value of h, where anyone below it will choose θ=1.
CS 2110 - Fall 2024 Project 2: LC-3 Datapath 1 Introduction Welcome everyone to Project 2! Please read through the entire document before starting. Things are often explained in more detail after they are introduced, so reading the entire document first can give you a better understanding of how the document is structured and of the overall goals for this project. Start early, and remember to utilize Office Hours and Piazza as needed. In this project, you’ll be doing the following: • Implementing some basic sequential logic elements (3.1). • Building three missing pieces of the LC-3 hardware (3.2). • Writing the microcode for the bulk of the LC-3 instructions (3.4). 1.1 Latches The first thing you’ll implement are logical elements that allow us to retain data in a circuit, and — in other words — “remember” previous values. This allows us to build more complex systems that depend on current input values along with previous events such as the previous state of the system - in other words, sequential logic. The goal of this section is to produce a register. A register is an edge-triggered sequential logic circuit element that can store a binary value. For more on this section, see Implementation: Latches 1.2 The LC-3 The LC-3 datapath we’ve discussed in class contains a lot of pieces very similar to circuits we’ve seen or even made before (e.g. an ALU, a register file with 8 edge-triggered general purpose registers, a RAM unit, etc.). One piece that we’ve mostly referred to as a “black-box” in the past is the micro-controller. It’s responsible for controlling the entire datapath and getting it to properly execute the instructions that we give it. That’s a big task! So, how does the micro-controller actually work? In this project, we’ll build a few datapath components to develop some familiarity with the LC-3, and then we’ll actually write the “microcode” which allows the micro-controller to function. The micro-controller, shown above, is a finite state machine. It has 59 possible states (holy cow!) and, thus, needs 6 bits to store all its possible states. It also has 49 output bits composed of output flags. 10 of these bits are used to determine the next state and the remaining 39 extend throughout the datapath to control other pieces of the LC-3. That would be a lot of very complex hardware—if it were built entirely in hardware. As it turns out, there is an easier way. We can actually use a ROM (Read-Only Memory) in order to specify the behavior of each distinct state in the state machine (e.g. each instruction will map to a series of entries in the ROM. Each entry in the ROM represents a micro-state, which is an individual state of the finite state machine and a component of a larger sequence of states known as a macro-state. There are 3 macro-states in the LC-3: FETCH, DECODE, and EXECUTE. Each macro-state requires between 1-5 micro-states to complete, depending on the complexity of the instruction. What does a ROM entry look like? We encourage you to go ahead and open up microcode .xlsx, on the microcode sheet, to follow along. A ROM entry is basically a long binary string. The last few bits of it handle the transition to the next state—you don’t need to worry about this at all during this project, so we’ve covered it in dark grey on the right. Do NOT modify the NEXT bits. Each of the other bits corresponds to a signal asserted onto the datapath during that clock cycle—this is what you will fill out!. We’ve simplified and removed a number of micro-states and control signals which aren’t directly a part of the LC-3’s main instructions. You are given the microcode for some of these micro-states, along with entries for the DRMUX and SR1MUX—do NOT change these. Your task is to fill in the rest and finish the LC-3 micro-controller! 1.3 Assignment Files • latches .sim - a CircuitSim file in which you will build sequential logic components. • LC3 .sim - a large CircuitSim file containing the LC-3 AND a “Manual LC-3” which does not need to be modified in any way but is simply present as a tool for you while writing microcode. You should only modify the PC, CC-logic, and ALU sub-circuits in this file. • microcode .xlsx - an Excel document in which you will write your microcode. Do not touch cells that have been blacked out. • ROM.dat - a text file which you can paste your microcode into and then import into the LC-3. • tests/ - a subdirectory which contains a number of test cases you can use to verify the functionality of your circuit and microcode. • proj2-tester.jar - a local tester that you can use to test all of the components of the project. This official tester is available on Gradescope (where it will be return your grade). • LC-3InstructionsDetail.pdf - a PDF with descriptions and pseudocode for each instruction. 1.4 Elements to complete 1. Implement the RS Latch, Gated D Latch, D Flip-Flop, and Register sub-circuits in latches .sim. 2. Implement the ALU, PC, and CC-Logic sub-circuits in LC3 .sim. 3. Complete the LC-3 microcode in microcode .xlsx. 2 Setup The software you will be using for this project and all future circuit based assignments is called CircuitSim - an interactive circuit simulation package. CircuitSim is a powerful simulation tool designed for educational use. This gives it the advantage of being a little more forgiving than some of the more commercial simulators. However, it still requires some time and effort to be able to use the program efficiently. Please follow the instructions in the file ”CircuitSim Installation Guide.pdf” in Canvas, under ”Files > Course Software > CircuitSim” to install CircuitSim. All .sim files should be opened in CircuitSim. Please do not move, rename, or remove any of the provided inputs/outputs. 3 Implementation 3.1 Latches For this part of the assignment, you will build your own register from the ground up. For more information about each sub-circuit, refer to your textbook. All work for this portion of the assignment will be done in the latches .sim file. 3.1.1 RS Latch In the “RS Latch” sub-circuit in the latches .sim file, you will start by building a RS latch using NAND gates, as described in your textbook. RS Latch is the basic circuit for sequential logic. It stores one bit of information and has 3 important states: 1. S=1, R=1: This is called the Quiescent State. In this state the latch is storing a value, and nothing is trying to change that value. 2. S=0, R=1: By changing momentarily from the Quiescent State to this state, the value of the latch is changed so that the output Q now stores a 1. 3. S=1, R=0: By changing momentarily from the Quiescent State to this state, the value of the latch is changed so that the output Q now stores a 0. Note: In order for the RS Latch to work properly (for testing purposes, not for building), both R and S cannot be 0 at the same time (illegal state). Once you set the bit you wish to store, change back to the Quiescent State to keep that value stored. Notice that the circuit has two output pins; one is the bit the latch is currently storing, and the other is the opposite of that bit. 3.1.2 Gated D Latch In the “Gated D Latch” sub-circuit in the latches .sim file, using your RS latch sub-circuit (in the Circuits tab), implement a Gated D Latch as described in the textbook. You are not allowed to use the built-in SR Latch in CircuitSim to build this circuit. The Gated D Latch is made up of an RS Latch as well as two additional gates that control when a value can be stored. The value of the output can only be changed when Write Enable is set to 1. Notice that the Gated D Latch sub-circuit only has one output pin, so you should disregard the inverse output Q’ of your RS Latch. 3.1.3 D Flip-Flop In the “D Flip-Flop” sub-circuit in the latches .sim file, using the Gated D Latch circuit you built (in the Circuits tab), create a D Flip-Flop. A D Flip-Flop is composed of two Gated D Latches back to back, and it implements edge triggered logic. Your D Flip-Flop output should be able to change on the rising edge, which means that the state of the D Flip-Flop should only be able to change at the exact instant the clock goes from 0 to 1. 3.1.4 Register In the “Register” sub-circuit in the latches .sim file, using the D Flip-Flop you just created (in the Circuits tab), build a 4-bit Register. Your register should also use edge-triggered logic. The value of the register should change on the rising edge. 3.2 Completing the LC-3 Sub-Circuits All work for this section of the assignment will be done in the LC3 .sim file. Note: DO NOT move the position of the inputs and outputs in the LC-3 sub-circuits; this could break the rest of the LC-3 simulator. 3.2.1 ALU You will need to build the ALU (Arithmetic Logic Unit) sub-circuit in LC3 .sim. Note: You may use the built-in adder under the Arithmetic tab! The ALU will perform one of 4 functions and Output it depending on the ALUK signal: 1. ALUK = 0b00: A + B 2. ALUK = 0b01: A & B 3. ALUK = 0b10: NOT A 4. ALUK = 0b11: PASS A You should be able to use what you learned from Project 1 to populate this sub-circuit easily. 3.2.2 PC You will need to build the PC (Program Counter) sub-circuit in LC3 .sim. Note: You may use the built-in multiplexer under the Plexer tab! The PC is a 16 bit register that holds the address of the next instruction to be executed. During the FETCH stage, the contents of the PC are loaded into the memory address register (MAR), and the PC is updated with the address of the next instruction. There are three scenarios for updating the PC: 1. The contents of the PC are incremented by 1. Selected when PCMUX = 0b00. 2. The result of the ADDR (an address-adding unit) is the address of the next instruction. The output from the ADDR should be stored in the PC. This occurs if we use the branching instruction (BR). Selected when PCMUX = 0b01. 3. The value on the BUS is the address of the next instruction. The value on the BUS should be stored into the PC. For example, this is used during the JMP instruction. Selected when PCMUX = 0b10. The PC should only be loaded on a rising clock edge and when the LD.PC signal is on. Ensure that you don’t reach the unused case (PCMUX = 0b11) of the circuit, or else spooky stuff might happen (undefined behavior in LC-3). 3.2.3 CC-Logic You will implement the CC-Logic sub-circuit in LC3 .sim. The LC-3 has three condition codes: N (negative), Z (zero), and P (positive). These codes are saved on the next rising edge after the LC-3 executes Operate or Load instructions that include loading a result into a general purpose register, such as ADD and LDR. For example, if ADD R2, R0, R1 results in a positive number, then NZP = 001. The CC sub-circuit should set N to 1 if the input is negative, set Z to 1 if the input is zero, and set P to 1 if the input is positive. Only 1 bit in NZP should be set at any given time. Zero is not considered a positive number. Bit 2 (the MSB) is N, Bit 1 is Z, Bit 0 is P. With that in mind, set the correct bit and implement this circuit in the CC-Logic sub-circuit. Hint: you can use a comparator for this sub-circuit! They are found in the Arithmetic tab in CircuitSim. 3.3 Using Manual LC-3 (for testing only) • Once you have your PC, CC-Logic, and ALU complete, you can begin playing around with the in- structions and control signals to understand how the LC-3 works. • In lecture and in lab, we have covered the control signals in the datapath and how to use them when tracing an instruction. • The first thing you will want to do is use the Custom-Bus, GateBUS, and LD.IR signals to set a custom IR value on the next rising edge. Until you figure out the micro-states for FETCH, you can use these steps to set your IR with any instruction you want to work on. • Once you have an idea of how FETCH works, you can load some instruction(s) in the RAM. In order to do that: 1. right-click the RAM near the bottom of the Manual LC-3 circuit 2. select “edit contents” 3. click “Load from file” 4. locate and select one of the provided test files in the project (ex: add.dat) 5. close the edit contents menu • Now that you have loaded the RAM with a program, you can fetch instructions into the IR. • Now that you have an instruction in the IR, you can start executing it. In order to do that you can turn on the different signal pins on the right in order to control the datapath and move data around like we did in lecture/lab. Once you think you know how an instruction is executed, you can enter it into the microcode spreadsheet, the process for which is outlined below. – Tests inside the tests/ directory have a comment at the end of the .asm file which explains the system state after the end of the program’s execution. You must ensure that this is the system state after you have run every instruction sequentially through the simulator. – To test whether the program acted correctly, go to the ‘LC-3’ circuit and double-click into the ‘REG FILE’ element that is placed in the datapath. Note: you should not just click into the ‘REG FILE’ subcircuit, as this will not properly load the state of the specific ‘REG FILE’ element that’s built into the LC-3, just some generic REG FILE. – Bonus tip: Use Ctrl-R to reset the simulator state and easily clear RAM and registers to 0 in order to test again. • If you are familiar with LC-3 assembly (you will learn it throughout the next few weeks), you are welcome to write your own test programs to verify your code. Make sure that your programs do not start at x3000, as in this project only we will start execution at x0000 for simplicity’s sake. You can compile LC-3 assembly projects to machine code by following the LC-3 ISA, and then create your own RAM.dat files. However, we can’t guarantee that we’ll be able to help with these test cases in office hours. 3.4 Writing the Microcode ONLY open this Excel Sheet in Microsoft Excel. Do NOT use other applications. If need help accessing Excel: https://oit.gatech.edu/email Note: We recommend you read this entire section before starting to fill out the microcode sheet. • DO NOT CHANGE: – NEXT bits (the 6 greyed out bits on the right side) – Values for DRMUX and SR1MUX (aleady filled in) – Any microstates that are already filled out for you As a general rule of thumb, try not to break the spreadsheet :). • Throughout this assignment, you will find the LC-3 Datapath Reference.pdf to be very helpful (Canvas → Modules → Shared Resources → LC-3 Datapath Reference.pdf). • For information on filling out LDSR and STII, see below. • For more information on each of the control signals, see the Appendix at the end of this PDF. • For the Store instructions, ensure that you prioritize loading the address we want to write to into the MAR over loading the value we want to write with into the MDR. The autograder is expecting that you load the address first and then the value. • Now that you’ve developed some familiarity with the datapath, gotten a chance to “act as the micro- controller”, and run the execute stage of instructions on the Manual LC-3, you can complete the final part of the project: writing the microcode for a number of micro-instructions on the LC-3! • In the microcode .xlsx Excel spreadsheet, there exists a number of macro-states. Among them are FETCH, DECODE, and the EXECUTE stages for most instructions supported by the LC-3 (we’ve removed several macro-states related to trap and interrupt handling). For each macro-state, we’ve provided space for the micro-states which will make up that macro-state. For each micro-state, we’ve handled all of the logic related to transitioning to the next micro-state. We’ve also implemented a small subset of the macro-states in order to provide inspiration to you, our students (you’re welcome). • You should complete all the remaining macro-states by filling in their micro-states. – We recommend that you fill out the micro-states for FETCH first; however, the autograder does not depend on FETCH being implemented first. – If you notice that the output column to the right of the blacked-out columns (column AH) is showing artifacts like #NAME?, try opening the Excel spreadsheet in your Georgia Tech Office 365 Online Excel workspace. Sign in to Office 365 with Georgia Tech credentials, select ‘Excel’, and then choose ‘Upload and open’ to edit the excel sheet online. • STIRSUM – Your fellow TAs are trying to create a new LC-3 instruction! We’re calling it - STIRSUM - Store Indirect Register Sum. – This instruction accesses memory twice–the first loading the value at address SR1 + SR2 and second storing SR1 at the address found from the original load. – Mem[Mem[SR1 + SR2]] = SR1 – The 16-bit instruction is formatted in the following: [ OPCODE | SR1 (3 bits) | 000000 (6 bits) | SR2 (3 bits) ] – Note: We do not care about the value that is left in SR when this instruction is completed; however, the values in any of the other registers must remain the same. – Note: You do NOT need to set the condition codes (CC) for this instruction. • PCHOP – The TAs got together again (yikes) and decided to make another function - PCHOP - Program Counter Hop. – This instruction sets the program counter (PC) to the value in SR multiplied by 8. – PC = SR * 8 – The 16-bit instruction is formatted in the following: [ OPCODE | SR (3 bits) | SR (3 bits) | 000 (3 bits) | SR (3 bits) ] – Note: You do NOT need to set the condition codes (CC) for this instruction. We could really use your help! Fill out the micro-states for STIRSUM and PCHOP in the microcode .xlsx excel sheet. 3.4.1 How to Test your Microcode At any time that you want to test your microcode, you can export it from the .xlsx file and apply it to LC-3 hardware by following these steps. IMPORTANT NOTE: Passing all of the tests provided does not guarantee that you have a functional datapath, any number of coincidences could cause you to get the correct output with incorrect functionality. As always, we reserve the right to grade with additional test cases. To manually test your microcode in the LC-3: 1. At the bottom of the microcode .xlsx file select the output tab. 2. Copy all of column D from row 1 through row 64. 3. In CircuitSim, open LC3.sim. Navigate to the ‘Fsm’ sub-circuit. This circuit contains the micro- controller. To enter the sub-circuit, either double-click on it inside the LC-3 sub-circuit or right-click it in the LC-3 and select “View internal state.” 4. Right-click on the ROM in the ‘Fsm’ and select “Edit contents.” 5. Select the top left cell and paste into the ROM. You should see the values in the ROM change. 6. Navigate to the ‘LC-3’ sub-circuit. 7. You can now load a program into the RAM, following the instructions above in the Manual LC-3 section (Section 3.3 of this PDF). 8. To run the LC-3, you can manually click through the CLK signal or use ‘Ctrl-K’ to start or stop the automatic clock. After your program has stopped executing (you can tell when it’s finished running because it will HALT and the datapath will stop changing). 9. Tests inside the tests/ directory have a comment at the end of the .asm file which explains the system state after the end of the program’s execution. To test whether the program acted correctly, go to the ‘LC-3’ circuit and double-click into the ‘REG FILE’ element that is placed in the datapath. Note: you should not just click into the ‘REG FILE’ sub-circuit, as this will not properly load the state of the specific ‘REG FILE’ element that’s built into the LC-3, just some generic REG FILE. 3.4.2 Tips, Tricks, Resources This is all pretty crazy to take in at first. But it’s not the end of the world if you make sure to use the resources available to you. Here are some options: • LC-3 Datapath Reference Sheet: Canvas → Modules → Shared Resources → LC-3 Datapath Refer- ence.pdf • LC-3 Instructions Detail Sheet: LC-3InstructionsDetail.pdf, in this project.zip • The Manual LC-3! • Your friendly TAs (Office Hours, Piazza, etc.) • Textbook • Appendix on Datapath Control Signals in this PDF. 4 Autograder To run the autograder, run java -jar proj2-tester.jar at a command prompt in the same directory as all your deliverable files (see next section for a list). This is the same autograder that Gradescope uses, but is much easier and faster to use. The autograder should be run from the terminal. Additionally, the autograder will grade each instruction individually, so issues will not cascade. 5 Deliverables Please submit the follow files: 1. latches .sim 2. LC3 .sim 3. microcode .xlsx to Gradescope under the assignment “Project 2 - LC-3 Datapath” . Note: The autograder may not reflect your final grade on this assignment. We reserve the right to update the autograder as we see fit when grading. 6 Demos This project will be demoed. Demos are designed to make sure that you understand the content of the project and related topics. They may include technical and/or conceptual questions. • Sign up for a demo time slot via Canvas before the beginning of the first demo slot. This is the only way you can ensure you will have a slot. • If you cannot attend any of the predetermined demo timeslots, email one of the Head TAs before the beginning of the first demo slot. • If you know you are going to miss your demo, you can cancel your slot on Canvas with no penalty; however, you are not guaranteed another time slot. You cannot cancel your demo within 24 hours or else it will be counted as a missed demo. • Your overall project score will be ((submission_score * 0.7) + (demo_score * 0.3)), meaning if you received a 100% on your submission, but a 50% on the demo you would receive an overall score of 85%. If you miss your demo you will not receive any of these points and the maximum you can receive on the project is 70%.
ELEN 4810 Final Exam 1. Z-transform. A discrete-time LTI system has transfer function Please answer the following questions: Part 1. Please plot the pole-zero diagram of H(z), labeling all poles and zeros. You can use the axes on the next page. Part 2. Which of the following best describes this system? HIGH PASS LOW PASS ALL PASS BAND PASS Please justify your answer. Part 3. Could the system be stable and causal? Why or why not? Part 4. Could the impulse response h[n] be real valued? Why or why not? Part 5. Please determine the transfer function H′ (z) of a system which is causal and stable, and ensures for the following system which applies H and H′ in series, the output y[n] satisfies |Y (ejω )| = |X(ejω )| for every input x[n] : 2. Generalized Linear Phase Systems. Consider an FIR generalized linear phase system, with real valued impulse response h[n], transfer function H(z) and zeros ζ1,...,ζL−1 . Please answer the following questions: Part (i). What are the poles of H(z)? For any repeated poles, please indicate their multiplicity. If it is not possible to determine the poles from the given information, please explain why. Part (ii). Set h′ [n] = (−1)n+1h[n]. Does the resulting system have generalized linear phase? Why or why not? Part (iii). Please give an expression for the zeros ζ1(′),...,ζL(′)−1 of H′ (z) in terms of ζ1,...,ζL−1 . Part (iv). Suppose we wish for h′ [n] to be a low pass system. What types of canonical generalized linear phase system/systems should we not choose for h[n]? Why? 3. Spectrograms. A continuous-time chirp signal xc (t) = cos(αt2 ) is sampled with a sampling period Ts = 0.03seconds to produce a discrete time signal x[n]. We compute the Short-Time Fourier Transform (STFT), X[r, k], using a time stride of R = 10 samples, N = 512 frequency samples, and a window w[n] of length L. We plot the magnitude of the Short-Time Fourier Transform (STFT) |X[r, k]|, for r = 0, . . . , 2400 and k = 0, . . . , 255: Above, the graph at right plots the vertical slice X[1000, 0], X[1000, 1],..., X[1000, 255]. (Note, that N = 512; here we only show X|[r, k]| for k < N/2). Please answer the following questions: Part (a). Please estimate the chirp parameter α . Justify your answer! Part (b). Why does the spectrogram exhibit a rising line and a falling line? Part (c). Which of the following windows was used to determine the spectrogram? RECTANGULAR HAMMING Please explain your answer! Part (d). Please estimate the length L of the window, based on the available information. Note: your estimate does not need to be perfect, but please explain how you arrived at it. 4. IIR Filter Design and Bilinear Transform. We generate a discrete time IIR filter by applying bilinear transformation to a continuous time system, with transfer function Hc (s). Here is the magnitude response |H(ejω )| : Please answer the following questions: Part A. Which of the following best describes this filter? BUTTERWORTH CHEBYSCHEV I CHEBYSCHEV II ELLIPTIC Part B. Our continuous time filter Hc (s) has poles at and a zero of multiplicity three at s = ∞ . What are the poles and zeros of the discrete time system H(z)? Part C. What is one advantage of FIR filters (e.g., designed by windowing or optimization) com- pared to the IIR design in part A? Part D. Suppose we generate a new system, by setting H′ (z) = H(z)H* (1/z* ). What is the phase response ∠H′ (ejω ) of the new system? Part E. The system H′ has order six (six poles and six zeros). Based on your answer to Part D, please describe one advantage to the system H′ , compared directly designing an order six system using bilinear transformation. Part F. Can the system H′ (z) be both stable and causal? Why or why not? Part G. Please give an expression for the impulse response h′ [n] in terms of the impulse response h[n].
ESE 419/572: Analog Integrated Circuits Semester: Fall 2024 Topics: Analog circuits serve as the interface between the physical and digital worlds. Analog circuits amplify and digitize sensor signals such as those from a microphone, accelerometer, chemical sensor or even the small RF signals received on an antenna. Analog circuits also turn digital signals into analog outputs capable of driving physical devices such as antennas, speakers, and robotic motors. Analog circuits often define the limits of system level performance, determining the smallest signals that can be sensed, the largest signals that can be produced, and the overall efficiency of sensors, radios, and actuators. In this course you will learn to design and analyze analog integrated circuits using both hand calculations and circuit simulations. We will focus on the design of CMOS analog circuits supplemented with material pertaining to the design of analog circuits using FinFET and Bipolar junction transistors. You will learn how to design and analyze single and multi-stage amplifier circuits including the gain, frequency response, stability, noise, and linearity. The course will conclude with a project to design an amplifier to a set of provided specifications. Background Knowledge and Prerequisites: It is assumed students have basic knowledge of field effect and bipolar junction transistors and in analyzing simple passive and active electronic circuits using Kirkoff’s voltage and current laws. Primary Text: Design of Analog CMOS Integrated Circuits, 2nd edition, Behzad Razavi, McGraw-Hill, 2017 Supplemental Text: Analog Integrated Circuit Design, 2nd edition, Carusone, Johns, and Martin, John Wiley and Sons Inc, 2012 Lectures: Live lectures will be held every Monday and Wednesday from 1:45-3:15 Eastern Time in Fagin Hall Rm 114. Lecture slides will be posted on Canvas prior to each lecture. The lectures from Fall of 2020 will be posted on Canvas as reference and study material only. The online lectures are no substitute for the in person lectures and should be used only as reference material. In person lecture attendance is highly encouraged. Some materials have been updated since 2020. Class Rules: 1. Class participation is encouraged. If you have a question, please ask it, others likely have the same question. 2. Do not disrupt the class. Silence all cell phones. Do not do other activities (e.g. eating, talking) that may disrupt the ability of other students to participate in the lectures. Grading: Your grade will depend on homework (15%), quizzes (10%),a midterm exam (20%),afinal exam (35%), and a final project (20%). Details are given below. Homework: Throughout the semester a series of homework assignments will be periodically assigned and will be due at the beginning of class approximately 2 weeks after assignment. There will be 5 homework assignments throughout the semester. Working on homework in groups is permitted but each student should turn in an individual homework assignment. Since homework is preparation for the exams please be sure you understand all of the homework solutions you are submitting. Also, be sure to show ALL of your work to receive partial credit. Submission of homework should be in pdf format through Canvas. Should you need to submit an assignment late due to illness or other emergency please let me know as soon as possible so we can make alternative arrangements. Schedule for Homework: Assigned Due Homework #1 9/9 9/23 Homework #2 9/23 10/7 Homework #3 10/16 10/30 Homework #4 10/30 11/13 Homework #5 11/13 11/27 Exams: The midterm is planned to be administered during class on 10/14 and you will have the full class period to complete it. The final exam will beheld during the time allocated by the University. For all exams you may bring a single 8.5x11 sheet of paper containing any notes, equations, problem solutions, etc. that you believe would be useful in completing the exam. Use of electronic devices/media, except for calculators, is prohibited during the exams. Communications with anyone other than the instructor is strictly forbidden during the exams. All exams must be turned in at the completion of the session and may not be photographed or otherwise recorded for further dissemination. Please note that academic honesty is of paramount importance to insure that the class and grading is fair to all students. Academic dishonesty will not be tolerated. You are expected to abide by the PENN Honor Code for all of your conduct in this course, especially the midterm and final exams. If you have any questions regarding what is and is not allowed, please contact me prior to the exams. Should illness or emergency impact your ability to take the midterm or final as scheduled, please let me know prior to the exam if possible. We will need to work together to schedule a makeup time for you to take the exam. I will review with the class the midterm schedule at the beginning of the semester. Should the initial midterm date of 10/14 fallon the same day as other midterms taken by students enrolled in the course, we will work together to try and move the midterm to a date that can be accommodate everyone. Schedule for Exams: Date Midterm 10/14 Final (Per University) TBD Quizzes: Throughout the semester there will be a series of short (10-15 minutes) quizzes administered during the class period. You will receive 1 weeks’ notice before the administration of any quiz. The tentative Quiz dates are given below Schedule Date Quiz #1 10/2 Quiz #2 11/6 Quiz #3 11/25 Project: This course will have a final project where you will design an amplifier to a set of specifications and write a report describing the design decisions leading to your amplifier design and architecture, and the amplifier simulated performance. The design will be performed in Cadence. The project will be an individual assignment. Parts i and ii will be due by the end of the day on 11/27 and the entire project will be due by the end of the day on Dec. 9th. There will be no extensions for the project deadline.
CSE 101 Introduction to Data Structures and Algorithms Winter 2025 Description: Introduction to abstract data types and basics of algorithms. Linked lists, stacks, queues, hash tables, trees, heaps, and graphs will be covered. Students will also be taught how to derive big-Oh analysis of simple algorithms. All assignments will be in C/C++. Prerequisites: CSE 12 or BME 160; CSE 13E or ECE 13 or CSE 13S; and CSE 16; and CSE 30; and MATH 11B or MATH 19B or MATH 20B or AM 11B. Required Text: Introduction to Algorithms (4th edition) by Cormen, Leiserson, Rivest and Stein. MIT Press 2009 ISBN 978-0-26-204630-5. On reserve (2-hour loan) at theScience Library. Recommended Texts: Open Data Structures (pseudo-code edition) by Pat Morin.https://opendatastructures.org/ Data Abstraction & Problem Solving with C++ (6th edition) by Carrano & Henry. Pearson 2013 ISBN 978-0-13-292372-9. Coursework: 50% Programming Assignments: Eight projects due at roughly 8 day intervals 15% Midterm Exam 1: Friday, January 31, 2:40pm - 3:45pm 15% Midterm Exam 2: Friday, February 28, 2:40pm - 3:45pm 20% Final Exam: Tuesday, March 18, 12:00-2:00pm All scores are rounded to the nearest 10th of a percent. They will not be rounded further. No scores are curved. The following letter grade boundaries will be used to determine your grade in the class. Grading scale: A+ 99.0% - 100% A 94.0% - 98.9% A- 91.0% - 93.9% B+ 89.0% - 90.9% B 84.0% - 88.9% B- 81.0% - 83.9% C+ 79.0% - 80.9% C 70.0% - 78.9% C- 68.0% - 69.9% D+ 65.0% - 67.9% D 61.0% - 64.9% D- 59.0% - 60.9% F 0% - 58.9%
