代做CMT304 Programming Paradigms 2024–25帮做Python程序

Assessment Proforma 2024–25

Key Information

Module Code	CMT304
Module Title	Programming Paradigms
Assessment Title	Differentiable Programming
Assessment Number	Part 3 of the 4-part portfolio coursework
Assessment Weighting	25% of the portfolio coursework
Assessment Limits	Hand-out: 6th of February 2025 Hand-in: 10th of April 2025, 9:30am Feedback expected by: 13th of May 2025 Limits are per task as set in the instructions

The Assessment Calendar can be found under ‘Assessment & Feedback’ in the COMSC–ORG– SCHOOL organisation on Learning Central. This is the single point of truth for (a) the hand out date and time, (b) the hand in date and time, and (c) the feedback return date for all assessments.

1    Learning Outcomes

The learning outcomes for this assessment are

•  Explain the conceptual foundations, evaluate and apply various programming paradigms, such as logic, functional, scripting, filter-based programming, pattern matching and quantum com- puting, to solve practical problems.

•  Discuss  and contrast the issues, features, design and concepts of a range of programming paradigms and languages to be able to select a suitable programming paradigm to solve a problem.

2 Submission Instructions

The coversheet can be found under ‘Assessment & Feedback’ in the COMSC–ORG–SCHOOL or- ganisation on Learning Central.

All files should be submitted via Learning Central. The submission page can be found under ‘As- sessment & Feedback’ in the CMT304 module on Learning Central. Your submission should consist of these files:

Description		Type	Name
Coversheet	Compulsory	One PDF ( . pdf) file	coversheet. pdf
Task 1	Compulsory	One Python (. py) source file	task1 . py
Task 2	Compulsory	One PDF ( . pdf) file	task2. pdf

If you are unable to submit your work due to technical difficulties, please submit your work via e- mail to [email protected] and notify the module leader (and ideally the setter, if different).

Any code will be tested on a Linux system equivalent to COMSC’s Linux lab machines and must run there.

3    Assessment Description

The attached measurements . csv file contains measurements obtained from the analog circuit below. It is a csv file where the first column contains the time and the second column the voltage measurement at x in the circuit (1, 000 datapoints from time 0 to 2π ; time in arbitrary units).

This circuit consists of analog computing components we used with LTSpice in the module (ic indicates a non-inverting integrator, + a summer, - an inverter and AB a multiplier).  The data in the csv file are noisy voltage measurements from this circuit.  Information about the values a, textttb, and x0 is not available.

This assignment is about analysing the data with differentiable programming techniques and some questions about the approach as given by the tasks below, aligned with the contents of part three of the module.  It is a simple example to demonstrate your understanding of the programming paradigms involved. There are of course other approaches one can use to analyse the data, but this is not asked for.

The example has been kept simple to avoid the need of high computational resources. This should be executable with reasonable CPU resources without GPU; you can of course use a GPU to solve

the task.

Task 1: Write a python program using differentiable programming techniques to approximate the measurement data in measurements. csv with a parameterised function fp  : R → R, t ↦→ x. You may use pytorch, tensorflow or jax for this (or maybe a combination of these pack- ages; using numpy and matplotlib for supporting functionality is fine – any other packages are not needed; check with the coursework setter if in doubt). You are free to choose any function type (some analytical function, a neural network, etc), but you may want to consider the circuit diagram to choose a suitable function and its parameterisation. Your code can produce the results in any suitable format, on the terminal or in files (do not submit these). Submit a single python file solving this task.  Assume measurements . csv is in the directory the python file is executed from.

Task 2: Justify the function and the parameterisation you have used to match the data and explain what you can learn from this about the circuit given your results from the previous task. Or, if you cannot conclude anything, explain why not. Write a short report about this of up to 400 words (this is an upper limit, not a target).

4 Assessment Criteria

Task 1 worth 25% of the coursework

High Distinction 80% - 100%	Code is valid without syntax errors.  An efficient attempt to approximate the measurements using differentiable programming has been made. It pro- vides and excellent approximation of the measurement data which clearly provides insights into the operation of the circuit.  The attempt has been well documented clearly stating the idea to solve the problem and how it has been implemented and performance optimised.
Distinction 70% - 79%	Code is valid without syntax errors. A suitable attempt to approximate the measurements using differentiable programming has been made. It pro- vides and excellent approximation of the measurement data which helps to understand the operation of the circuit.  The attempt has been well docu- mented clearly stating the idea to solve the problem and how it has been implemented.
Merit 60% - 69%	Code is valid without syntax errors. A suitable attempt to approximate the measurements using differentiable programming has been made. It pro- vides a reasonable approximation of the measurement data, but the results do not help much to understand the circuit. The attempt has been well documented, stating the idea to solve the problem and how it has been im- plemented.
Pass 50% - 59%	Code is valid without syntax errors. A suitable attempt to approximate the measurements using differentiable programming has been made, but it does not work for the measurement data provided. The attempt has been reason- ably documented.
Marginal Fail 40% - 49%	Code is valid without syntax errors, but is not capable of approximating the measurements with differentiable programming. An attempt to document the code has been made.
Fail 0% - 39%	Code does not run or is not suitable to approximate the measurements with differentiable programming. There is little to no relevant documentation.