# 4th Year EE Module Feedback

# Autumn Term

Check the module page for details of examination method (coursework / open book / closed book examinations). Please note that examinations for these modules will take place in the summer term and NOT at the end of autumn like 3rd year modules.

# Advanced Communication Theory

Imperial Module Page: here (opens new window)

# 2020-21

Manikas is a great lecturer although the content itself may not be that interesting for those who won't go into COMMS that much. High marks almost guaranteed and a very organized module.

# Coding Theory

Imperial Module Page: here (opens new window)

# 2022-23

# Quick Summary

Average module scores from 3 respondants.

  • Content: 4.33 out of 5
  • Organisation: 3.67 out of 5
  • Lecturer: 3.33 out of 5
  • Overall: 4.0 out of 5
# Comments

I'd say that overall the module was quite interesting. Especially if you have a more theoretical mathematically inclined background, or have interests in that area. It honestly made a nice break from some of the application which is often used in other modules, and you got to pick up on a lot of theory to make it abstract and interesting, refreshing some of that FP3 Group Theory knowledge, while also being able to see it in application and gain an appreciation for it. The notes for the course were really well written and gave a clear insight into lecture by lecture content. The actual schedule for it was super chill also, just one Monday evening lecture that gave you a relaxing insight into everything under the hood.

This module is all coursework assessed, which can be a good May lifeline, if you already have enough exams going on during that time, so I'd recommend picking it up to get yourself a good grade by December. That being said, the courseworks are not exactly a cakewalk as time goes by. There are usually 3 question-based courseworks to do, the first couple will not take long, but the last one usually takes the most time, and will challenge both your theoretical learnings as well as having to code up some of the grunt work and algorithmic work in Julia (the programming langugage of choice in this module, sort of like a BTEC Python but with beefed up cryptography capabilities), honestly Julia was a bit weird at first but easy to pick up - make sure that you're in a good group and divvy up the work if possible as the courseworks usually have something that plays to everyone's strengths, whether you're more theory or code inclined.

The last coursework is even more interesting, as you get to research a relevant topic supplementing with learnings from the course and present to a small panel about it, answering a couple of questions. This is a great way to show interest, and just a chill and nice way to assess the course.

Overall, I'd say the module was very chill, striking the right balance between challenging and advanceable, while maintaining interest. If you're just trying to fill in an EE slot as an EIE, take this one. (There is considerable carryover into the DoC Spring Cryptography course if you want).

I heard some negative reviews about this module from previous years, and almost didn't choose it as a result, but this module ended up being one of my favourite EE modules I took this year. Wei (the lecturer) seems to be really keen on improving the module based on student feedback, with the module itself has undergone a number of changes in the last few years, and I think that its current state is fantastic. The module essentially covers the mathematics behind encryption and encoding schemes (number theory) in quite a high level of detail from the ground up, meaning no prior knowledge is needed, touching on topics which are in use today, although isn't as application focused as DoC's Cryptography Engineering. Assessment takes place via 4 courseworks, where you can choose your own teams of up to 4 members (I personally did them in a pair and found them all very doable in a reasonable amount of time). The first 3 CW all follow the same pattern, with a 50-50 blend of pure math and programming questions that you fill out in a Jupyter Notebook, using a (probably unfamiliar) programming language called Julia, which I'd describe as a modern version of Matlab (as it uses 1-based indexing 😭) and works really well with Matrices. The lecturer provides a "Worksheet 0" to help you learn the language, and I found it relatively easy to get used to it, the most challenging part is learning how to use the Nemo library needed for CW 2 and 3. Initial versions of the worksheets are typically published 2 weeks prior to the deadline, with the final version published 1 week before, and as long as you put in the work it's very doable to get close to full marks (although they might make them more challenging next year as they were a bit too easy this year which caused the module to get moderated down). CW4 has a different format, where you essentially choose a topic related to the course which hasn't been covered in the lectures and present it to a subset of the class, incorporating 50% peer and 50% lecturer marking feedback, and helps balance out the super high average of the first 3 CW. My session had a large variety of topics presented, and I found it to be an interesting way of learning about different topics, often times using the existing content we learned. The lectures themselves are decent, although Wei has a tendency to get a bit lost in some of the more challenging topics towards the second half of the course, but fortunately the class notes provided are fantastic, and given that this is a CW-only module you can do all of your learning entirely based off the notes if you find that the lectures aren't working for you. This was also Wei's first year lecturing using the class notes, so lectures may even be better in future years. Overall, a really solid, high-scoring, low-stress module with some fun mathematics and programming tasks, would highly recommend to anyone looking for a good standalone EE module, or who's interested in communication / encryption. - Simon

# 2021-22

# Quick Summary

Average module scores from 1 respondants.

  • Content: 4.0 out of 5
  • Organisation: 1.0 out of 5
  • Lecturer: 3.0 out of 5
  • Overall: 1.0 out of 5
# Comments

Don't do it, too much work and little reward, marking was late and disorganised

# 2020-21

Lecturer is not always clear, but high marks were given.

The last CW is hard and takes a lot of your time during the busiest time of the term which is the end.

He says it isn't a "coding" module but it is. Make sure you get a group that is good at coding because it saves a lot of time. The content is interesting and he is decent as a lecturer. - KG

# Digital Image Processing

Imperial Module Page: here (opens new window)

# 2022-23

# Quick Summary

Average module scores from 3 respondants.

  • Content: 3.33 out of 5
  • Organisation: 4.0 out of 5
  • Lecturer: 3.67 out of 5
  • Overall: 3.33 out of 5
# Comments

The module was mid. Hella mid, exceedingly mid, nothing but mid really. The lecturer was mid, the timeline was mid, everything about it, mid. The lectures were usually on time, but I found it dragged on a lot, the lecturer herself spoke quite monotonously so it made the lecture content more boring than it actually was tbh, because it did seem quite cool. This is one of those modules you take if you just want to have a plainass standardised exam, with little challenge, which is in itself a benefit, depending on your purview. The content is split up into different sections depending on restoration of images, compression of images, enhancement and different fundemental filter types. The course materials does include a lot of the past papers, as well as past paper questions based section by section and their full answers, which is great to practice with. Other years have stated before that the exam is very standard, and if you can do these you can do decently in the exam. Overall this module is something you probably don't want to do as a 'passion pursuit', unless you are angling towards DS/AI/ML, otherwise if you're like me you'd probably just want to take this as a filler at best.

I didn’t find many modules I was interested in during autumn term so I honestly only took this module because past exams seemed to be consistent with not much variation in the parts of the course you are examined about. The content ended up being a lot more interesting than I expected especially the sections about transforms and image compression. Tania was a pretty solid lecturer and was very helpful and eager to clarify any doubts one might have. Pretty fair module overall.

# Older

2019-20

A really good module. You may doubt Tania after 2nd year linear systems, but when teaching a subject she's actually interested in she's very engaging and a good lecturer. I wish there was a practical component to this module to put into practice the techniques talked about. The exam is very doable, and has remained fairly consistent over the last few years (i.e. past papers are your friend).

# Probability and Stochastic Processes

Imperial Module Page: here (opens new window)

# 2020-21

Very similar to 2nd year module, however exam is much harder in my opinion. If you confident in maths it is a good module to take, but otherwise would avoid it if you looking for amazing grades

# Stability and Control of Non-linear Systems

Imperial Module Page: here (opens new window)

No feedback available

# Systems Identification and Learning

Imperial Module Page: here (opens new window)

No feedback available

# Optimisation

Imperial Module Page: here (opens new window)

# 2022-23

# Quick Summary

Average module scores from 6 respondants.

  • Content: 3.5 out of 5
  • Organisation: 4.33 out of 5
  • Lecturer: 4.33 out of 5
  • Overall: 3.83 out of 5
# Comments

I'd say Optimisation is a fun module to take in the same vein as DIP, a filler module. I took this because I wanted to maximise DoC choices in the Spring, and because there was not a great selection of DoC modules for EIE in Autumn. The lecturer himself is very engaging, interesting and dare I say funny, even though this was my only lecture, on a Friday, at 9AM, in many senses the engagement and passion with which he brought to what would have felt like a mundane subject made the content and the tuberides worth it. I'd say that the content taught in the notes is very rich with information, and actually almost to a fault. This is because you are only really assessed on a plurality of the notes of which you are reading, and the rest is spent deepdiving into various proofs, which can help you conceptually. The exam itself is quite standard if you check out any of the previous past papers, there are a couple curveballs here and there but presuming that you have an understanding of what is being taught, there should be no issues, the actual content itself is useful in a multitude of industries, but I'd say especially if you're angling towards ML. I would recommend this module to anyone who is just looking to supplement with an EE module.

Astolfi is a pretty good lecturer. Exam not too difficult, about average. Would recommend if you have a busy Autumn term as you can basically study all the content during Easter.

The exam was surprisingly difficult and very different from all past papers. Hard to prepare for.

The fact that all lectures were not recorded properly at the beginning of the term meant that a lot of switching between last year's (2 years ago actually) and this year's recordings was required.

Disclaimer: I took this module in combination with DoC Computational Optimisation which has substantial overlap in terms of content, so my opinion on the difficulty of the content might be slightly skewed. This module is a really enjoyable "maths-only" module with a fantastic lecturer, covering various methods of how to optimise functions (typically minimisation problems), which is applicable to a huge number of fields (notably ML), going over the underlying theory. If you took DoC's Operations Research last year, you'll already be familiar with the formulation of the kinds of problems being solved, but it's by no means a pre-requisite for taking this module. A strong understanding of linear algebra is necessary to do well (e.g. taking the derivative of vectors / matrices), but this is something you can learn during the course (doing DoC's Computational Optimisation is a big help in this). The workload is quite small during term time, with only a single 2-hour lecture slot per week, but you'll probs have to spend a fair amount of time revising in May (even if you took Comp Opt), so be careful when choosing this module in combination with other exam-based modules. I found that it worked really well to help frontload my modules in Autumn, whilst spreading out the assessment for later on as my only EE exam module. The lectures themselves are really engaging, and Astolfi is an amazing lecturer, who's also quite responsive on EdStem, which is very helpful leading into exams, although his responses sometimes lack depth to properly answer the question. He also provides accompanying notes, although these are significantly overcomplicated and dense in my opinion, and I didn't find them particularly helpful - I think the lectures are all you really need. They also contain a large number of practice exercises, usually taken from old exam papers, but I think focusing on the more recent past papers is probably better in terms of practice. The exams themselves are relatively similar from year to year, so doing 5+ years of past papers puts you in a strong position to do well in the exam (I personally did 9). They're also open book, and you're allowed to take as much as you want with you into the exam, meaning there's no need to waste time memorising formulas (although in depth notes aren't really needed). One issue I had with older papers is the (sometimes outrageous) level of number crunching, but this is something that the lecturer is aware of and has rectified in recent years, and wasn't a problem in this year's exam (provided you could read the question correctly which I was unfortunately unable to do 😑). In terms of differences with Computational Optimisation, convexity isn't really covered much in this module, and its importance is a little understated imo, and different penalty methods are considered, but Gradient descent, Newton's algorithm, and the KKT conditions are present in both. If you're looking to cover as much content as possible this year, taking both modules is probably a waste, but I think that the pair are highly synergistic and will help you understand the other better, with Autumn Term EE Optimisation lectures -> Spring Term DoC Optimisation -> May/June EE Optimisation exam. I think I would not have done as well in either module had I not taken the other, and would recommend both (unless you hate math). - Simon

Average module, do if you’re interested in the topic

# 2021-22

# Quick Summary

Average module scores from 2 respondants.

  • Content: 4.0 out of 5
  • Organisation: 4.0 out of 5
  • Lecturer: 4.0 out of 5
  • Overall: 4.0 out of 5

# 2020-21

Very good lectures, really interesting and covers the concepts of optimisation really well. The exam is average, neither hard or easy. But it has a good structure and it is unlikely it will have weird or unexpected questions.

You can take this course just for the lecturer himself!

# Wavelets, Representation Learning and their Applications

Imperial Module Page: here (opens new window)

# Modelling and Control of Multi-Body Mechanical Systems

Imperial Module Page: here (opens new window)

No feedback available

# Power System Economics

Imperial Module Page: here (opens new window)

# 2022-23

# Quick Summary

Average module scores from 4 respondants.

  • Content: 3.5 out of 5
  • Organisation: 2.25 out of 5
  • Lecturer: 2.5 out of 5
  • Overall: 3.0 out of 5
# Comments

Not exciting but fairly easy content and easy to revise. Unfortunately insane time pressure in the exam. Content easy just volume a lot. Practice timed papers as much as possible or hou will not finish.

Very middle of the road module, not great not bad. Exam is pretty consistent throughout the years with very minor changes if any. Coursework is pretty simple, but marked really weirdly. Feedback is very vague and haphazard, many of us had similar graphs, obtained the same numbers and provided pretty similar explanations but got very different grades (A-C)? Content is interesting for anyone who doesn’t know much about power systems though!

# Topics in Large Dimensional Data Processing

Imperial Module Page: here (opens new window)

# 2022-23

# Quick Summary

Average module scores from 3 respondants.

  • Content: 3.67 out of 5
  • Organisation: 2.33 out of 5
  • Lecturer: 2.0 out of 5
  • Overall: 2.67 out of 5
# Comments

Coursework is all about solving and doing some very complicated maths that is not really taught well (or at all) and implementing optimization algorithms in Julia. Lectures and notes are quite bad and not very useful, you resort to reading really badly written research papers about the algorithms in hopes of understanding how to implement them. Not much time given in between notebooks so you’re constantly working on this module. Would recommend having a good group and dividing the parts of the notebook amongst yourselves as it’s too much content for everyone to try and cover simultaneously. Weird module overall, didn’t take much out of it.

The organisation of the lectures was awful, but the organisation of the CW was okay.

Very high marks can be achieved if a lot of hours are put into it.

Module topics (!) are interesting, but module delivery (in 2022-23) limits how easy it is to learn. Lectures are frankly hard to follow, since they jump from point to point of a massive pdf of notes without any clear script. Coursework questions are solved from googling and past knowledge more often than with lecture content. Some questions were basically impossible, but marking was generous so it was fair to have those as a bonus challenge.

# 2021-22

The module itself is quite hard and Dai tries to condense and simplify it to a level that we can understand. Other than the first coursework, the rest aren't that difficult in terms of implementation but conceptually they can be a bit more difficult. - KG

# Self-Organising Multi-Agent Systems

Imperial Module Page: here (opens new window)

# 2022-23

# Quick Summary

Average module scores from 2 respondants.

  • Content: 4.5 out of 5
  • Organisation: 2.5 out of 5
  • Lecturer: 4.5 out of 5
  • Overall: 3.5 out of 5
# Comments

The group coursework is "fun" with everyone working on the same project. Would recommend if you want a less technically demanding module - just requires basic programming skills.

# 2021-22

# Quick Summary

Average module scores from 2 respondants.

  • Content: 4.0 out of 5
  • Organisation: 5.0 out of 5
  • Lecturer: 4.5 out of 5
  • Overall: 3.0 out of 5
# Comments

It you don't want a Christmas holiday take this module (unless it's changed) and make sure you choose a good group

So... this was an experience. 47 people all doing the same coursework. I took two days off during the entire Christmas break cos I spent all of my time working on this module. Was it worth it? Probably not. But also I did enjoy it (maybe I'm a masochist) and I learned a lot from it. Gained lots of software engineering experience and learned a lot about project management. Inevitably with a group this size you will have people who do fuck all for it. Students decide everything – programming language, version control, how to split into groups, which groups do what work, etc. You get out what you put in for this module. JZ (EIE)

# 2020-21

Time consuming but a very interesting module.. Do it if you can sacrifice your XMas break.

# Hardware and Software Verification

Imperial Module Page: here (opens new window)

# 2022-23

# Quick Summary

Average module scores from 6 respondants.

  • Content: 4.17 out of 5
  • Organisation: 3.33 out of 5
  • Lecturer: 4.0 out of 5
  • Overall: 4.0 out of 5
# Comments

Software part is quite theoretical, focusing on proof assistants and verification oriented languages rather than testing of regular programs. Very interesting, but important to know what to expect.

Hardware was awful, but is being changed completely. Unlike SW the focus was on testing real hardware, though the content taught in 2022-23 was outdated.

This was honestly my most disappointing module of Y4. The module is essentially made up of two components (hardware and software), which focus on the idea of verification in both contexts. Pete Harrod, a visiting lecturer from Arm, covered the hardware component of the course, and it is honestly difficult to understate how painfully boring his lectures were. However, as someone who has very minimal experience in hardware, the content covered in the course seems very relevant to verification in industry, and is an excellent way to (re-)learn SystemVerilog. Note that the coursework component for this section of the module, in which we needed to apply a range of verification techniques to a simple GPIO, and a VGA block, was incredibly time consuming, and I would estimate requires at least 2 weeks of full-time work (as a pair) to complete to a reasonable level of quality (and most likely take longer). Although it was much too long in my opinion, it does allow you to take quite a lot out of the module, making it very useful (I'd imagine) for hardware-focused people. Note that in future years, Wickerson will be taking over this part of the module as well, so content / coursework may be quite different. The software part of the course was essentially a complete opposite to the hardware part. Wickerson's lectures are (as anyone who's taken Y2 compilers knows) entertaining and easy to watch, but the actual content covered has incredibly slim links to how software "verification" is done in practice (particularly on the second half of this component focused on theorem provers), and I think that a lot of what was taught isn't particularly helpful outside of narrow research areas. The courseworks were made up of 2 worksheets which were quite frankly a complete joke, with each doable in a single day. I don't think these really allowed you to fully appreciate the material taught in this part of the module, as you only had to implement extremely simple functions (with the most complex being a sorting algorithm and a logic optimiser), meaning that you can't really see the value being added by verification in these cases which could be quite easily asserted correct using more conventional unit testing or fuzzified testing approaches. Another issue, particularly for Isabelle (Theorem Prover), is that the language itself is incredibly obscure, making it very difficult to debug if when you get it wrong. This made the entire software part of the course feel like a bit of an afterthought, which obviously isn't ideal if you took this module as someone who's more software inclined (DoC's Software Reliability is probably better if you're looking for something like this). Overall, as someone who doesn't really care about hardware that much, this module felt like a complete waste of a slot, gaining very little from the software side, and needing to spend substantial amounts of the Autumn term working on a large (and stressful) SystemVerilog project. DO NOT take this module just because you really liked Wickerson in Y2 as you will likely be disappointed. - Simon

Software part a lot more entertaining than hardware

# 2021-22

# Quick Summary

Average module scores from 3 respondants.

  • Content: 4.33 out of 5
  • Organisation: 4.0 out of 5
  • Lecturer: 4.0 out of 5
  • Overall: 4.33 out of 5
# Comments

Very good module, it's fairly new, but lecturers keep improving it and I like the state of it (2 years ago it was apparently quite rough). Teaches lots of SystemVerilog which is a must have for any hardware role, but also the software part gives a very different view on writing 'trustworthy' code, which is also starting to get attention in industry

This module is split between two lecturers so it's harder to give at overall rating. Some organisation for the hardware part was lacking and lectures weren't too engaging.

It is what you sign up for: proving that software works (which is not testing software). Dr. Wickerson has the lectures up on his Github and they're quite straightforward to set your expectations. You have to 1- reason about a proof of why your algorithm works and 2- "write" this proof to be interpreted and confirmed by a verification tool. The hardware part can be improved by a lot. While you learn about practical verification methods (e.g. you have to build testbenches), it lacks a structure and could be much better organized. In terms of marks, I think the marking scheme was fairly generous VS the amount of work you have to put in. - Jaafar (EIE)

Partnered with Jaafar for this one – the hardware coursework is really open-ended – you can do ‘enough' and come away with a quite decent mark, but you also have the opportunity to really stretch yourself and get a really good mark. JZ (EIE)

# Spring Term

Check the module page for details of examination method (coursework / open book / closed book examinations). Please note that examinations for these modules will take place in the summer term and NOT at the end of autumn like 3rd year modules.

# Digital Signal Processing and Digital Filters

Imperial Module Page: here (opens new window)

Prerequisite: ELEC60010 - Digital Signal Processing

# Adaptive Signal Processing and Machine Intelligence

Imperial Module Page: here (opens new window)

# 2021-22

# Quick Summary

Average module scores from 1 respondants.

  • Content: 5.0 out of 5
  • Organisation: 1.0 out of 5
  • Lecturer: 2.0 out of 5
  • Overall: 3.0 out of 5
# Comments

Highly demanding coursework that will require a lot of dedicated time. Perhaps disproportionately too much time given the weighting of the module in the overall year.

# Speech Processing

Imperial Module Page: here (opens new window)

Prerequisite: ELEC60010 - Digital Signal Processing

# Digital Control Systems

Imperial Module Page: here (opens new window)

# 2022-23

# Quick Summary

Average module scores from 1 respondants.

  • Content: 5.0 out of 5
  • Organisation: 5.0 out of 5
  • Lecturer: 5.0 out of 5
  • Overall: 5.0 out of 5
# Comments

In the first lecture, the lecturer Giordano showed module feedback from previous years. One of them said that this was the best module the student had ever taken, and honestly this is true. It is taught very well, Giordano clearly cares, and the content is interesting.

If you like control, take this module because most real controllers are implemented digitally.

If you like signals, take this module because it will give you in a couple lectures what you should have learned in control year 2, and you'll find the Z-transform content easy.

If you like software/digital, take this module because you will get the background you need to implement controllers in SW/FPGA.

If you like none of those things, then you're in the wrong department because EEE only cares about FPGAs, signals, and optimisation related stuff like ML and control.

# Design of Linear Multivariable Control Systems

Imperial Module Page: here (opens new window)

No feedback available

# Information Theory

Imperial Module Page: here (opens new window)

# 2021-22

# Quick Summary

Average module scores from 1 respondants.

  • Content: 4.0 out of 5
  • Organisation: 3.0 out of 5
  • Lecturer: 2.0 out of 5
  • Overall: 3.0 out of 5

# 2020-21

Interesting and well-taught module! Although the content is very much math-heavy, exam was decent and the lecturer was very approachable so that you could ask any questions about the content. Would take it again! This course is especially for those who wouldn't mind proving a couple of mathematical theorems and/or those who are at ease with making mathematical arguments

# Sustainable Electrical Systems

Imperial Module Page: here (opens new window)

# Predictive Control

Imperial Module Page: here (opens new window)

Prerequisite: ELEC600008 - Control Engineering

No feedback available

# Discrete-Event Systems

Imperial Module Page: here (opens new window)

# Human-Centered Robotics

Imperial Module Page: here (opens new window)

# 2022-23

# Quick Summary

Average module scores from 1 respondants.

  • Content: 2.0 out of 5
  • Organisation: 2.0 out of 5
  • Lecturer: 4.0 out of 5
  • Overall: 3.0 out of 5
# Comments

What you learn entirely depends on your CW project and what you do within your team. Lectures are essentially useless. The professor and the GTAs are very nice and approachable.

Demands high efforts, but marks are appropriately rewarding.

# 2021-22

# Quick Summary

Average module scores from 2 respondants.

  • Content: 5.0 out of 5
  • Organisation: 5.0 out of 5
  • Lecturer: 5.0 out of 5
  • Overall: 5.0 out of 5
# Comments

Was good, a lot of work but fun, choose team wisely.

Really good module, really open-ended, lectures are really interesting. You have the freedom to build pretty much any sort of robot. Make sure you pick a really good group who you know will put in the work and who you work well with. Try to get a good mix of people with different experience as well. Highly recommend this one – you won't get anything else like this. JZ (EIE) ^(.+)$

# Wireless Communications and Optimisation

Imperial Module Page: here (opens new window)

No feedback available

# Signal Processing and Machine Learning for Finance

Imperial Module Page: here (opens new window)

# 2022-23

# Quick Summary

Average module scores from 1 respondants.

  • Content: 4.0 out of 5
  • Organisation: 3.0 out of 5
  • Lecturer: 3.0 out of 5
  • Overall: 3.0 out of 5
# Comments

I don't really like the style of submitting one big coursework at the end of term, with little feedback except for one initial submission at the start.

With that being said, I think this module and many others definitely benefit from being 100% coursework based, just maybe the structure could be better.

Content was fine I think.

# 2021-22

# Quick Summary

Average module scores from 2 respondants.

  • Content: 4.5 out of 5
  • Organisation: 2.5 out of 5
  • Lecturer: 2.5 out of 5
  • Overall: 3.0 out of 5
# Comments

Applies signal processing and machine learning ideas to finance. Lectures a bit boring. 100% coursework that each year seems to have deadline much after the term end means you don't do much during the term, which reduces the workload. CW itself is in Python, which is quite nice, but you also need to make the (very long) Jupyter Notebook into an actual self-contained report without code, which is really stupid and doesn't add any value.

Coursework is highly demanding and will require a lot of your time to do experiments, generate figures and write the report. This module was slightly better than ASPMI in that the GTAs were more willing to offer help.

Lectures are overall very poor in quality. Some slides even have incorrect theory on them. Just like 3rd year ASP and 4th year ASPMI, you have to submit one report at the end, but this time you have no page limit. GTAs are helpful only if you have questions about the theory. The actual Python code is on the whole relatively easy to complete, the main thing to ensure a good mark is that the final report is written well. The main things the GTAs said were (in order to do well):

  • Presentation, formatting and high-quality figures are very important.
  • The final deliverable should be a standalone pdf report, as if you are writing a scientific paper. No python code should be included and please try to clearly introduce each concept and analyze critically the results you are getting. A mere presentation of the results is not enough. Would recommend doing everything locally, I.e. Jupyter notebooks and LaTeX report on local PC, as it saves a lot of time in compiling the report (Overleaf has been crashing a lot more frequently this year). Letter grades came back a lot quicker this year than for 3rd Year ASP (grades came on 18th May) - AB

# Computer Vision and Pattern Recognition

Imperial Module Page: here (opens new window)

# 2022-23

# Quick Summary

Average module scores from 3 respondants.

  • Content: 4.0 out of 5
  • Organisation: 3.67 out of 5
  • Lecturer: 3.67 out of 5
  • Overall: 4.0 out of 5
# Comments

Pretty average module. For EIE students the content is very different to that covered in the DoC CV course. Pretty light workload, particularly during Ad's part (his lectures are incredibly short). Marking is normally pretty generous, so would recommend if looking for some easy marks.

the module is good

# 2021-22

# Quick Summary

Average module scores from 4 respondants.

  • Content: 3.0 out of 5
  • Organisation: 3.0 out of 5
  • Lecturer: 2.75 out of 5
  • Overall: 3.0 out of 5
# Comments

First half a bit boring, overlaps with 3rd year DoC Graphics as well as Computer Vision modules. Second half more interesting, but still very easy. Both courseworks done in MATLAB which is far from enjoyable - in the first one most time is spent on taking good photos. Both CWs also have stupidly low page count for the report, so significant time is spent resizing images and cropping text so that it is actually in the limit

Not fun, mainly just grinding out report, apparently will run like Deep Learning from now on which is way better than how it was run this year

The lectures from the computer vision part of the course could be improved as they were not explained well

Easiest module I have ever done at Imperial. Mikolaivic seems really uninterested in the lecture so some concepts are poorly explained even though they seem really interesting. The new lecturer, Spiers teaches the second part of the module in pattern recognition and he goes over very simple concepts without much depth. He is very engaging however and if you ask questions will be very easy to learn more. The coursework is really easy, just follow instructions and submit your matlab results. Would recommend this module as a way to waste little time in Spring term.

This was really dull. Wasn't too hard I suppose, but I really didn't enjoy it that much – especially the first half. Not sure if I really learned anything about CV. Also the lectures appear to have a lot of plagiarised content. Lectures didn't have much context with them so I didn't really understand what was being taught. PR half is a lot better – Ad Spiers is really engaging. Both cw are quite doable, first one I'm not sure what the point of it was. Also they introduced a test at the end of the term for this year, a lot of it was Menti questions from the live lectures. Stupid test but whatever, it wasn't too difficult. JZ (EIE)

Agree with what JZ is saying. The first half of the module (CV) is taught by Krystian and its pretty much the same manner that he lectures for 3rd year DL. The CV coursework is pretty bad in my opinion, as it doesn't really link in with the slides at all. GTAs are helpful for this part if you organize 1-1 meetings with them to discuss your questions. For the final Menti test, be sure to go over all the Menti questions he does in lectures. Also, he doesn't really go over his final lecture (and it doesn't come up in CW1) but it comes up in the test, so be sure to go over it once. Spiers teaches the 2nd half (PR) and his slides align a lot more nicely with the coursework. There's no GTAs for this part, but he is generally pretty quick to respond on Ed. His Menti questions were on the whole ok, but he did throw a couple of random unseen questions in that just require common sense. - AB

# Advanced Optimisation

Imperial Module Page: here (opens new window)

No feedback available

# Non EE

# DESE96008 – Design for Additive Manufacturing

# BIOE97120 – Human Neuromechanical Control and Learning

# BIOE97152 - Biomimetics

# BIOE97156 – Animal Locomotion and Bioinspired Robotics