ABCDEFGHIJKLMNOPQRSTUVWXYZAAABACADAEAFAGAHAIAJAK
1
Please take a look at the end of this sheet (lower left corner for numeric average), please scroll
2
TimestampHow well did we adapt to the COVID-19 situation in the context of this class without compromising the learning goals? What else could we have done to make the Zoom-based class better or closer to the actual classroom exp?Rate the overall learning experience for this class? How interesting or intriguing were the materials presented in the class overall?How challenging was the class overall?How challenging were the labs?How challenging was the exam?Rate the impact that the following aspects of the class had to expand your learning boundaries? 5: Highest impact, 1: lowest impact [Lectures]Rate the impact that the following aspects of the class had to expand your learning boundaries? 5: Highest impact, 1: lowest impact [Labs]Rate the impact that the following aspects of the class had to expand your learning boundaries? 5: Highest impact, 1: lowest impact [Project]Rate the impact that the following aspects of the class had to expand your learning boundaries? 5: Highest impact, 1: lowest impact [Exams]Overall, how effective was my instruction style in making the topic interesting to you?How would you rate the help that you received from TAs and lab instructorsHow satisfied were you with the quality of grading of labs?How satisfied were you with the quality of grading of the exam?How comfortable did you feel in asking questions in the class?Rate the labs. Consider challenges and learning opportunities that each of the labs offered. [Performance Evaluator]Rate the labs. Consider challenges and learning opportunities that each of the labs offered. [Basic Gait Authentication]Rate the labs. Consider challenges and learning opportunities that each of the labs offered. [Keystroke Dynamics Challenge]Rate the labs. Consider challenges and learning opportunities that each of the labs offered. [Swiping-based Authentication]How effective was the use of the GitHub classroom, consider the fact that it helped you build your coding profile which will only grow from here?How effective was the use of the Piazza discussion board?How responsible, respectful, friendly, and accessible I was overall inside as well outside of the class? The topic you understood the most:The topic you understood the least:The topic you enjoyed most:The topic you enjoyed the least:What would you tell a friend who was considering taking this course and asked you about it?What would you tell a friend who was considering taking my course (other than 263) in the future?Anything else that you would like me to know?
3
30/04/2020 11:27:446Maybe almost enforce, or more strongly encourage people to turn their cameras on. I definitely could focus better when I had my camera on, but I also often didn't feel like I should because so many other people had their cameras off. It definitely also makes you feel more pressured to participate when you have your camera on. 867873551634855544686The general process of how a biometric system works, such as the general flow chart of it, the typical problems in one, and how you might make one.Things that were only in our test and not in our labs. Most of them felt like we were learning them just to be tested on and then to be forgotten. Things like huffman coding and Naive Bayes Theorem. I think I forgot these things because I never actually had to use them in a lab or saw why it was important that we know or learn them, other than that I was going to be tested on them. Probably looking at spoofing techniques, and our initial discussions in learning how to create biometric systems. Again, things that weren't in our labs: Fourier series, Naive Bayes, and then also all of the distance metrics felt confusing and gone through very fast. The pair labs and realistic course work are great design choices for this class, and probably will give you real applicable experience in coding for your future in CS. Sometimes lecture can feel disorganized, or not too related to the labs, but working on the labs were really what made the class great and interesting. I would also say that this was the first iteration of the class, and is on a new topic, so of course it is going to be disorganized, and that I think it would likely to be better in future semesters. The labs and projects are great. I think with a designed syllabus and or textbook the lectures could be pretty good and interesting too, but sometimes in feels a bit confusing why you are learning certain topics. Overall the pair-programming and labs are really good experience. Pair-programming was awesome and easily the most meaningful experience I have had working on labs at Haverford. I am also glad you decided to keep the final project and agree that that was also meaningful to our learning. I think with a more definite syllabus and idea of topics outside of the labs, some of the things we learn in class that aren't directly related to our labs could feel more meaningful and useful. But whenever we were learning about things that we were using in our labs, they mostly made sense and felt meaningful. Also I think it could be helpful to have a little more direction for some labs, although I do appreciate the open nature of the labs and how they forced me to think thorugh the problems of the lab myself, and solve them, but I think students could still get this benefit if you eased them into that process. Thank you for a great semester!
4
30/04/2020 11:34:118I think overall Zoom-based might have been someways better. I felt like I was more engaged and was more willing to talk. I think to improve and ultiize our online platform is to add more polls and interactive aspects to our zoom sessions. 871010744346559553449108Classification methodsThe math behind certain distance metrics such as DTWKeystrokesGaitI would say be prepared to put a lot of effort into this class. I feel like overall I initially hated it because I was so imitated and felt I didn't understand anything, but after a while I began to understand more. Especially after reading research papers about biometrics, I feel like I have a good understanding on the concepts. I would say you expect a lot from students, but you will definitely learn. I think you are very understanding when it comes to grading, but I think the workload was a lot at times.
5
30/04/2020 12:00:0410I think you should feel free to cold-call on people--if a few seconds pass and no one has responded to your question, don’t hesitate to ask a specific person what they think! It was kind of awkward sometimes when no one wanted to chime in, but this way everyone would be more alert I think. 9107873524867109524481010The performance evaluation stuff in the beginning of the course--histogram, FAR/FRR/ROC curve, etc. I know it’s small but it is helpful for the entire course and many others!Machine Learning--it wasn’t until the midterm review lecture that I understood what you had been presenting about for a few weeks. Keystroke analysis, once we finally figured out how to clean the data!machine learning/your lectures on it--felt like information overload from slides but no simple examples to get me engaged. I did not like learning via snapshots of other professors’ lectures! But, in studying for Exam 1, I actually learned a ton about ML which I am thankful for!In all, a super cool class! I didn’t realize going into it how much I’d learn about ML methods and statistics and math. Lectures were sometimes a bit confusing/dry and labs were a little disorganized at times (TAs didn’t know full expectations, rushed timeline, etc.) but I’m sure in future iterations of the class these wrinkles will be ironed out.Rajesh cares about learning new things above all else, which I really appreciate. He is a caring prof which means he is flexible when you need! Feedback comes from TAs slightly more-so than in other Haverford classes, but Rajesh is always available for office hours, before/after class, exam feedback, etc. Thank you! I sometimes feel bad that my peers and I were sometimes shy to participate, so I hope you know that I really appreciated this course and learned a lot!
6
30/04/2020 12:45:517Not anything I can think of.368842433233723343888Confusion Matrix, ROC, DET, DTW, Naive Bayes, KNN are some examples.Fourier transformations, and the machine learning algorithms that were not listed above.DTWFourier transformations, and the machine learning algorithms that I did not understand.The professor genuinely cares about your education and is very open to input from the students. However, I found that the course was hard to follow as the Professor tried to cover a plethora of topics which oftentimes required a knowledge base that was outside of the prerequisites for the course. For example, we covered a number of machine learning algorithms in a short span of time, perhaps a week or two. The coverage of these methods was much too short to acquire an adeuquate understanding of the methods. Even in the machine learning class, which has Linear Algebra as a prerequisite, the syllabus has one algorithm covered per week. This pattern of sparse coverage of too many topics continued throughout the semester (for example, entropy and fourier transformations). Furthermore, for most of the semester the expectations of our grading and what we hoped to learn was not clear. Most of the lectures felt as if the course was being planned week to week instead of a series of lectures that built upon one another. In terms of assignments, little guidance was given until late into the semester as to what we were being graded on. On one lab in particular, we had trouble decoding what exactly we were being docked on.The professor is nice and is open to input. But overall, I think the courses are quite hard to follow, with unpredictability in the difficulty and expectations of the courses.Thank you for the semester and for working together with the students to make the difficult adjustments that we had to make.
7
30/04/2020 13:37:058I actually think it may have been better than H204. That room is terrible for lectures because no one can see anyone's face, the board is far away and poorly lit, the heater is loud, etc. I recommend you never teach a class in there again.788882412567771243988Information theory because I'm learning it in another class as well.Window selectionInformation theoryWindow selectionHe expects way too much out of sophomores in a 200 level class, which is good if you want to push yourself. The material is challenging and he moves fast and he also isn't great at explaining things so you may have to teach yourself a lot, but if you're ready to handle that then you will get a lot out of this class.He expects his students to be at a research level in the topic which I think is a good thing, at least it's better than coddling us and moving too slow. He's not great at explaining things.I think H204 shouldn't be used for lectures by you or anyone else, and you should get better at explaining things and measuring our understanding. If you want help I recommend asking Sorelle, she is an incredible lecturer.
Also the exams and labs could have clearer guidelines. On lab 2 and 3 I felt you imposed too much structure on our code and I would have structured my classes differently. The lab guidelines were often unclear, so I just interpreted them as soft guidelines and did what I felt was right. I recommend giving clear expectations for what the final product should do and giving us freedom and advice for how to get there.
On the midterm I felt some of the questions were unclear. Like you asked for the hamming code of a text passage and at first I thought that meant encoding the words with distributions drawn from the passage, and it took me a while to realize you meant the letters. I wasted a lot of time on that. I think you should use more clear language because "The hamming code of this passage" is not well defined, but something like "The hamming code for the alphabet with distributions drawn from this passage" is more clear.
8
30/04/2020 15:15:528I think that if it were possible to decide on a reasonable coursework plan sooner, it would have been helpful for providing more structure and less stress. Also for the lectures, writing on the sides as a white board was slightly messy and difficult to follow. However of course I know that this is a new course and unexpected situation, so all things considered I think the class adapted pretty well.710998454379101075455101010Feature standardizationDistance metricsSupervised and unsupervised machine learningDistance metricsIt is fast paced and has a very heavy workload, probably more than most 200 level classes. But, the professor is nice and considerate, and grades things very fairly.It will probably be difficult and have a heavy workload, but he is considerate and fair. Maybe not the best at teaching but passionate and knowledgeable.Thank you sincerely for all of your efforts! We can all tell you have worked hard to develop the course and want us to learn as much as we can. We all appreciate how you have dealt with the circumstances and adjusted the course plan with our needs in mind.

I think my only small piece of advice would be to keep in mind that not all students enter your class with the same background. And so while some students may be able to jump straight into the concepts (ie machine learning, classifiers, using very specific packages) with no problem, it may be helpful to slow down and provide more background or resources for students who do not have as much knowledge. Either that or add more prerequisites to the class.

Thank you for all that I have learned. Your passion in the topics is very inspiring.
9
30/04/2020 15:50:255Made use of polling/raise hand features686765433578810443310108Keystroke BiometricsFourier TransformGoing through the different types of biometricsFourier TransformThe course is basically machine learning applied to some interesting authentication problems. Concepts from statistics and computer science come together.Professor Kumar listens to students when they bring up concerns, but they need to speak up. The lectures are unstructured however, which might not work for everyone.
10
01/05/2020 03:11:5710It was overall pretty good I think898984 star4 star3 star4 star810910105 star4 star5 star4 star101010Distance measures and metricsI think machine learning was a little rushed - maybe make it a prerequisite so you can go deeper into the biometric applicationsHonestly just enjoyed the classCan't really pinpoint 1I found that it was a really interesting class, and if it was taught again I'm sure the experience will only be better due to improvements from the first timeProfessor Kumar is very open and always strives to improve. He responds to feedback well and adjusts his teaching or lab to suit students needsThank you for being very understanding as a professor, enjoyed this class a lot
11
01/05/2020 19:28:358It was good8107893 star5 star5 star3 star8661095 star5 star5 star4 star9810Performance metrics / evaluationattacks on systemsfinal project/paperdata compressionDefinitely take it if you have any interest in applied data science. It's a very interesting elective but it's not for people who don't get excited about their own work. The work can seem tedious if you're not interested.The labs are well thought out and interesting. If the content is interesting to you, engaging in the lectures is easy. Otherwise it can be tough to stay motivated if the content doesn't spark your interest. Rajesh responds well to demonstrating your interest in the topic. I think GitHub classroom keeps our repositories private (so other students can't copy our code) but it doesn't help flesh out our profile much (other than show that we are actively committing to some kind of repository. It just doesn't show which one. I think.
12
02/05/2020 14:19:496we could use the breakout room feature.336751 star3 star3 star3 star127884 star4 star4 star4 star678keystroke authenticationhard to say, a lotentropy,information gain and random foresthard to chooseThe lectures are not helpful at all, and you should be prepared to self-study.The professor is accessible,nice and smart, but not good at teaching. So prepare to self-study a lot and receive a high expectation from instructor.You should definitely improve your lecture by making them more organized and interesting.
13
02/05/2020 16:04:468Synchronous online lectures are not very good in general - we can't slow down/speed up/replay like recorded lectures and be as focused as in person lectures236761 star4 star4 star4 star128974 star4 star3 star4 star858KNNFourrier transformfeature selectionKeystroke dynamicsIt's mostly self-study, watch others' lecture videos, read others' slides and papers, and some textbooks. Lectures only touch on the concepts briefly.sorry to be honest but I would not recommend so
14
03/05/2020 12:05:478This is hard because I feel you worked to get us to talk but we as students didn't talk much. I don't know how to improve that besides forcing people to talk a certain amount during the semester (assuming they can make class). I sincerely hope you never have to improve this skill.898874 star5 star4 star4 star879994 star4 star4 star5 star899I really understood the biometrics systems Ideas. The entropy concepts were hard. I really enjoyed the attacking systems ideasI disliked some of the more math stuffI'd tell them it is a good amount of work and outside research but can be very funI'd tell them you'll learn but it can be disorganized at points. the labs will be the most helpful parts and if you are willing to do the outside reading you can learn as much as you'd like
15
03/05/2020 20:48:066Rather than using mouse to draw graphs, use something else547753 star3 star4 star3 star644753 star3 star2 star2 star778Machine learningFourier transformN/AN/ATake machine learning and other data-related course before thinking about taking this one
16
03/05/2020 23:48:4210Use breakout rooms for small discussions before sharing what we have discussed with the class886654 star4 star5 star3 star8888104 star4 star5 star4 star81010Feature extraction and selectionImproving accuracySwiping labPerformance Evaluation labIt's fun and interdisciplinaryGo all out and learn. Don't worry about gradesThank you for the great semester!
17
04/05/2020 17:52:406It would have been nice to make the meeting reoccuring on zoom789964 star4 star4 star3 star736735 star2 star2 star2 star788classifiersthe topics in the 2nd half of the semesterkeystroke dynamicsfusion I would say it was an interesting topic to learn about in the classroom but the labs were too difficult for me to effectively learn as I was caught up in the minute details of coding.
18
09/05/2020 01:20:028It was a bit hard following the slides and discussion, but I honestly don't know how it could have been done differently. I am a TA for a class and found it hard adapting the material to zoom. 8107774 star5 star5 star2 star851010104 star5 star5 star5 star101010Gait Biometrics :)StatisticsPhysical BiometricsI thought all topics were interesting. Go for it! Many cool things to learn.It's a research based class so you should be willing to read papers do some research. But if you don't know python, ml and statistics it might be very challenging to you!Overall the class was enjoyable. I personally learned a lot. I wish we had more in class activities (well before covid19 happened) like collect data in class (I know you mentioned this before) or maybe try to spoof a biometric.
19
09/05/2020 15:09:578The labs in the second half of the semester could be easier998974 star5 star5 star3 star991984 star4 star5 star4 star8910The biometric system designMachine learningThe biometric system designMachine learningPros: Interesting topics, accessible professor. Cons: lack of test cases for the labs. Gradings of labs are vague and unhelpful Very accessible and responsive professor. Always wants to offer intersting coure topics within the cs departmentThe lab grading is not very reasonable. Purely based on result. For lab 3 and 4, points were deducted for no reason specified.
20
12/05/2020 11:43:168-10106794 star5 star5 star4 star99101084 star5 star3 star4 star101010Confusion Matrix, KNN, Decision Trees, Series comparison (gait verification)OneClass SVMsPretty much all of them. I wish we were able to practice face recognition as it is something that is very interesting for me- I already told all such friends that I think this was the most interesting class at Haverford that I ever tookJust like I said before, I would say this was definitely the most exciting class I took at Haverford. The class work and home work require a lot of time (my partner and I have spent 20+ hours for each lab) but the labs are incredibly interesting, you learn a lot, and Rajesh's theoretical explanations in class make sure that you understand all nuts and bolts of what you are designing in the labs. I would definitely recommend this class to everyone interested in Data Science, Biometrics, and researchI absolutely loved this class and I don't understand why some people from 107 didn't like Rajesh's lectures as he mentioned in one of the lectures. I think Professor Kumar made sure to provide us with theoretical understanding before diving into the labs which was incredibly important for me (for example, before implementing frequency comparison, I really liked that professor spent a lot of time explaining Fourier transformations to us in lectures. Without these explanations, I would still finish the lab but I would just be shooting in the dark and just guessing why everything worked the way it worked). I think we were able to cover an incredible amount of material over such a short period of time and I really loved that we not only just did the coding but we also had to write reports, analyze our findings, read scientific papers. I have never done this before, so it was a very useful experience and now I even became interested in doing research and data science! Also, maybe I just got lucky with my partners but I think pair programming was an amazing idea. In the beginning, it was hard, but in the end, it taught me how to better communicate my ideas to other people, how to step back and analyze my code, how to take advice and how to give constructive feedback. I think it will really help me during coding interviews with communicating my thought process and reasoning behind my choices to interviewers. At the beginning of the semester, I had to choose between Compilers and Biometrics, and honestly, choosing Biometrics was one of the best choices I've made. I am really happy I got to take this class as it has taught me not only a looot of technical material but also let me practice my soft skills and even try a little bit of research.
21
13/05/2020 18:15:428It is good now886774 star5 star4 star5 star757764 star5 star4 star4 star888SecurityFourier seriesSecurityFourier seriesIt is great.
22
14/05/2020 02:37:547Maybe cover a little bit less material and add more Q&A sessions688954 star5 star5 star3 star686784 star4 star4 star2 star7910Gait and feature extraction in generalRelation with machine learning (I think it is because we didn't do hands on labs with machine learning)Spoofing and the last lecture on facel and fingerprintsentropyThis is a challenging course but the material is extremely interesting. This course requires a lot of time commitment outside of class, but is overall worth it.Thank you for the flexibility. It really helps me get through this hard time.
23
14/05/2020 11:46:269Requiring video presence. The class felt less engaging with everyone's video turned off7109974 star3 star4 star4 star7651083 star4 star3 star5 star989I understood the topic of linear regression and regularization the mostFourier coefficients were and still remain the most confusing topic for meI enjoyed learning about the method of biometric security attacksFourier coefficients were and still remain the most confusing topic for meJust go with the flow, things can be hectic in terms of the class. Always talk to Rajesh or the TA's and get clarification, additionally, always ask the rubrics for the labs. Go with the flow, things can be hectic. Rajesh is a nice professor, but can leave out things that bring a lot of clarity to assignments. Ask when things seem unclear and ask for rubrics before the labs so you know how things are being graded.While the class felt a little hectic at times, I did enjoy the class a lot and learned quite a lot! I'm very happy this class was offered this semester
24
15/05/2020 10:59:207I think more interactiveness would have been better.468983 star4 star4 star3 star544885 star3 star3 star5 star10104Facial BiometricsKeystroke DynamicN/AKeystroke DynamicsI would tell them that I would not recommend it.
25
20/05/2020 20:00:527More student participation 897863 star4 star2 star4 star775774 star3 star4 star4 star9610Swiping and Gait Authentication Face biometrics Spoofing Not sureTake it!
26
21/05/2020 11:52:094I feel like we tried enough to make the class experience good10107875 star4 star1 star4 star9798105 star5 star5 star5 star101010Fusion based biometricsKeystroke dynamicsFusion based biometricsKeystroke dynamicsThe class is interesting and and the labs are challenging in a way that will make you learn a lotThe professor is understanding and is willing to help you both inside and outside the class just to make sure you succeed in the class
27
21/05/2020 12:03:3110recording all lectures977843 star5 star4 star4 star671010105 star5 star4 star5 star10109Classification Pipelines and steps to building an effective biometric systemsome machine learning topicsswipe-based authentication and its featuressome statistics topicsIt's a great course to get you started on machine learning and to give you a solid background in biometrics and how to set up one.
28
29
Average of overall responses7.67.047.847.486.66.365.926.728.527.568.568.728.76
30
Grand Average7.542857143
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100