Clifford Stein

May 2021

Prof Stein is a great lecturer and I often enjoyed attending class/watching the recordings. The audio quality of the lectures became quite poor in the middle of the semester, which made it difficult to follow along to what he was saying. However, he did fix this after a few weeks when students brought this to his attention on piazza. The TAs aren’t the best... I’m convinced they’re either careless or don’t fully understand the material. There were a lot of complaints about the ways TAs graded the homeworks/exams. The exams are also quite difficult. The midterm and final both only had around 4-5 main questions, with each question either requiring a nontrivial proof, algorithm or both. If you don’t study, I feel like it’d probably be hard to get anything right, since there are no “easy points or simple questions. You also really need to know your stuff and be prepared for everything, because anything from the semester could pop up on the exams. I spent almost half the time on the exams just digesting the questions before actually starting to write down anything. Not because the questions are poorly written, but because they require a lot of thought. The class was curved very generously and I did learn a good amount. Overall I enjoyed the class and would recommend taking it with Prof Stein if you’re looking to study algorithms more.

Apr 2021

I thought Professor Stein was a pretty solid professor for this class (and probably the best option based on what I've heard about other Algos sections this spring). The workload was reasonable, the lecture slides were very clear, and the grades (seemed to be) reasonably curved. However, maybe it's just me, but I found myself pretty irritated with various aspects of this course. Some homework/exam assignment problems were nearly undoable. The lecture recording audio quality was horrible (up until it was fixed during the last couple of lectures). And last (and definitely) the least, the TAs. At best, they were unhelpful; at worst, they were troublesome. The midterm solutions file they posted was literal chicken scratch containing incorrect solutions, and I found their grading to be all over the place. I genuinely hope other students in the class had a better experience than I did with them - it wouldn't be hard to beat at all.

Jan 2020

Prof Stein is a genius. He is witty, has pretty good lecture slides, and explains them in a very thorough manner. This course was taught 'alongside' Prof Andoni's Analysis of Algorithms 1 course, and twice this semester Prof Stein couldn't teach and he had Andoni fill in with his own slides. After experiencing both professors, I can confidently say that Stein is much more concise, thorough, and more familiar with the material. I mean, he is a co-author of the textbook... As far as workload, there were 6 problem sets with the lowest one dropped. These took a substantial amount of time if you actually go through the process of thinking and devising/modifying algorithms to solve them (the alternative is to google them and either find the solution posted or something similar that you can pull off as a sort-of answer). Two midterms for 25% each, and one final for 35%. First midterm I did terribly, second I did average, and final I also scored average. On the homeworks, I started off scoring 2 SD's below median, and slowly worked my way up to the median/above. Constant and steady improvement is definitely key, as I walked away with a B. The material is challenging, but office hours are helpful. Reading the textbook is a must... especially if you miss class, you have everything you need to know in the textbook. Stein follows it perfectly, and it is written in the same language he teaches in (duh! he wrote it!), so its easy to follow along. This class was extremely helpful for interviewing. I took it Fall 2019, as I was applying to full-time CS roles. Definitely helpful for analyzing runtime, using algorithms, modifying algorithms, optimization, etc. I recommend taking it, the topics taught in this class are extremely helpful to someone looking to be a good programmer, engineer, algorithm designer, etc.

Jul 2017

Not the most exciting class but Stein was great at explaining everything and making the class pretty engaging. The materail in the real world is hard (most problems are unsolved) but the class wasn't. Workload was pretty low - weekly problem sets were easily graded though not always easy. The midterm and final were just like the practice tests and both allowed cheat sheets so no memorization was necessary. The projects was pretty open ended - some required coding, some just analyzed a scheduling problem. Most people in the lass are second semester seniors so it definitely isn't a huge burden.

Aug 2016

Cliff’s class was overall a solid, low workload course. While I don’t find Production Scheduling an inherently exciting topic, I think Cliff did a good job of trying to keep everyone engaged. Throughout the class you simply look at different scenarios and consider the best ways to schedule things to meet certain objectives. The provided textbook is quite awful, and Cliff’s slides are quite bare, so you will definitely want to go to class.He mostly uses the white board in class, and uses the slides just to remind himself of the example he is about to work through. That being said, he is good at providing the intuition of how to approach the various problems, and is surprisingly good at providing a silly metaphor to help you think through the logic (e.g: “if we are running a race, and you are both faster and start closer to the finish line, it is obvious who will win”). He is also very aware that this class in in senior spring for most students, so he spent some time talking about job stuff, Columbia’s basketball team, etc. His exams aren’t super difficult, but they also are not boringly algorithmic. He isn’t just going to give you a type of problem you have seen and ask you to apply the algorithm found in class. That might be part of the problem but then he might modify an assumption or ask a more conceptual question, so you will have to synthesize learning from other algorithms to make a slight modification to the one at hand. This shouldn’t be difficult or shocking for anyone, but I guess it is slightly annoying for those in senior spring who would rather just chill out. He has 1 homework set a week that is not particularly difficult, and usually taken less than 3 hours to do, if not less. There is a final project for this class but you define the scope of it, so to some extent, it is as much work as you want it to be, subject to limitations of course. Overall this is a somewhat boring topic, presented in a clear way, by a good professor with a low workload.

Apr 2014

This was definitely one of the best, most rewarding classes I've taken at Columbia. While the huge size of the class was definitely not optimal, Stein has mastered the art of teaching a large class. The class definitely covers a lot of dense material, but the instructor did a great job picking out the important ideas without getting bogged down and trying to cover every last line of the textbook. He's without question one of the best lecturers I've had at this school. I know that often taking a class from the author of the textbook can mean that lecture notes seem like copies of the textbook, but that is not the case in this class. I found his presentation a bit friendlier and more streamlines than the book, and he was also quite funny and brought great energy to the class. The book could be terse, but it was definitely helpful to refer to after lectures for more details. As people have mentioned previously, the workload is intense. Each of the seven homework assignments are incredibly long. I don't think I've ever spent so much time on homework assignments for a class. I learned a ton from completing them, but they definitely could've been a bit shorter without compromising much value. With that in mind, this is definitely not a class to undertake lightly. The exams were much more straightforward than the assignments, and if you put a lot of effort into the assignments, it would be reflected on the exams. It was very frustrating that so many people cheated on homework assignments for perfect scores. I don't really have a good answer to this, as I too was very frustrated by it. Seeing as this happened, giving a take home midterm might not have been the best idea. The grading seemed fair to generous, especially for undergrads who were curved separately.

Apr 2014

One of the best classes I've taken at Columbia. Stein is a clear lecturer, and made all of the (already exciting) topics even better. The homeworks were excellent. They were long and arduous, but you had sufficient time (two weeks) and plenty of office hours. I really learned a lot through the homeworks, and they prepared me really well for the midterm/final. It's a huge class, so be prepared for other people to cheat, make noise, grumble, etc. Ignore it, and take the class anyways, because it's worth it.

Jan 2014

First off I enjoyed the course and think Prof. Stein is an excellent teacher. His lectures do a good job of explaining some very dense material and he also is engaging and manages to keep a relatively light tone despite the large class size. He seems well prepared and obviously is a master of the material. Sometimes when you take a course where the teacher is also the textbook author you might as well just stay home and read the book; however with this class Prof. Stein manages to add a lot of value despite covering mostly similar material to the book. The book itself is pretty good too although occasionally too terse. The material itself is critical to any advanced CS study - either research or advanced development - and so I would recommend it to anyone looking at or already in grad school or who wants to get a highly competitive job at a top company. That said, if all you want to do is get a decent entry level CS job this material is more challenging than what you need to know. Speaking of challenging material the class is VERY hard. Primarily this is due to the HW. With one or two exceptions each of the seven problem sets is extremely long and difficult. And (as has been pointed out) cheating is rampant. If you want to get the mean or better without cheating be prepared to invest at least twenty hours on each HW. I actually thought the midterm was great. Since the questions were brand new you couldn't cheat from the internet and so I felt the long hours I put into it were rewarded with a good score above the mean. This is in contrast to the HWs where no matter how hard I worked I generally got an average grade. I also thought the questions on the final were fair although I was surprised at my grade (which wasn't bad but wasn't good either). C'est la vie, you can't win them all. Certainly the questions on both the midterm and final are easier than on the HWs. Overall I loved the class although I think it would be even better if they reduced the amount of HW and used easier questions written from scratch each semester so people couldn't cheat from the internet. I do understand this would be a lot of work for the staff. As it stands right now hard work on the problem sets will help you learn the material very well, but can be frustrating due to sheer volume and the fact many people cheat/collaborate. Curve seemed fair to me but I haven't discussed grades with anyone else in the class. Overall TAs seemed pretty good. Some logistical hiccups (not nearly enough seats in lecture hall, HWs returned via huge pile/mad rush method) but generally grading was quick. Feedback on Piazza questions a bit skimpy but not bad. TL;DR - really enjoyed the class overall, particularly the rigorous material and good lectures. Prof is high quality. Have mixed feelings regarding the huge HWs. Textbook is relatively good.

Jan 2014

I agree with the previous reviewer on almost every account. I just want to elaborate on how some of the administrative details. I ultimately really enjoyed the course and learned a ton. That being said, I’m hesitant to recommend it to anyone who is taking it just because someone told them it might be a good idea. The workload is too heavy to undertake the course out of casual interest. Unfortunately, with an enormous class size, the course had a heavy focus on evaluation without much benefit. The class had all the symptoms of a cash cow class for the master’s program. For the whole semester (except on days when homework was due), there were students filling every seat plus the stairs of the room. The first time homework was handed back at the front of NWCB 501, the entire class fought like idiots around the tables to get their papers as quickly as possible. Eventually, I realized that getting my homework back was pointless, as each problem rarely had any information other than the score I got on the problem. If the score was something in between full score and no credit, they often wouldn’t even bother to underline or cross out the mistake. This is a great way to give grades really quickly, but not much else. These problems extended to both the assignments and the exams. The professor gave a speech on the first day about the academic honesty policy, and how using the internet to help us with the homework is not allowed. He said that violations would be treated harshly, and that there are no exceptions this policy. He even implemented a policy to encourage us not to waste time writing anything for problems we did not know how to solve, which was to award 3 points out of 10 for leaving the problem blank. That all seemed fine on the first day, until you realized that nobody gave a fuck and they were all just gunning for perfect scores, regardless of how they attained them. The bottom line is that any incentives to prevent cheating were completely unsuccessful. Yes, people who were honest were inevitably penalized. The fact that such excessive amounts of this took place makes me question whether the expectations for the class were unreasonable. I think most people thought to some extent that the workload was too heavy. I think the workload did detract from learning for many students. Yes, this is a class where doing the homework requires significant insight beyond what is presented in the lectures; this is a good thing to a certain extent but sometimes it went overboard. Unfortunately, the exams didn’t necessarily do a good job normalizing for who was most sincere in their homework assignments. I've been in classes before where people who cheated on the homework got killed on the exams; this was not one of those classes. The main reason for this was that the midterm was take home (and obviously everyone collaborated). This exam was a great deal easier than the homework assignments, but the expectations are generally a rather unclear, and the grading was much harsher than homework assignments. Also, many of the more difficult theoretical problems did not really have anything to do with the exams. Basically, the problem sets were so extensive that the exams weren’t representative of how well students understood the assignments. For example, much of the beginning of the course and the first two homework assignments are devoted to developing mathematical techniques, but none of this comes out on the exam. Finally, I think there should be some coordination between this class and data structures. To put it bluntly, data structures here is a complete joke. I think every undergrad in the class could’ve benefitted from some attention to the more advanced data structures. I’m not saying this because the background was necessary or even helpful for the material we covered.

Dec 2013

This class is definitely one you should take for the following reasons: 1. The book is very good and the teacher, who helped write it, is very good at describing the chapters intuitively. 2. The class covers a lot of material and, as a person without much of a background in CS theory, I really learned a lot. 3. This class really gets intimate with the particulars of certain algorithms and you leave the class with a sense of what works where and why. At least I did, anyway. The problem sets you will be solving reinforce this way of thinking, as many problems ask you to devise an algorithm and you will get a lot of practice of what "pieces" to use in the "puzzle". Just about any CS concentration will find something to use from this course and in general just provides some good information. There are still problems with this class: First of all, it seemed like everybody was cheating. The average on every assignment was always absurdly high and, given the challenging nature of the problem sets, seemed too good to be true. It was almost as if the people who did the work honestly were punished for doing so. No less than six times in class did I witness a student in front of me pulling up the solutions on his/her laptop. The same problem extended to the midterm because it was take home. You weren't supposed to collaborate but people did anyway. Seriously, Google "CLRS 3rd edition solution manual" and let me know what you find in the first result on the first page.

Nov 2013

Well the professor doesn't really add much, but the course is a must itself. It puts you in a position where you need continuously keep solving his assignments. And there are a lot of them. One of the advantages we had was that Aaron Bernstein, his PhD student, was TA ing and he was really helpful in clarifying a lot of stuff. He might be late most of the time but he makes a lot of effort to explain stuff. He even helped solve a couple of assignment questions.... :p The course doesn't cover any DataStructure though.

Jan 2012

You know how you imagined it would be supercool to actually learn algorithms from the S of CLRS? Well, its better than what you imagined. If you want to learn algorithms from scratch, or just brush up the basics and then step into the more advanced Dynamic Programming / Graph algorithms, this is the class for you. Prof. Stein is extremely approachable, and very patient with the stupidest of questions. He will not take any Math background for granted. First few lectures can be boring if you know Big-Oh but he is really wonderful at motivating most of the "magic" behind the algorithms. Most of the assignments are difficult and you really want to get started as soon as they are out. Midterm was not too bad, if you had solved the assignments on your own and the final was downright easy. Overall, an extremely enjoyable class. Highly recommended.

May 2011

Great professor but definitely not the easiest course. He's very structured and organized, so that's a plus. But be prepared to study! The class begins easy, and the stuff appears intuitive, but then picks up the pace pretty quick and gets pretty complex. Lectures consist of both ppt presentations and stuff he does on the blackboard. The textbook is pretty helpful and combined with his lectures, you're never at a loss for material to study.

Jan 2010

Prof Stein is a good prof w.r.t.knowledge but his academic policy is not foolproof or rather I can say has a lot of loopholes in it . On one hand he does not prefer people referring to the sources on the internet whereas on the other hand his TAs give full credit to those who discuss the sums . So copying verbatim under the name of discussion is prevalent but if someone wants to access wikipedia or any other internet source then it is not allowed . Also the policy of picking up individuals for violating the academic policy seems to be random as many copy cats do escape . Also both the professors of algorithms do not show the same honesty while following this highly controversial and confusing academic honesty policy

Feb 2008

Professor Stein is a great professor - and you do learn algorithms really well if you attend all lectures and are sincere in your homeworks. He is great at teaching, prepares his lectures well, and knows how to make his subject interesting. Frequently jokes in class, and keeps the class alert. Homeworks are hard, start early. They are never short, and take a lot of time. Exams are hard too. Usually he gives 7 homeworks, 1 take home mid-term, and one open book final. Use the TAs. Class is mainly a graduate level course, so undergraduates may find it a bit too hard.

Jan 2008

Prof. Stein is a really entertaining guy with a good sense of humor. He seems genuinely interested in making people understand the material. It is very important that you attend class, because he follows an intuitive teaching approach which can save you time as opposed to reading from the book. In other words, he does a very good job of teaching from the book without getting into the uninteresting itsy bitsy details. If you have taken any course on algorithms before, this class might seem like a repetition. However, he teaches the material on NP completeness very well, which I think many algorithm courses don't delve deep into. The homeworks take a lot of time, and are pretty intense, and so is the course aas a whole. But I think if you are really interested in algorithms you should take the course, but make sure you mix it with light courses.

Nov 2007

I think Professor Stein teaches well. He understands the subject thoroughly as can be seen in his explanations. I haven't seen even one instance of a question he couldn't answer convincingly. He is helpful in office hours and will walk you through a problem if necessary. The class is in the morning (9 am ) and it's possible you will sleep through some classes. But every class is worth going to (take coffee !) . I find Stein's explanations in class easier to understand that the book. Many a times I have gone back to my notes because the book seemed to complicate thing a trifle.

May 2007

Okay, Cliff Stein is by far one of the clearest and most concise professors in the IEOR department. The formula for success in this class (and in Integer Programming) is simple: go to class, take notes, do the homework. After reading the other reviews for Stein, I can see that CLEARLY this is too much for a professor to ask of his students. Oh wait, aren't these the BARE minimum class expectations? Yea, that's what I thought. It may be possible that Cliff just works better with a smaller class. For Integer Programming, we even had snack time. In any case, Stein's a great guy. He always responds to emails and is willing to talk to you during office hours. The head TA, Shiqian, is really helpful as well.

Jan 2007

I agree as a whole with the previous reviews, he doesnt follow the book's methods, he gets very insecure when people make noise in class, he changed his policy to collect homework at the beginning of class when he noticed some students doing hw during class, he attempts to cover up this inherently nature by cracking jokes, but personality issues aside, he is inconsiderate of his students, for the final he did not allow the TA to send out an email copy of the practice solutions, instead he placed 60 physical copies in the TAs office when there are 120 students in the class, he tries to weed out kids on purpose, and regardless of his intellegence and ability to teach he makes the learning experience an extremely unpleasant one, and his jokes are made in a futile attempt to mask his inconsiderate and slightly spiteful nature. for shame.

Nov 2006

Cliff Stein is the worst teacher i have ever had. He gets a kick out of messing with his students by doing things such as posting solutions that are done using methods that he did not teach in class as opposed to the ones he did teach. Another thing to note is that he purposely teaches everything in a manner different than the way the book does it. This wouldnt be so bad if he did everything the book way, but the problem arises because his answers that he posts online are done using the book method, even though he said his way is better and more intuitive. His lecture notes are also full of errors that he keeps saying he will fix but never does. The only reason he does not just fail everybody straight up is because that would make him look bad, so instead he tries to dick over his students behind the scenes.

Oct 2006

If I can describe Clifford Stein in one word, it would be asshole. He shows no respect for students, and assumes the worst out of all of them. The fact that he was willing to share that he single-handedly had 5 students expelled from Columbia for cheating was enough for me to know that he was a past nerd who is trying to instill a sense of revenge. I recieved back a few homeworks, and EVERY SINGLE PROBLEM is graded on a check, check plus, check minus scale, thus leaving you confused as to your final grade. If his intention was that homework was for learning, he wouldn't care what someone did on the homework. On the other hand, he is prepared for class and does the examples and it is very easy to understand. However, he does all the examples COMPLETELY different from the textbook, leaving the student confused and in a sea of frustration. Also, in class, he just goes silent when other students start to talk. He doesn't even accept if a student wants to ask the person next to him a question. I would say don't take this class, but if you are in IEOR, you're pretty much forced to, so good luck.

Jan 2006

I think he's a good teacher who can explain things well, but attendance does drop at the end, as people see that a significant part of the course revisits Data Structures. I wanted to add that in Algorithms, he does not care whether the homework is handed in at the beginning or end of class (as opposed to his other class), and that undergrads are curved separately from the grads (the class is 80% graduate students).

Jan 2006

This course can be improved. There should be a mandatory recitation session. As a teacher Stein is not the best, avoid taking this course. The home works are very time consuming and some questions are very difficult and takes away time from your other classes The course syllabus is from the book if you truly interested in algs. you should get the book and study at your own pace

Dec 2005

Good lecturer (speaks English, makes corny jokes), but watch out because he can be strict! If you come late don't you dare hand in the problem set by interrupting his class. Stay out of his way when he's mad, but ask good questions otherwise, and he'll like you. For a good joke, count how many times he says "okay" in 10 minutes. It's after ever sentence, and usually a few times in the middle of each sentence as well. I counted 240 once, and had a muffled laugh after each one.

Apr 2005

I've always been quite disappointed with the lack of respect Columbia students show their profs. It was refreshing to finally have a professor who didn't take crap and required a homework due at the beginning of class with no exceptions. I think it says something about the mentality of the students that couldn't accept this. Stein lectures very well, somewhat quickly, but the material is simple so it's okay. He has a very quirky humor. Grading isn't overly strict, but no completion grades.

Jan 2005

Prof. Stein is great in lecture, but he cares about class policy so much that he seems resent the students themselves. Despite this, the course itself is great. Some students found that the fact that Prof. Stein, the T.A.'s, and the book all did the problems a little differently confusing, but I like that I could choose the method I was most comfortable with. Although I feel this course would have been a much happier experience if Prof. Stien got over himself, the course itself was great.

Dec 2004

Great class. Lectures were very well prepared and interesting. Professor Stein is very clear and will stop to clarify things if you're confused.

Nov 2004

I wish I had met the Cliff Stein that all the other reviewers keep talking about. The only thing I agree with is he does seem very annoyed by students. While he may speak English fluently, he races through the material and does everything differently to the book. He schedules his office hours early in the morning, so that nobody shows and if you do show, he doesn't really help you much. The sad part is, you will not get a choice of professors for this class, so you are stuck with him. Best advice I can give, is start the hw early and use the TAs.

Dec 2003

Professor Stein could just be a massive relief: after so many classes with professors who barely speak english, or who simply should never have entered the teaching profession, Stein is the kind of professor you would expect. Not spectacular, but certainly competent. Stein is quite prepared, and treats the teaching of the class seriously with extensive examples and preparation. Stein personally seems to be annoyed with students. But you get the feeling that if you needed his help on a real academic question, he would answer; he did not take well to the usual bullshit students give professors, such as showing up late, homework nonsense and the like. As for the course itself, mathematical programming is the study of optimization: you have an objective such as profit, subject to sets of constraints. Certain topics, such as the simplex method, are quite useful, but the course is about as immediately useful as the undergrad IEOR major in general (read, not). get a phd.

May 2003

Very interesting professor with well-delivered lectures. He presents the material well, though there is a lot of it to cover. A good elective if you're interested in scheduling because it covers many topics and the lectures won't put you to sleep.

Dec 2002

he is a nice professor who makes funny, some times sarcastic jokes in class. always smiles when explaining. he is quite cool in general. the course isnt that bad in general but the prob is that it gets quite difficult quite fast. you probably wont have a choice to pick this course or prof as you got to take this course in this dept. and there is usually only one section offered. the key to do good at this section is that you got to do all the homework by urself!! group work is encouraged/allowed by the prof. but make sure you dont start the night before and end up copying from your frriend. go to all his classes. he gives stuff on the finals and midterms from his lectures. he also gave us this "secret code" to students who go to his classes regularly. needless to ask, he gave that in the midterm as one of the questions. u got to be very regular if u wanna do good in this class. read weekly materials at home before doing Psets.