<p>I’m planning to concentrate in computer science and eventually go into research (right now I’m interested in neuroscience research, but this could change in the future). I’m wondering which linear algebra course I should take. I have the option of taking the “regular” course or the honors course. At first, I assumed that it would be more beneficial for me to take the more proof-oriented honors course. However, I’ve been told by a few people that I should look into taking the regular version of the course, since it apparently focuses a lot more on practical applications and computations which will benefit a CS concentrator more than proofs will.</p>
<p>What are your thoughts about this? Should I take the more computational “regular” course or go for the more abstract honors course? Which will benefit a computer science concentrator more in the long run?</p>
<p>Given the description you provided, and as a current practitioner in software, I would recommend the more practical course.</p>
<p>Linear Algebra is a fundamental component of many Machine Learning techniques, which are an important aspect of “Big Data” and analytics software that is so in vogue.</p>
<p>But if you think you would prefer the honors course I doubt it would limit you from a CS perspective.</p>
<p>So just to clarify, the computational linear algebra course would be the better background for Machine Learning, AI, and so on, right? </p>
<p>I disagree. The hardest classes in computer science are the proof based theoretical classes. Proof-based linear algebra is one of the best places to develop proof skills because the theorems just make sense. The computational aspects are pretty easy and with the Honors Linear Algebra background, I doubt that you would have any trouble picking up Strang and working through it. </p>
<p>Futhermore, if you get into optimization, which everybody eventually does, the more abstract your thinking the better off you will be. </p>
<p>Well you can take my opinion with a grain of salt- I do not have a CS degree, but I have been tasked with implementing various ML-lite tasks from scratch (circa 2007) and a practical/computational course would have helped me a lot more than a proof-based course (plus I probably would have washed out of the honors course).</p>
<p>But yes, if my daughter were asking my opinion, I would recommend the computational/practical course. I don’t think there’s a wrong choice here- you should decide which course you will enjoy the most.</p>
<p>There are others on CC that are more qualified to answer- @ucbalumnus, @PurpleTitan do you agree with @ClassicRockerDad? Or maybe @DrGoogle?</p>
<p>If I do go with the computational course, do you think it would be feasible to self-study all the proofs stuff in the future, or is it something that can only really be taught by a (good) professor?</p>
<p>The reason for my indecision is because I have had no exposure to linear algebra before, so I’m not sure what to expect from either course.</p>
<p>I’m in the exact same situation. Stuck between linear algebra and the honors version as a first year CS major. I have no idea what to expect from either one.</p>
<p>I thought the first math class that is taught with proofs in college is discrete math. I think since it’s a first class, it’s taught much more slowly, not sure when it gets to linear algebra, a student is expected to know the basic proofs already or not. Maybe this depends on different college. But at my daughter’s college it will be at disadvantage if not discrete math not taken previously because linear algebra more of an advance math class.</p>
<p>Check out this link from Cal Newport about learning proofs for Discrete Math.
<a href=“Case Study: How I Got the Highest Grade in my Discrete Math Class - Cal Newport”>http://calnewport.com/blog/2008/11/25/case-study-how-i-got-the-highest-grade-in-my-discrete-math-class/</a>
and more link from Harvard of why we do proofs in CS.
<a href=“https://www.seas.harvard.edu/courses/cs20/MIT6_042Notes.pdf”>https://www.seas.harvard.edu/courses/cs20/MIT6_042Notes.pdf</a></p>
<p>Good point about the discrete math being the course used to teach proofs. So I concede that my argument for taking the honors linear algebra is without merit. </p>
<p>I’ll just say that linear algebra from a theoretical point of view is just a thing of beauty, but it probably wouldn’t hurt you to take the computational version, and might actually be helpful. I found the computational linear algebra among the easiest classes I’ve ever taken, but it could be because Stang’s book is so well written. </p>
<p>Thanks! I’m actually sort of relived now, haha. I looked more closely at my scheduler and realized that all the honors linear algebra slots overlap with other courses that I want to take (like CS) so it turns out that I’ll be forced to take the computational course after all. I’m just hoping that I will be able to self-learn the abstract side of linear algebra in the future, because it certainly seems interesting. Thanks for the help, guys! </p>