Video 0: Welcome, teaching mode, overview, assessment, academic integity, Campuswire, Slido, exam question gallery

0:00 Module introduction

0:04 In the first parts, I'll tell you a little bit about the organization of the module. Um, some fairly straightforward things, but then also about assessments and uh how we'll conduct the module. Um, so let me just repeat what I said before. The goal for today is to show you what the module will be like and then we'll restart a little bit in terms of the material next week. Um, but for most people, I think for everyone, this should be an elective module. I think it didn't exist for a while, so nobody can have this as a compulsory module. Uh, if you're here, then because you want to. Um, and I want to make sure that you're here because you want the thing that's actually coming and not something else. Uh, it's always a bit hard to tell from the few sentences on the module uh, handbook. Okay? And we'll talk a little bit about assessments as well.

0:58 So hi again um I'm Sebastian M uh wilt uh the German version um module is going to be in English because the master module uh of course you can also answer any question or ask any questions answer any exam questions etc in German. Uh our uh tutor will be Muhammad who is uh who was here saw you earlier.

1:26 Um I'll I'll do the tutorials. We'll we'll talk about that later and um Mohammed Omar will look at the submissions of your exercise sheets. Um most of the not at all a copy paste error. You haven't seen this. This is obviously alk bio.

1:52 Uh but the link is also on the Ilias course that you will have seen as always I'm um trying to get you connected outside of the lecture uh as well outside the class where we meet. Anyways, and we'll use uh this tool campus wire for that. um where you can join. You need this fourdigit pin to join otherwise um the link is is just public but you also have the link and the pin on Ilia so don't need to remember it now um you've already seen Slido I'll come back to that and as always as mandated by the regulations the final exam is the final mark and there's formally at least no other option uh to be admitted to the exam, you have to do the exercises as the standard kind of policy. Looking at the size of the group or the size of the cohorts, I'm leaning towards oral exams. Um, but we'll decide that in a week or two.

3:01 Communication and digital tools

3:01 Okay, that was the module in a nutshell. Um, in a bit more detail, some of the bits. Um, Slido, uh, many people may have seen it before. The goal is to keep you awake a little bit more than you would otherwise be. And uh we'll use it for for two separate things. Uh one is I'll show you an example in a second of uh asking polls to the audience. So you can tell me a bit where you are, but you can also use it to ask questions during the lecture. You can raise your hand or just uh shout it out. That's fine as well. Uh if you prefer not to, you can ask anonymously on Slider or you can just type it on Slider whether you use your name or not. Um that's another option. It's a little less disruptive and I can like cover the material I wanted to cover and then send to questions. I'll try to be attentive to the Q&A on the side, but I'm not always good at that. So sometimes uh if I miss your question and you feel the point has passed where it would made s would have made sense to discuss it then maybe feel free to point it out.

4:11 Um so let's let's test this briefly. Um there's the tools that we use in this lecture. the first three and there's some AI tools that I'm just curious so I included them. Nothing else really behind the question.

4:58 The other benefit for me for from Slido is I don't have to count heads. I can just count uh votes or can let Slido count for me. That's what computer science is all about, letting technology do the job. Um and because people can join on the live stream. So if you haven't uh noticed from the infinite tech setup at the beginning, if you if you must, you can watch from home and join on Slido. It's always nicer if there a larger group here on on site, but still I think on balance the benefits of having a live stream is larger than of not having it and forcing everyone to be sleepy and in the room. Uh so that's that's an option you have. Uh so 16 might roughly be the number 17.

5:47 So people are well aware of Ilas. That's good. Everyone has used it. Um would you know about the meeting today the lecture if you haven't seen it on Ilas? I guess you can see it on Marvin. The uh Ilas is not needed. Um but you probably don't survive long in your studies if you've never touched Ilas. So I guess and everyone has used chat GPT but not so much some of the other tools. Uh that's uh whatever whatever you want to make from this. Um none of the above. Okay. I was I mixed up the two. That would have been surprising.

6:29 Uh what are the LLM based chat systems? Do you want to shout out what what are which did I not include? Grock X AI perplexity. Yeah, I guess now nowadays any kind of web search has some sort of AI built in, but it's a bit limited. Other big ones, >> finding stuff like >> Deep Seek. That's right. Yeah, that's right. Um don't really use the uh Alibaba ones. Anyways, um there's a bunch more and there's new popping up every day. It's going to be fun to watch this again in a year or so and laugh at, oh, that was the good old days when this was still around. Okay. Um let's focus.

7:17 The important thing was that you know Elias. Uh the other thing I should have pointed out is Slido on campus wire is somewhat well known I guess from other modules uh that I was doing but also half of you don't know it. Uh and so uh I'll briefly talk about what we what these are and what we'll use them for. They're all fairly intuitive. You'll find your way through it.

7:45 Course content and objectives

7:47 But let me talk briefly about the material in this module before we spend more time on organizational bits. Um it's a bit of a an unusual module at least for me. Uh it is still an algorithms module by and large. So a theoretical computer science module. Uh but all of the questions that we'll look at have some motivation from from computational biology, bioinformatics, biology uh depending exactly where you want to make the cuts. So we'll have to dip a little have to dip the little toe into biology and especially some lab techniques and some molecular biology. uh I called it backdrop which really just means it's there in the background uh for the scene but it will not be the main focus. Uh we won't cover a lot about biology itself and I'll try to keep the sources that we use somewhat informal and non-scolarly. So people who study bio biology would probably not be happy with some of the simplifying versions of that. I think for the module here and what we need to know to use the algorithms it's going to be good enough um is a fine fine line between being uh sloppy and imprecise and being approachable.

9:09 Uh so I'll I'll have several sources and I'll rely on you to also point out if you have others. Um we'll we'll come to that uh next week in more detail. I want to stress that the main material that we'll do is your standard kind of algorithms theory lecture. So we'll define problems. There's going to be a part before the definition that motivates why do we do this. Then there's usually a a disclaimer box saying oh and now we simplify from all of these problems that you would have if you solve the the actual problem from the data that you get from from wet lab experiments. You have to make some simplifications to start and then in this abstract model we can decide what algorithms to use. We can prove things and at the end we might come back and try to see how big is the gap between what we can achieve and what a biologist would want. And sometimes the gap is not as big as it first seems and sometimes it is a sizable gap that's left and uh some of these things are evolving. It's um still a relatively young field um bioinformatics uh with some some results fairly fairly flexible and young and uh still still moving and sometimes people come up with better ideas.

10:34 I hope that's roughly um the picture. If you have questions uh you're of course just shouting out or put them on Slido. So um we'll survey a bit um what algorithmic contributions have happened in the past to computational biology and hopefully equip you with the tools to contribute to those if you so choose. Uh we'll learn just a little bit of biology as as much as needed uh as is needed but not much more. Although I'll point you to kind of as a digression when whenever I work on this, there's so many interesting things to learn because this is about about life and us and everything is uh there's a lot of exciting things that are um well worth knowing and learning about but you won't have those on the exam for this module. So you can uh dig into that as much as you want and I think we should use the group to kind of have some fun together in exploring some of these topics. The module will not uh go much beyond the the very basic things we need. Okay.

11:43 And um yeah another key goal is I mentioned before uh we'll take a step back make simplifying assumptions define some nice mathematical problem and then try to solve it. You should keep in mind where the limitations of these things are. And the goal of the module is also to make you sensitive for these kind of limitations.

12:06 Thematic curriculum units

12:09 Preliminary plan um is a bit too much. So I'll probably have to kick out one topic, but we'll see. Um maybe I can give you a little bit of each topic instead of kicking out one. Um we'll start today with this little example of one puzzle that might come up in such a setting and then we'll look at uh how to find patterns in in genomic sequences which are just long strings without there's no commas there's no points no spaces >> and still they have some meaning. How do you figure out what that meaning is? And then we'll work our way to comparing sequences. What does it mean to be uh similar in that sense? Uh we'll look at how can you uh from these lab techniques that don't give you at all a possibility to read your genome, how can you still compute it? There's a lot of uh computer computer science techniques behind this.

13:06 Um and then we'll look a bit take a step back um go back to some some basics from computer science that we'll need for the last bits. uh spec specifically is searching in strings and indexing strings and we'll apply that uh to searching in whole genomes. And then there's a few um two two other topics that are bit separate from the rest um just to show more of the breadth of uh of bioinformatics.

13:36 uh RNA is another type of molecule that serves roles in your cells and u expresses genes and phoggenetic trees are about the tree of life. How close are different species to each other and uh how can you compute those? Okay, that's that's roughly the plan. As I said, maybe uh one of these topics will have to go. I'll try not to uh to skip one entirely.

14:07 Assessments and requirements

14:09 Okay, assessments. Uh, we'll have a final exam in one form or another. Um, oral examination I think is nicer in M's modules. This is more flexible. But, uh, this will mainly depend on uh the numbers.

14:28 Most of the material is relevant for the exam. There will be some specifically marked parts that are just for your interest that go deeper than I would want to ask you about in an exam situation. Um what is included though is everything we do in the tutorials and um I'll come back to that um in a bit. Uh for for these for this lecture I've also made some some slides that have a green background. That's whenever we talk about the biology part and for that I guess it's uh only superficial knowledge is needed. You should you should be able to explain the problems that we worked on. So you should uh makes make sense of the narrative but I won't quiz you on the specific uh names of some chemical compounds or I don't know uh what does DNA stand for. This is this is not what I'm interested in in the exam. Okay. To be admitted to the exam, uh as I said, standard rules. Uh we'll have a uh weekly exercise sheets. Well, the first one will come out today or tomorrow and then you have a week to work on it. So next Friday is the submission and then at the moment we have a slot on Fridays which will start in two weeks uh where we meet and discuss the solutions.

15:53 Um it's also again this lunchtime slot which I think is not great. Um, if in a week or two we have to I I'll ask around. If it's possible to move that to a more convenient slot where all of you have time and we find a room that's large enough, then I would suggest moving it. But um you'll you'll hear from me about that. If not, then it stays on Fridays. And so no exercise class tomorrow. Nothing to discuss from this week. No exercise class next week because you'll just be working on your first sheet. In the third week of classes, we'll start. Okay.

16:40 Uh the exercise sheets themselves, they'll pop up on the public module website, the hand corrected link from the first uh slide. Um you'll submit it on in groups. uh try to make it not more than four people. I think three would be better. Um if the group size stays like this, I think that should be manageable for the marking. Um standard kind of uh kind of things. Um I'm not insisting on students presenting solutions in the tutorials, but it's encouraged. I think it's nicer instead of hearing the same person talk that already does the lecture talking. It would be nice to hear some other voices and we'll we'll have an eye out for submissions where you've done a good job and um then we would ask you to do it but if you don't want to I'm probably not going to force you. Okay.

17:44 Um as in in master's modules I tend to see the exercise classes as an offer for you to engage with the material and uh the requirement of these 50 point 50% of the points is intended not to be a threshold that you cannot pass but rather an encouragement to spend some time with the module. So we'll mostly mark and give points for sensible attempts at the question even if you don't fully solve it. Okay, it's different from a written exam though the marking would be would be stricter. Uh I think that's that's good enough in this case. Um it makes our job for marking also a bit more efficient but it means you're a bit more um responsible for your own learning. um you can probably make your way through with group submissions and so on without having done much and still get the exam uh admission. Uh but the exam you'll have to do on your own so don't cheat yourself, right?

18:45 Okay. Uh also some some slides about AI stuff. Uh we talked about it earlier. It's fast evolving and these tools are really cool but uh they can't really help you learn and the more advanced the modules get the trickier the distinction becomes. Yeah, I think the tools are also not as good with the advanced exercises yet. At least still um basically don't deprive yourself of the opportunity to solve some nice puzzles. If you're sitting in this room and you don't at all enjoy puzzling over some math questions, maybe the module is not the ideal choice, just to be honest. uh um many of the many of the questions we study and many of the problems in the exercises will have a sort of a puzzle component and uh the fun really comes when you enjoy these kind of things. You can also work your way through if you don't enjoy them all that much. That's fine. Um try not to destroy the puzzles by using AI on them. specifically uh it's okay if you use it for background research and for this module in particular asking it some gen generic questions how does this whole uh protein synthesis in cells work again I think this is amazing uh use cases for for jet assistance because you can adaptively quiz about the parts that you're interested in and then stop when you have had enough Um, it's not okay if you do major parts of the assessments. Uh, if you don't acknowledge what you've done, um, just don't do it. Be reasonable.

18:46 AI usage and ethical conduct

20:44 Same, of course, applies to copying other people's work. It's not just about plagiarizing AI stuff. Hope that goes without saying. at the level you are um you have to work things out yourself or at the very least if you do if you copy someone else's work means you cite it then you would have put a proper reference and say what you did I think I put this on the exercise sheets if you if you feel like you're better working in a larger group on some exercise problems then do so but every group has to do has to do their own write up and you should acknowledge who you who you've worked with just for this kind of uh we're moving more towards a scientific mindset. So the uh culture of proper credit for who achieved some some fact is important give proper credit and site people that you worked with.

21:42 Student habits and success

21:44 Let's do this quick one to wake you up. Um, I know this kind of gets old. Uh, if you've seen it before, then um bear with us for 5 minutes. The idea behind this is if you anonymously ask people various um various character traits or various life situation criteria properties and also as one of them whether they self-reported as cheating on some assessment then you can do a correlation analysis what kind of property maybe if a person okay what can I pick that's uncontroversial uh hardly anything I guess um a person that prefers blue pullovers is more likely to cheat. It's just a correlation. So, you don't know which cause is which, right? But there was one thing um in in some studies that popped out as strongly correlated with whether people were more likely to cheat. And that's what I'm what I'm asking about here. >> And I want to I want to give it a positive spin, but I I'll I'll let you think for a moment.

23:45 I f the answers so that you're not too influenced by what the others already wrote. Show them in a bit. Okay, there's some some fun ones. Let let me just show them. And uh there's a bit of an overlap between people in this in this module and other modules I've done where I use the same question. Uh yeah, it's never about cheating, right? It's about getting caught. Um laziness. They sit on the last row. Okay. So the kind of right answer is is in in a way among them ask GPT. That's an interesting one. Are people that heavily use AI tools more likely to cheat? I I don't know. So, the study is too old to tell. I think this is an interesting one. I'd be surprised if the correlation was that strong because 100% of people in this room reported using chat GPT. So, it's a you can't get much signal. Comments. I mean technically uh there is already a study that shows that doing a lot with JPT uh well you unburned to think so basically if you encounter hard problems again afterwards you will at some point have to ask because you're >> forget how to think about problem itself.

25:52 Okay, the comment was there's there's a study apparently that shows using chatb heavily um makes it harder for you to solve problems afterwards on your own. That's not implausible. I I don't know the study, so don't know um what they've what they've looked at in detail. That's not unreasonable um to expect. I guess it depends a bit on how heavily you use it and so on and so forth. Uh but yeah, so in this case um AI use was not the focus. Um what they what they've pointed out was uh whether or not people procrastinate or self-report to suffer from procrastination was the number one predictor. So the fact that you spend time on things where you in hindsight think oh this was stupid I would have wanted to do something else. uh combined usually with a with a deadline driven mode of working that you push things because you procrastinate on on doing other stuff that's much more fun or maybe just um let's not psychoanalyze why people do this. Um but it leads to then having to do it last minute. Um and I guess it just means once you become desperate uh you start looking at all options even if you know that they're not good or morally wrong or whatever else. Which also means the best way to get out of this uh this problem is to better manage your time so that you don't fall into the last minute uh the last minute desperation.

27:31 And ideally, don't procrastinate all that all that much either. Uh it's a bit easier said than done, but there's simple tools to better manage your time. I'm always tracking, for example, what I work on. Not so much because I'm interested what I spent my time on, but when you have to click what you now want to spend your time on already is a bit of an incentive not to do something different. And there's many small things like this. is not really part of computer science but it's part of life and growing up and so on. Um that's my little kind of nudge. Uh think about whether you're you have a system that works for you in this respect. Let me go very briefly over the other tools. Campus wire is essentially a Reddit for lectures. You can ask questions, you can answer other people's questions, you can upvote answers and questions.

28:12 Campuswire and communication

28:25 Um the main distinction is that you can ask things anonymously and it's a kind of closed space just for this lecture. Plus uh you can write things in markdown and use latte formulas which is just a nicer way to do it compared to other forums uh whether or not onas or elsewhere. Um, it supports I I would mostly want you to use public questions for anything that other students might be interested in. Use a public question. Only really use the private questions when it's spoiling some assessment or it's really really something that only concerns you. Um, most of the things should be a public question on campus wire. So in a way when we're here together use Slido for in immediate questions or find me after class. Um for any time outside of the qu lecture slots if you need anything asynchronously the default will be campus wire. If you send me an email with a question that I think other people would also uh have have something from the answer then I will just redirect you to campus wire anyways.

29:39 Okay. Um, otherwise I think it's it's relatively straightforward to use. There's also a chat room feature that you can use. Um, but for the lecture I'm I'm not going to heavily work with it.

30:00 Let me just reiterate why I think Compass Wire is worth the hassle. I know it's another account. It's another another thing to check. You can set it up so that you get emails if someone posts a question. So you don't have to monitor it or use their app. I think it's important that everyone gets the same answers if you ask questions. That's what I don't like about office hours. Uh it's fine for some personal things, but then if it's a question about the material of a lecture, I'd rather have everyone see the same answer. And I want to nudge you into teaching others. If you help others understand the material, that's the best way for you yourself to understand it. And by answering other people's questions, that's what you can do.

30:49 So the only thing that really means is I'll redirect you to campus wire by default. If you ask a question where I think some other student might be able to answer, I'll not answer right away and give other people a chance. And if they do answer, I'll endorse the answer or augment it if there's uh some parts missing. So even if it's some some other student answering, you'll know whether you can rely on the answer.

31:18 Um any questions on this? There was one um question on Slido. might as well talk about this now. Uh, is the module classified as theoretical or practical? Does it have a formal meaning? I think it's a theory module. I don't know if the module handbook has a specific uh binary property.

31:46 The thing is that also the updated module description I couldn't yet find even though I wrote one a year ago when we updated um the new uh exam regulations. I think what we have on the website is still the previous one.

32:07 By all intents and purposes, I would consider it content-wise to be a theor the theoretical module. Everyone knows Elas. Everyone loves Elias. Um, so I'll restrict usage to the minimum I can. Namely, everything that has to do with marks or with personal information. uh if on campus wire if you want to sign up with a fake name that's fine right I don't need to be able to identify people it's nice to know your names a bit and so um uh you don't have to go anonymous um but if you prefer not to use your real name and uni address that's all right um Elas is the official thing anything with marks will go there and the rest is more um informal Um, and whatever I can I will put on the public website. That's more convenient in my opinion.

32:15 Exam resources and literature

33:14 Uh, let's talk about this another time in more detail. I want people collectively collect questions for possible exams. What these will look like depends on the exam, uh, whether it's oral or written. um and you probably won't start right away anyways. It's just a Google doc that's linked on the website and it worked in previous modules quite quite nicely at least towards the end of the lecture when the exams loomed a bit closer. Um I think it's a a useful resource to help you prepare and over the years we can keep that collection growing. More on that at a later point. Uh the last thing the physical copies are there in the front row. Feel free to have a look. Uh there's no required reading or so. Uh everything you need will be covered in the lectures. But sometimes um it's good to hear something explained from another perspective. And uh books are really good uh if they're good books. And these two are ones I like. Uh we don't we don't cover nearly all of what they do. And there's some there might be some bits that neither of them really cover in that way. Um there's some some other uh sources that I'll point out on the module website.

34:28 You'll find a subpage for each unit with some some other links. Uh but these are the main the main two things that cover most of the material in as I said slightly different way. Now, especially the the first one here is is a very modern one. Um, and it is available as a book, but their main thing is an interactive online version of the text. Now, uh, they're in the US where it's standard for people to buy textbooks, and a book of this sort would very very clearly be above $100, and it would be required for everyone to buy it. I'm I'm just saying this to put into context why they charge for access to their website. It's a different system. Uh it's not a very nice one and um I'm not going to encourage you to spend money outside of this. The books um should be available via the library. At least the uh biological sequence analysis we already have and for the other I'm I'm trying to get one uh ordered but you can also speed this up by asking. Um, still the interactive online text has a lot more than what we'll do. In particular, it has lots of programming exercises. If you want to go um a good deal more into this and maybe become more involve yourself in more practical questions of this sort, I'd recommend it. And they shared some discount codes that I'll put on on Campus Wire for people to check out. Maybe just give it a try uh and see if it if it helps you. Um, it's very very nicely done and has lots of fun extra side stories and so on. Much more than I have time to tell you. It's just one lecture a week.