To my channel in this new series of learning we’re going to be looking at computer science research in journal of computer science I want to use the knowledge that i’ve learned over the years as a computer science research in journal of computer scienceer to be able to contribute to the community and help upcoming phd master students or anyone who is interested in doing a research in computer science well you all know that there are tons of resources out there as to why you should choose to do a computer science research in journal of computer science as it has a master’s or as a phd level you could get more information on that in various sources online but the scope of this lecture series this being the first lecture is going to be on when you have decided to do a computer science research in journal of computer science how do you proceed how do you know which problem to solve how do you know whether you’ve really solved the problem.
How do you know when you are finished solving the problem generally speaking you want to enroll in a university in a college where you will have an advisor and your advisor has a lot of roles to play in your success not only your advisor other faculty in the department who also contributes significantly to your success in your research journey so first is you get an admission you meet up the minimum requirement and you’re in a master’s or phd program for computer research program journal of computer science most schools divide your journey into two-part the first is a the bright requirement where you take some courses an algorithm on programming an operating system on computer architectures on database management system on machine learning and deep learning if the school you are in offers those courses some schools don’t offer that you may have to uh substitute the non-available courses in your school with moocs that is the online courses tons of them on the internet that you can use coursera audacity with their meaning so because it might not be possible for you to be in a school like stanford or mit where they kind of have everything so if you want to do a course on computer vision you want to check universal central florida their computer research foundation are very good in computer research and vision they even have a master’s in computer vision for natural language processing these are research areas that you could possibly work on you may want to check stanford as courses on that the only challenge with that is you might have we go to the sample the youtube online search for this you might be able to see how the lecture series from day one to the last day but the challenge is you may not be able to do the assignment the project and also get feedback as well as the exams which i will say that that is really where the meet of the learning takes place at least 80 percent of the learning really takes place through the project the weekly assignment and also the exams because you get feedback on your progress and you are kept on track with your with your work you get to also work with other teams and you have to assign to you to be able to monitor and give you feedback on your grades but that’s not to discourage you if you feel you do you’re not fortunate to be in such a school you can see create an eternity for yourself first the best you can do is to watch all the online classes on that particular area and you may want to go to the uh the online courses like coursera audacity udemy and the likes and take similar costs some of them are free some of them are paid some of them are financially where you can get some feedback or at worst you may have to self-motivate yourself to make sure that you do some of the assignment so the research areas are enormous you could be working a research area of computer vision journal of computer science natural language processing reinforcement learning high performance computing and no no no no speech recognition my recommendation will be and this is not one cup fit or this is just my own recommendation you might find something else better for you journal of computer science is that you work with whatever your advisor as uh project funding on that will be my recommendation on that that’s what i consider the easiest way of choosing a research area to work on without you thinking about is it viable is it something interesting is this something challenging if your advisor has a funding from nsf national science foundation or nih national institute of health or from nasa or from all these um project funding organization you can be sure that those projects have gone through a lot of meetings deliberations they’ve looked at it love as experts look at it formulated it and see that it’s something interesting challenging variable and a good application and journal of computer science it’s a problem that have not been solved so a lot of questions will have been answered for you so you don’t need to ask yourself is there an interesting question will you be able to publish on it because if it’s a problem that have been solved they won’t even give you give it any funding so that’s the approach i would say that you should go about selecting a research area one way just work on at least for the start for the start for the first two one or two project start with that so did we start with that you are guaranteed the irony save please save journal of computer science area then how do you build more knowledge once you have chosen that research area conferences conferences are uh indistinguishable you have to attend the conferences now tanks most conferences are now running online so you go with a little font with support from advisor you should be able to attend those conferences online what do you do when you get to the conference you are just doing what you call exploration initially you don’t even know what is what you just go there and just keep jumping from one one session to another keep listening you listen to some talk you see oh this doesn’t make sense to you you listen to some it looks like it makes sense you just keep listening just take your time keep doing that from one conference another be consistent so conferences like no rips their top tier conference icml iclr cbpr acl all this conference irrespective of even if you have chosen a research area saying computer vision just from start go to these conferences and just get to see how they do things and just listen you can’t contribute initially just listen you see okay what people are working on what they are publishing on so excuse me now your knowledge will begin to converge after some time of going to conference you begin to see that there are similar projects that are being taken as maybe workshop or tutorials and conferences or paper in journal of computer science that people are published publishing from posters that are similar to what your advisor are spending on then you begin to see that oh this is a current area of research that people are working on damage is still actively being researched so there is an unsolved problem there okay so i’ll be rounding up now will i be able to solve a research problem well that’s dependent on you we don’t know nobody knows your advisor does not know no one knows whether you’ll be able to solve a an open research program they’re able to contribute so it’s a journey that we don’t know whether you’re going to arrive at the destination however there are signals there are signals that tells the university the admission community as to your viability of success that’s why they have requirements admission requirement there are some examinations you take your performance on those exams will tell you a good candidate and even after you have been admitted journal of computer science you take courses in the university those courses algorithm courses you’re taking your performance and your understanding of those courses and those projects and those exams give signals as to whether you’ll be able to adapt the knowledge you have lent into a noble research area for instance let me take an algorithm like quick sort match sort selection sort yeah or other certain algorithm if you find it difficult to understand this certain algorithm there might be an indication that you may be weak in being able to formulate a step-by-step solution to problem journal of computer science but if you find yourself being able to understand how this algorithm works and you’re able to close your notes and put them down in codes and explain them then it means that you have the kind of brain like the people that originally created those algorithms you have the capacity to understand the problem that the algorithm is solving and capacity to understand the solution that the algorithm is providing so if we do then we can be sure that you have capacity to be able to solve and unsolved problems.
That’s just the best case journal of computer science so as you proceed we have kind of answered that you proceed in an iterative fashion there’s no linear way of going and powering there could be even many more ways that i don’t even know so it’s an iterative fashion initially you keep moving from what your advisor is doing checking conferences checking what other students in the lab in your group are doing small senior student those who are finishing those who are graduating start watching what they have done or what they are doing that begins to give you signals as to what kind of things you two should expect at least something similar may not be exactly then there are presentations seminar presentations in your group in your department what other phd students are doing you listen to them then it begins to make sense to you one or two begins to make sense to you then you may want to explore that a little more in rounding up for this session journal of computer science I would strongly advise that you start solving a problem you start working on a problem that you understand the problem I’m not saying that you understand the solution the first step is do you understand the problem if we sit down in a lecture or in a conference or in a tutorial a workshop and in the first 15 minutes you don’t you’re blank out then that’s not an immediate area of research that you should embark on because you don’t even understand the problem and if you don’t understand the problem there’s no way you can understand the existing solutions and there’s no way you can contribute new approach to solve a meaningful aspect of the problem all right thank you all for listening and see you in the next one.