I am a CS undergrad student from India and I would love to continue with academics/research. But getting into top institutes in India requires high percentiles in competitive exams which are not my cup of tea.
I was in a similar boat as you many years back. Did not get CS at the older IITs, I was too interested in coding and entrepreneurship to be super involved in JEE.
I persisted and got an internship at a top CS place in India. Ended up publishing more research than most PhD students in just a year. And got admitted to all top 5 CS programs I applied to. Was advised by a very well known professor in grad school and even published with one of the grandfathers of machine learning.
I now mentor several undergrads who are going through the same phase. I think just partnering with the right professor helps. Having a vision helps. Most research students just blindly follow whatever the professor tells them to do. Being able to think independently really helps; most of my classmates who ended as Professors still crave for such students. And find it hard to encounter such folks.
To repeat, independence and persistence. You will make it. Good luck!
Publishing research without any affiliation could be difficult because of various reasons beyond your control. Having an advisor or a team can significantly boost your publication quality. Think in terms of how quickly you can validate your ideas, get proper feedback while preparing the publication, and get proper visibility for your work.
Keep in mind that several large tech-companies (e.g., Google, Facebook, Yahoo, IBM, Xerox, etc.) look for research engineers, or software engineers that work with research teams. You could start looking for such jobs and collaborate with teams/team members that publish their work at top journals/conferences. Such teams tend to be highly selective, but that is definitely something different from taking competitive exams. You'll probably have to create a strong portfolio to indicate that you can "independently" conduct research (e.g., showcasing some novel ideas on your website, sharing implementations on GitHub, listing your contributions to open-source projects, etc.).
Another option is to look for any research associate positions with labs/professors.
I'm a CS professor, so I've a fair bit of experience publishing research. A few things to think about:
It's easy to publish research, but the first lesson is that the places it's easy to publish in, no-one will read your research. I was one of the authors of DHTs back in 2001. That paper has clocked up nearly 10000 citations over 16 years. That's a 16 year mean of around 2 citations per day, and that's just one paper! It should be clear that the total number of CS papers published is huge. It's likely that many of those papers will almost never be read. If you want your research to have any influence on people you need to publish in places that are not easy to publish in.
The second lesson is that the hardest part of doing good research is coming up with a good problem. There are three parts to a good problem: it must be solvable with the resources you have, it must not have already been solved well enough, and people must care if you succeed in solving it. I see people fail on all these three, but especially the third. "Make something people want" applies to research as much as to startups. Knowing what will be a good problem is hard without people around you to guide your problem taste, and even then it's hard. Rejection letters from conferences and journals rarely say the biggest problem with your work was that it wasn't interesting; you have to read between the lines if this is what the feedback should have been.
Most people think that the hardest part of research is coming up with a good solution, but usually this ends up being the easiest part. If you find a novel question and really understand it well, solutions tend to present themselves.
The third lesson is that evaluation is hard. Usually harder than coming up with a good solution. I've never yet met a new PhD student who could devise a good experiment plan, and properly interpret the results. It's just not something that is taught properly in schools or at the undergrad level. It's all too easy to evaluate what you can evaluate, not what you should evaluate, or to only run experiments to confirm what you think you already know. You should be looking hard to convince yourself that you're wrong (you often will be), and only when you've failed to convince yourself you're wrong, will you have a convincing evaluation. Good PhD students eventually learn how to do this, mostly by osmosis from their advisor. Some fail to learn this even then. As an independent researcher not surrounded by people asking critical questions, you'll likely struggle on this part.
In my field (social science) you do a PhD in order to learn research methods and familiarize yourself with the literature you are going to relate to in your research. Without this background, it would be virtually impossible to get punished in a big journal.
If you are serious about research, you should do a research Masters / PhD somewhere. What if you join a non-top university? Or can you work on your exam skills? Many universities offer support with that.
"non-top universities" actually publish contents that are not up to the mark in my opinion, however, there are exceptions. Also in many cases, I have seen that mediocre universities of India have a very poor environment for research that actually demotivates the bright candidates and hampers their performance.
EDIT: most universities / institutes in India focus and excel in undergraduate programs.
I can speak of theoretical computer science: if you have work that is worth publishing (as evaluated by your peers), then yes, it is possible to publish without any affiliation. I would be very surprised if a good result is rejected just because the author is not affiliated.
The operative word in the first sentence above is "if". In my opinion, unless you are in the Ramanujan or Erdős class, then it is highly unlikely that you come up with a publishable manuscript on your own. Briefly put: you need a lot of "context" to be able to do publishable work, and you are unlikely to get this context unless you interact with the research community.
The corollary to this: most correct work is not worth publishing. And most people bitter over their work not being accepted were not rejected because their work was incorrect, but because it was not notable. So at the end of the day, it's all a popularity contest, and that's what you're really engaging in.
It is always possible to publish without affiliation, but I'd recommend you try and "get" some affiliation nevertheless. Many universities have interest groups, reading groups, loose projects, etc., that in principle anyone can join, and in most cases all you need is one project leader or professor interested in your publications who allows you to mention the affiliation. For instance, at our institute we have the status of a collaborating member, which includes anyone from the outside who works together with someone from the institute on a project. There are restrictions of our funding agencies of what we can count as a publication of our institute, of course, for that you need to be an integrated member, but that's another issue.
At least that's how it works in many places in Europe, I imagine some places in the US have stricter conditions, especially the top universities. Anyway, contact some people at universities and see if they are interested. That's my advice - assuming that your planned research and publications are of high quality. If not, you're going to have a hard time convincing anyone, of course.
First, please understand that publishing a research paper is not an easy task (at least at reputable venues). I'm writing this mainly because lot of undergrads have this feeling that they are someone special endowed with some magical capacity to envision beautiful ideas that no one else has thought about. Many of them tend to think that they don't need to study prior art, do advanced studies or be upto date with any recent works. It seems to me that some of this thinking in young students have developed thanks to constant onslaught of movies like Harry Potter, Lord of the Ring or even Star Wars where main character is great just because they were born that way and they don't need to put in any work to get there. They are just destined to be great. They are the chosen ones to crack the hardest problems and save the world :). This is then further re-enforced by stories of Faraday and Edison who did not had PhDs and ended up doing great discoveries. The world has changed big time since then!
Typically these days to publish research paper at reputable venue requires lot of work. Count in 10-12 hour per day for 3-6 months per paper + lot of luck + competatnt collaborator(s). You will most definitely need to be intimately familiar with everything that has been going on in the field. Design experiments to prove how you compare to others. If you are doing research, you won't have time for anything else. It's a full time job, not a weekend hobby. If you don't know what were the major paper came out during last 3 months then it means you are not quite ready. If you can't name the 20 people who have published significant work during last 20 years in your field then you are quite possibly not ready. If you haven't been reading at least 1 paper a week (or if you are reading but have no clue what's going on) then you have quite some way to go.
So to summarize, being a researcher is a full time job of reading research papers with some occasional bouts of writing your own papers. In today's complexity levels, it would be very hard to do significant novel work without any collaborations at all (ala Einstein 1905). More importantly you need other eyes to enhance content and/or fix issues. The best way for undergrads to write papers is to get enrolled in Masters and then to PhD. There is almost no escape from this requirement and people who tend to think that there is another way would usually tend to regret it if research is what they really wanted to do all along.
Not to mention that the review cycle can take /years/. I've published a paper that was under various stages of review for a year and a half at JAMIA. And, that's actually kinda short. Be prepared to let go, revisit, refine, redefine, etc for potentially a very long time. Often it's not a deal of submit, revise, accepted. It takes time. A lot of it.
> Not to mention that the review cycle can take /years/.
In journals, yes. In conferences, you get your review within two months or so, and the decision is final (modulo a two-week author response period in some venues).
You can, because reviews in any reputed conference are anonymous.
You will not be able to get travel support funds though. That step assumes you have an academic apparatus behind you.
What are you interested in?
Can you demonstrate expertise in that field?
If you can, just email a professor.
Most of them, at least the young ones, will be happy to collaborate.
If you have made decent publishable research, then sure, publishing without affiliation isn't a problem.
Making decent, publishable research is the difficult part though, and this is where academic institutions are very useful - both in doing the actual research, and also in getting funding for the (many!) weeks you'll need to do this.
Are papers not published by the 'lesser' institutes?
I ask because you state that you can't test at the highest levels and so won't get into a 'top institute' but are you concerned you would not get into any graduate programs?
I ask because while the source institution might get you in front of the review committee more often, it is the quality of the work (in the ideal situation I know) that gets you into the journal. You can also publish papers as an undergrad with your adviser's support in the US, is that not true in India?
> You can also publish papers as an undergrad with your adviser's support in the US, is that not true in India?
It is just as true in India as it is in the US.
And: while it definitely helps if your college/department/advisor is there to lend you support, such support is not a prerequisite for getting published.
Yes as an undergrad I have 3 publications till now (in 3 years) in India. Still other institutes in India won't care how many publications I have when I am applying.
1. Identify research scientists (e.g: in Google/IBM/Microsoft/... research labs) and professors (in institutes/universities) that you would like to work with.
2. Write to them directly, attaching your publications and requesting an internship.
Depending on how your publications are perceived, there is a good chance that at least a few will choose to engage with you.
Indian institutes funded by taxpayer money are required to follow a prescribed set of rules for selecting PhD candidates, in the interest of ensuring fairness, preventing nepotism, etc. . The competitive exams that you seem to loathe are part of this process, and professors can do nothing about this.
They have much more leeway in choosing interns, though. The rest follows as per kubrickslair's recipe.
Publish it on arXiv.org or ECCC (Electronic Colloquium on Computational Complexity). Then, check whatever you have published on arXiv is novel, then submit to a 2nd/3rd tier conference.
That order is no good. You’d want to check what other people have done in that area before you spend weeks figuring out the details of your idea, let alone produce a decent paper and publish on the arxiv. Not to mention that the arxiv is for preprints, not half-baked ideas.
I agree that the order of Step 1 and 2 should be reverted, but how is Step 3 absolutely correct? The tier of the conference you should submit to depends on the content of the paper, not on the author/affiliation.
I don't think anybody's first paper will be accepted at a tier 1 conference. It takes practice to learn how to write a paper, not just in style but also the substance. In academia, your supervisors will typically help/guide you. It's going to be much harder on your own without any experience.
Honestly, the easiest way to get involved is, similarly to open source contributing, I think, is to build upon someone elses work. Find a paper you like and think you can use as a base for something. Then when you've done, publish it and contact the writer of the paper you used as a base. If you're lucky, you'll get some important feedback from him.
There are some very high quality resources out there hosted on WordPress or home-grown websites. Maybe not the same level as prestige as a published journal, but of all subjects, CS is less steeped in credentialism than most.
Do you mean doing academics and research on your own time outside of a job? Is it for fun or do you have some goal to accomplish in mind? I ask because getting published might be relatively easy to do. However whatever you publish may get lost in the million plus papers that get published each year. Perhaps there are more effective ways to let others know about your important ideas?
You can apply to the Ronin Institute [1] which helps people in exactly your situation. They allow you to be affiliated with their "institution" and also provide some support for access to publications.
As with anything , you will need to give it time and learn all the dependencies to be able to do some quality research.
They also have a gitter page where you can talk to like minded people .
As far as competitive exams are considered , i would highly recommend you go through the drill.
1. These are human artifacts of judging your credibility
2. The science you actually learn can be mind expanding .
3. Life is mostly about the people you surround yourself with. You shall be able to meet peers that may have similar interests.
Further, having gone through the education system myself I would congratulate you on the belief that you can achieve anything if you set your mind to it .
If it's love of the art that drives you then nothing really can stop you from doing what you want.
If you want academic prestige, on the other hand, you really need collaborators and mentors who can spar you and introduce you to the community and its culture.
Doing research and getting it published requires you to show how you have built-upon the work of researchers who came before you. In order to get to that point you have to read and understand a huge body of work. Much of that material is behind paywalls which are expensive to access unless you are enrolled as a research student at a university that subscribes to those journals. As a minimum you would need access to ACM, IEEE and Springer-Verlang conference papers and journals.
People without the depth of knowledge often "re-discover" stuff that has been documented a decade or so ago. So if you were to submit such material to any respected journal it will be rejected during the peer review.
As for calling yourself an academic, you would need to be part of some academic institution in order to do that. Passing exams as an undergraduate is the first of many hurdles.
What is missing from these three resources that would prevent anyone from doing a literature survey of Ph.D standard in any research area related to Computer Science?
Sure, and most of them require a huge GATE score. Even if as an undergrad I have several top tier conference publications, IITs won't even consider me without a very very high GATE score.
a) I think you're underestimating how tough it is to actually publish at top tier conferences. Professors at the IITs rarely get more than a few a year.
b) There are MS programs that are typically not as big on GATE score. In fact, at places like IIITH you don't even need a GATE score, you can talk to the professors involved and he/she can accept you as a graduate student.
c) If you did publish at even national level conferences, you can talk to professors from the IITs who will show up there to present their work. They can put in a word with the admissions head, and get you admitted if you really wish it.
Source : I study at one of the older IITs, and have seen it happen where a professor is impressed with students who have interned with him, and ask them to join the graduate program.
facetious answer: yes, due to computational trinitarianism! All you need to do is convert your proof from a logical form into an executable, type-checked and fully verified program - that should get you around the academic clout pre-reqs.. people love to use hobbyist-made software alll the time, but they equally hate to even look at hobbyist research :P /s
Possible but exposure would be a problem that said that is true for most affiliated research also, most papers are only read by their reviewers and authors and given what we know about the peer review process even the reviewer part is in question.
If you have good research the best way to get it published is honestly through talks, register for local or international conventions as a speaker and make a talk out of your paper.
It the talk picks up so will your paper.
On the other hand you can always “pay” to promote your research there are companies that connect popular media outlets with researchers.
I feel it's pretty arrogant to think that you can contribute to research entirely on your own. You really need to be around other researchers who are smarter and more experienced than you to share ideas with and learn from, you simply do not have the insight needed to make progress in your field on your own. If you were that wunderkind who could do it, you honestly wouldn't be asking.
If you truly want to get into research, grad school should sound like a dream (at any school with a decent program, not just top schools). You'd be surrounded by likeminded people, get to do crazy science stuff all day, and you'd push the frontier of knowledge a little bit. It's really the best and most effective way to immerse yourself in the material and become a good researcher.
This is not entirely true. Speaking from my rather long sojourn with CS Research in an industrial research lab setting. While it helps to start out with researchers, many fields of CS research are as much individual pursuits as they are a team effort. If you have a network of domain experts you can communicate with, and access to publications, a sufficiently motivated person can definitely make inroads into research.
A couple of other points to be noted,
1. open access to publications is still a utopian dream. Being affiliated with an organization or academic institution gives you access.
2. Infrastructure and Equipment is something you will have to arrange yourself if you want to do things on your own. This is prohibitive when it comes to the latest and greatest in deep learning, or systems engineering or anything biomedical.
3. Data. Organizations have access to data. period. Or atleast can arrange access to data.
4. It helps to have more hands on deck. So working in teams can accelerate the pace of your results.
Some fields which are conducive to lone work would be around cryptography, algorithms, essentially anything theoretical CS. or fields like ICTD where its less about a break through and more about a novel application.
> If you truly want to get into research, grad school should sound like a dream (at any school with a decent program, not just top schools). You'd be surrounded by likeminded people, get to do crazy science stuff all day, and you'd push the frontier of knowledge a little bit. It's really the best and most effective way to immerse yourself in the material and become a good researcher.
Are you speaking from experience? It honestly sounded like a dream to me but many who have gone have said that many parts of it really suck because you have to do a lot of things besides research.
Grad students usually teach so that they are paid, but you're not required to. Classes are required in my program, but only a handful. Your mileage may vary by university.
I think the only source of potential arrogance is unfamiliarity with prior and ongoing work. My experience is that people often comes up, on their own, with brilliant ideas that have already been explored. A lot of my uni time was a roller coaster of "wow, I have this amazing idea" (talk to someone in the field), "bummer, they beat me to it by random (1,200) years".
Of course being around smart people helps think out, crystalize and polish ideas. But many people do have insights on their own, some completely new even in 200 year old fields.
Plenty of people did exactly that before science became a professionalized endeavor. Sure, being in a community of practice, having experience guidance, etc would help a lot -- but by no means are they necessary conditions to do good research.
I persisted and got an internship at a top CS place in India. Ended up publishing more research than most PhD students in just a year. And got admitted to all top 5 CS programs I applied to. Was advised by a very well known professor in grad school and even published with one of the grandfathers of machine learning.
I now mentor several undergrads who are going through the same phase. I think just partnering with the right professor helps. Having a vision helps. Most research students just blindly follow whatever the professor tells them to do. Being able to think independently really helps; most of my classmates who ended as Professors still crave for such students. And find it hard to encounter such folks.
To repeat, independence and persistence. You will make it. Good luck!