I have done research on a topic that is very sensitive in my original country. I want to submit a paper to a (well-rated) Computer Science conferences, but I do not want to use my real name or university at all (neither for submitting nor publishing). Have you ever heard of such thing? How doable is it?
This is a highly sensitive topic, see Brian Borchers' comment:
As a practical matter, even if the publishers of the conference proceeding were willing to keep your identity secret it would probably be relatively easy for officials from your country to determine your identity. It's likely that the publisher will want to avoid having any responsibility for keeping your identity private.
My advice: Talk to people from your university what to do. Ask trusted colleagues (which work on the same topic), university lawyers etc.
This question is too complicated to be handled by you on your own.
You can write any name, and no one will check your ID. The reality is that you can submit a paper to any scientific journal under a fake name, and no one will ever check whether your name is real. Also, I do not remember having to show my ID to any conference organizers.
If you want to be able to later claim the authorship (e.g., after you become a citizen of a different country and totally independent of your country of origin), you have to collect evidence and retain the email box you will use for submission of the paper. You can take a video of the process of submitting the paper by using a program like Free Screen Video Recorder.
Consider choosing a name that closely resembles your real name, but is different enough to eliminate the risk of getting into legal trouble in your original country. You know, non-English names can be transliterated to English in different ways, and you can also shorten your first name. For example, if your real name is Aлександр Сидоров, which is normally transliterated as Alexander Sidorov, you can consider writing it as Al Seadorough. If someone asks you in your original country whether it is you, you can plausibly deny. Later, when you feel safe to claim the authorship, you will be able to simply include the paper in your CV and, if needed, provide explanations and evidence (e.g., the ownership of the email address indicated in the paper, the video of submitting the paper, etc.). Also, if you use just an altered version of your name instead of a totally fake name, you can hardly be accused of misconduct, especially given the circumstances.
If you want to ensure that the publisher won't be able to track down your university and city, you can use a proxy server to submit your paper. Otherwise the publisher may learn your real IP address.
You are not required to write an institutional address. You can just write any street address, thereby claiming to be unemployed by any university. I saw papers with such addresses. Just ensure that the address chosen by you won't cause any suspicions. For example, you can provide a valid street address of a certain apartment complex, but choose an apartment number that does not exist in that complex.
Submit your paper to a conference where you can and will pay the publication charge in cash, or to a journal that does not charge the authors. Otherwise you may be unable to find a way to pay without exposing your real name.
If your research is so sensitive that the authorities in your country of origin will do their best and utmost to find the author, then you should be especially careful and take additional security measures. First, it goes without saying that you should not go to any conference in person with your research. Publish your research in a journal instead. Second, it goes without saying that you should publish your article under a totally fake name, which does not resemble your real name in any way. Third, never store your research data in your country of origin in the unencrypted form. Use a program like TrueCrypt to encrypt data. (A comment below suggests that Veracrypt is a good choice and better than TrueCrypt.) Leave no evidence on your computers in your original country, because they may be searched if police come to you. Fourth, do not submit your paper from your original country using your home Internet connection. If you currently live in your country of origin, you can submit your paper from a public computer in an Internet cafe, additionally using a foreign proxy. Of course, this applies not only to submission of the paper, but also to registering and using the email box, returning proof corrections, and everything related to all this. You have to leave no trace leading to you.
Others have mentioned that in most cases there's no real issue with submitting under a fake identification. As user111955 points out, the technical details of how to handle this are complex and you should certainly seek help with this. But, while it's pretty much impossible for any individual (even security experts) to know everything they need to know about securing themselves against their attackers, you're also going to have learn enough about this yourself that you can spot the kinds of things that can be security issues for you, i.e., you need a good "security mindset."
Here are a few core issues that it's really important to keep in mind when implementing security measures.
Security is not an "is" or "isn't" thing: something is not either "secure" or "not secure." It's always a trade-off between the cost to you (most often in terms of convenience for you and others involved) and the cost to the attacker of overcoming those security measures, which is also a measure of the amount of risk you're taking on.
The largest component of security is not the technical means (such as particular encryption programs or protocols) but how these technical means are used. The best tools in the world provide no additional security at all if they're used poorly. Anybody helping you out with this should show great concern about training you to work in more secure ways.
Keep in mind is how security affects disaster recovery. As an extreme example, keeping all your work on a single encrypted drive can result in total loss of the data if the drive fails or you forget the passphrase. Always consider the various scenarios that could result in loss of access (for both you and others) and how these can be mitigated.
For doing this, you can find plenty of good resources outside of academia, as well as within it.
I would suggest starting with reading through the Electronic Frontier Foundation's Surveillance Self-defense pages, which will give you a better sense of the problems, techniques for dealing with them, and tools that can help with this. In particular, the Your Security Plan page gives an overview of what you need to be considering.
For specific questions and answers about various security tools and measures, the Information Security Stack Exchange is very helpful, and you can use the same SE account there that you're already using for this question.
Then see if you can find a professional in the field who can help you out. Especially given that your research sounds as if it may have significant value to the general public, a reporter who deals with stories similar to what you're researching and whose sources require similar kinds of security could be very helpful. I won't give any specific suggestions here, but reading through the news for stories covering topics similar to your research will likely lead to reporters who specialize in your area. Find a reporter or publication that offers higher-security methods of contacting them. (If they publicly and widely announce their PGP key fingerprint or SecureDrop information, that's a good sign that they do this and take it seriously). A good reporter will not only offer the technical means for secure communication, but explain how and why their systems work, and provide training on proper use of these systems.
Others already have mentioned the importance of opsec in this case and I have a very harsh thing to say about that: you will mess it up. I only know from what I read that drug dealers mess up, but I know first hand software pirates do. I was one for twenty years, after all, and while I wasn't high level enough to get arrested but I saw people got busted despite they were super careful.
Maybe you leave one copy of your incriminating data on an USB drive. Maybe you have scribbled some parts of it on a notepad and they can (easily) find out because the pen pressed into sheets below the actual drawing. Given enough resources these days they can read palimpsests, something as simple as this is no problem. Maybe you write in a similar style and you get matched by stylometry. The fundamental problem here is the classic assymetry of security: the defender (you) need to make every single step airtight, the attacker (your original country) only needs to have one lucky break.
You need to weigh the consequences as to what happens to you and potentially your family when you are found out. Not if -- when. Because if a country gets pissed at you, they will.
A small sidenote that hasn't been stressed enough (though Sandra's answer alludes to this):
Submitting the paper is one thing, but do keep in mind that unless you can find a coauthor, you'll have to attend the conference in person.
There may be other people from your original country in attendance and some of them might even recognize you. Perhaps they wouldn't report you on their own, but if somebody asks them, all bets are off.
There's also the risk of photographs - people sometimes take pictures of interesting talks or posters and many conferences also keep public galleries from past events on their websites. If your research really is that sensitive and someone suspects that you (or just someone from your country in general) might be the author, they may go looking for these. All it could then take to land you in trouble is one photo showing your face. And what's worse, this could happen years after the fact.
In short, if you do decide to publish your research, it may be better to either choose a journal rather than a conference, or to have someone else present the paper in your stead.