Jump to content

moolriaz

Members
  • Posts

    19
  • Joined

  • Last visited

Everything posted by moolriaz

  1. It will of course be helpful. If you have 10 years of experience in programming, you're not going to have a problem implementing Matlab or LISP scripts. Research is going to be more about coming up with the algorithms, and implementing them is usually not the hard part. This isn't always true though because some problems require very efficient implementations to work with big data, and may require using clusters, servers, distributed algorithms, or rewriting core functions in C++. If you want to focus on something, focus on learning statistics, mathematics. Take a MOOC course on machine learning, as many as you can. Then decide which subarea of AI you'd be interested in, there are few MOOCS and course resources on things like NLP, vision, bioinformatics..
  2. Hi, I would recommend the masters programmes offered by the University of Edinburgh: http://www.ed.ac.uk/schools-departments/informatics/postgraduate/msc/msc-ai Under their AI degrees you can specialise in Machine Learning, Natural Language Processing, Robotics, or classical AI. There are enough courses offered in just one of these areas to constitute an entire masters (for example, there are 6 courses just on ML, 5 just on NLP etc). I have not seen a masters programme that is as comprehensive. The fees are also much more reasonable than what you may have to pay in the USA. From my study in AI so far all I can say that statistics and machine learning are ubiquitous. The more mathematics and statistics you learn, the better.
  3. Please join https://www.facebook.com/groups/570324409719397/ this Facebook group.
  4. Yes, I applied and was accepted with scholarship. We have a FB group but there are very few students. You may find posting on thestudentroom is more useful for this.
  5. Grats guys! Rejected here, so look's like it's CMU for me.
  6. Go to NYU. I was at Penn CIS doing CS and in my own personal opinion I did not think the CS department lived up to the prestige of Penn at all. That is really down to Wharton. The CS department and courses offered seemed small and limited, and the facilities were not sufficient at all (often had to stand outside a lecture room just to listen to a lecture, insufficient space in the labs to do work). The students were not friendly, there were some good professors though. I sat in on some of the MCIT courses. It did not seem like there were many tech opportunities either. I was very surprised given the prestige. Penn would be a better choice if you wanted to move there for research in a specific field, like NLP which Penn has good faculty in.
  7. I'm currently at Edinburgh and know people in this programme and have taken courses with all of the advisors in ML @ Edinburgh ie Amos Storkey, Chris Williams, Iain Murray, Charles Sutton and a few more. Edinburgh has a very good rep in the US, when I interviewed at CMU etc and did exchanges abroad at some good US universities I have always had favourable comments about Edinburgh Informatics as it is if not the largest then one of the top three CS/ML/AI research departments in Europe. ICML was at Edinburgh in 2012. Informatics @ Edinburgh has more faculty than Oxford and Cambridge combined and high quality research in ML. I would not consider any of the universities you have offers from other than CMU, and certainly not University of Michigan. Many of the staff in Informatics carry out research with their US peers - Stanford, CMU, MIT - I would say more so than any other UK university, so I don't think you should worry about that factor at all.
  8. Apply to the top 4. I got into one of the top four with no pubs, lower GRE than you. It's all about the references I think.
  9. I think if you don't receive acceptance by the end of the today from CMU LTI you have been rejected, as they said all acceptance letters would be sent out by then, and have already mailed us info about visiting etc.
  10. I just got an invite to a Google Group for accepted applicatants to the CMU LTI. It is from a cmu email address and looks legit. However, no formal acceptance email. Anyone else get this - maybe its a mistake???
  11. I dont think MLT Masters decisions are out yet. Though I was told the admissions board meeting was last Friday, so not sure why it isn't out yet.
  12. I don't think HCI CMU went out, if you check the gradcafe tracker it looks like LTI PhD, ML PhD and Robotics PhD have gone out. Waiting on MLT here
  13. ? The UK has some great masters programs. Unlike (the majority) of US masters programmes, they can be used as a stepping stone into a PhD in the UK or US, at a fraction of the cost. Part 3 in maths at Cambridge is very popular for those looking to go into a PhD afterwards, the same in the advanced computer science MPhil. There are many US students doing masters in the UK, I work with them every day. Just make sure you go to a top 5 UK university - eg for Computer Science Cambridge, UCL, Oxford, Edinburgh, Imperial. I think the masters degrees in the UK are well regarded in the US from respected universities, indeed I was advised by my profs in the US to do a masters in the UK and the US masters are often 'looked down' on as many are just money generating programmes and for those who did not get into PhD programmes. In the UK a masters year is regarded as a stepping stone to a PhD, and in some cases in required. And to the previous poster, coming from Bulgaria you should know universities such as Edinburgh are *free* for EU citizens, who would normally be considered internationals in England.
  14. Owego, I agree with the above, but I would add the comment to choose Chicago over Penn if you had the choice. I studied at Penn in the CIS department and it wasn't a particularly big or active part of Penn (facilities were also surprisingly poor - often had to sit on the floor). There are many state and private universities with more active/well known CS departments than Penn, and I'd say Chicago is one of them.
  15. Hey guys just to let you know I rewrote with any study at all and got a higher score (161 quant) which I am very happy with. This time I got 3 verbal sequences and 2 quant sections and I found it much easier overall. I would advise to not overstudy for the GRE as there is a big luck factor, especially with the 3/2 section split. It's always harder if you get 3 of the section you struggle with.
  16. Yes, I am taking it again next Thursday. I really have a lot of university work though and don't feel like putting that time into studying again. It really didn't pay off last time! I think I'm just going to try rushing through the second quants section this time and see if it pays of... Which would you recommend are the best 'hard' practice questions? I have the Manhattan GRE books and these have hard sections at the back. I really don't feel like paying $80 for Magoosh...
  17. Thanks for the advice. I do get very stressed under time and always have. I don't know if it is possible to correct this through studying or practice, as I have felt this way for my entire life. In a similar way the reason I do well AW section is just because I naturally enjoy writing and reading and have always had above average competency in it for most of my life. I would normally get marked down for writing very long essays/responses in most contexts, but in the GRE it is rewarded. I never did any preparation for AW or Verbal. In AW I start writing and do not stop until the 30 mins are up, they definitely reward longer essays. I also do not do any kind of plan other than breaking similar thoughts up with paragraphs, and having some kind of intro and conclusion. You need to be creative and give examples for your position and waffle on and make as much use as possible of fancy language. In most contexts this isn't something good, but in the GRE they seem to love it, it's a silly exam really.
  18. I wrote the GRE a month ago, and yesterday. When I wrote a month ago I had to fly up to another city and got about 4 hours of sleep before the test. I was exhausted and scored 158 V 158 Q, 5 AW. I didn't take the test too seriously because I had already planned to write it again. I wrote again yesterday, after studying quite a lot for a week. I made sure I was well rested etc, but ended up with 160 V 156 Q 5 AQ! Totally embarrassed that my Quant score actually dropped and not sure what is going on. That quant score is less than the quant score I got in the 7 practice tests I took prior to the test (all in the range (158-164). I realize that studying isn't really a problem, it's more about time management. The thing is every time I wrote I got 3 quant sections. And both times the pattern was V Q <break> Q V Q. The last quant section is always the hardest, and I'm usually exhausted by then and don't manage to finish that section (always have to guess the last 4 questions or so). Is there any chance that if I rewrite I will get 3 verbal sections rather than quant? I'm much better at verbal and it isn't important for my major (Computer Science). I feel I'd do a lot better if I just got two quant. Also, how can I improve my score on the second section? I practiced timing, and 'moving on' after 1 minute, but I think when I get to that last section I either freeze up or slip up and forget about times. I've always had a problem working under time, but this is the worst so far. Can anyone recommend a way to do better on the final quant section? I'm definitely getting the 'hard' section and I finish section 1 with a few minutes to spare. Also I really want to say GREs aren't that important. I'm currently a finalist for Rhodes and I already have Fulbright (from outside the US to study inside the US). In my overall application, it looks like the GRE would be my weakest point. Do you think a 158 Quant would be enough to nullify all the other aspects of my application? My undergrad is a first class from a top 5 UK university. I'm applying for masters in the US but hoping for top places.
×
×
  • Create New...

Important Information

This website uses cookies to ensure you get the best experience on our website. See our Privacy Policy and Terms of Use