close
Log inSign up
user avatar

Fei Sha

@feishaAI
Researcher and engineers in AI and ML; now at Google Research, formerly Professor at USC
Cambridge, MA
feisha.org
Joined February 2009
332
Following
1,856
Followers
RepliesMedia
  • user avatar
    Fei Sha
    @feishaAI
    Aug 10
    I am waiting for the moment when a model declares “I have created a research problem that my fellow models cannot solve in O(X)-GW and I have solved it!”
    user avatar
    Andrea Montanari
    @Andrea__M
    Aug 9
    Very honestly, the current status of research in this area is more sophisticated than this result. This way of communicating science can make serious and long lasting damage. [This is not to say that the result isn’t new or that AI models aren’t amazing.]
  • user avatar
    Fei Sha
    @feishaAI
    Jun 1
    I tried this or similar a few times; I have given up. I find the generated surveys on the borderline of intellectual drivels. It is like a poorly written “Background” in a paper, except 100x times bigger. It generally feels like a fluid pile of “abstractive summary”.
    user avatar
    (((ل()(ل() 'yoav))))👾
    @yoavgo
    May 30
    Replying to @yoavgo
    - my long term goal is to map the field of 'X'. it should be very comprehensive. at some point i would like to write a survey. can you help me out with the literature search? - [long and kinda useful introduction to X, mapping it to subtopics and trends, could be useful if i
  • user avatar
    Fei Sha
    @feishaAI
    Mar 17, 2025
    Vladimir Vapnik once said "Nothing is more practical than a good theory." He also introduced the concept of privileged information in his book The Nature of Statistical Learning Theory. Glad to see it inspired this work. A great pleasure working with you, @JinPZhou !
    user avatar
    Jin Zhou
    @JinPZhou
    Mar 17, 2025
    LLM-as-a-Judge is very popular for automatic LLM evaluation. A natural question is: How can we trust LLM to grade themselves on tasks they don’t master yet? Excited to share our recent work where we let graders "cheat" by using privileged information arxiv.org/abs/2502.10961 🧵👇
    BERJAYA