@Hanlin

deep learning • computer vision • artificial intelligence

hanlingoh.com
Joined February 2008
  • Tweets15
  • Following88
  • Followers156
1Media
[1/2] #ICLR2024 Vision-language models still struggle with open-domain visual concepts, while finetuning without proper regularization tends to overfit. We introduce OGEN to improve OOD GENeralization of finetuned models by synthesizing OOD features and adaptive self-distillation
1
3
8
1,115
Our paper on uncertainty estimation for offline RL by our Intern Yue Wu (yuewu.ml) has recently been accepted to #ICML2021. It's a simple, yet intuitive method that results in performance gains and training stability. Check it out!
This Post is from an account that no longer exists. Learn more
New work on Geometric Capsules: Learning to group 3D points into parts & parts into the whole object in unsupervised way. Each capsule represents a visual entity consisting of a pose & feature representing "where" & ''what'' it is. arxiv.org/abs/1912.03310 w/t @nitishsr & @Hanlin
24
142
Check out our latest work on modeling the "what's" and "where's" of objects and parts with geometric capsule representations from 3D point clouds. arxiv.org/abs/1912.03310 (with @nitishsr & @rsalakhu)... I'm at #NeurIPS2019 all week if you want to chat about it!
8
1
29
The roars of #TheRSAF #BlackKnights... Happy Birthday #Singapore! #SG50 – at Singapore
1
A life is like a garden. Perfect moments can be had, but not preserved, except in memory. LLAP
8,385
189,764
1,178
162,172
Friday the 13th Honey Moon Over Paris Photo by Hanlin Goh — @NatGeo #YourShot on.natgeo.com/1nCdDRn
@JWvanWessel @Lytro has already changed photography...
@lytro a day: 3/365 - Discover what's hidden in the rains in today's living picture --> lyt.ro/wKIopz
Load more