Intelligence Representation & Reasoning Lab

Gwangju Institute of Science and Technology (GIST)

How to build a single accumulative intelligence
for more general and wider problems?

Main page img modified at Jun 23, 2021


Join Us!

We are always looking for strongly self-motivated students for Ph.D., and Postdoctoral Program, who want to dive into research for deep and fundamental understanding of AI. If you want to join in, please send an email (kangil dot kim dot 01 at gmail.com).

(Intership for GIST students) please watch this video for internship, and send me an email.


News

24.11.06: Congratulation! accepted paper by TMLR, “CFASL: Composite Factor-Aligned Symmetry Learning for Disentanglement in Variational AutoEncoder”
24.05.16: Congratulations! accepted ACL Findings 2024 “Structural Optimization Ambiguity and Simplicity Bias in Unsupervised Neural Grammar Induction”


Recent publications

Background image
TMLR, "CFASL: Composite Factor-Aligned Symmetry Learning for Disentanglement in Variational AutoEncoder"
November 06, 2024
Symmetries of input and latent vectors have provided valuable insights for disentanglement learning in VAEs. However, only a few works...
Background image
ACL 2024 Findings, "Structural Optimization Ambiguity and Simplicity Bias in Unsupervised Neural Grammar Induction"
August 11, 2024
Video will be uploaded after conference! Summary & Contribution We reveal two problem “Structural Optimization Ambiguity” and “Structural Simplicity Bias”...
Background image
ICLR 2024, "Fixed Non-negative Orthogonal Classifier: Inducing Zero-mean Neural Collapse with Feature Dimension Separation"
January 20, 2024
FNO classifier makes the LPM achieve the global optimality even in inducing the max-margin decision while satisfying the properties of...