Goutham Rajendran
9 papers · 2021–2024 · 3 conferences · across top CS/AI conferences
Achievements
Jump to papers ↓+2 more ↓ Show less ↑
๐ Cross-Pollinator (13) ๐ Interdisciplinary Bridge ๐ Conference Polyglot (3) ๐งญ Keyword Pioneer ๐ Renaissance Researcher (5)
๐บ๏ธ
Taxonomy Completionist
(18)
โ
The Questioner
Conferences
NIPS (7)
CLEAR (1)
ICML (1)
Top co-authors
Keywords
graphical model
(2)
causal representation
(2)
latent variable
(2)
causal inference
(2)
independent component analysis
(1)
structure learning
(1)
transformer architecture
(1)
dimension reduction
(1)
contrastive learning
(1)
sparse principal component analysis
(1)
bregman divergence
(1)
system identification
(1)
exponential family
(1)
model identifiability
(1)
generative model
(1)
computational complexity
(1)
algorithm complexity
(1)
latent variable model
(1)
associative memory
(1)
causal discovery
(1)
Papers
Do LLMs dream of elephants (when told not to)? Latent concept association and associative memory in transformers
NIPS 2024
On the Origins of Linear Representations in Large Language Models
ICML 2024
An Interventional Perspective on Identifiability in Gaussian LTI Systems with Independent Component Analysis
CLEAR 2024
From Causal to Concept-Based Representation Learning
NIPS 2024
Learning Linear Causal Representations from Interventions under General Nonlinear Mixing
NIPS 2023
Sub-exponential time Sum-of-Squares lower bounds for Principal Components Analysis
NIPS 2022
Identifiability of deep generative models without auxiliary information
NIPS 2022
Learning latent causal graphs via mixture oracles
NIPS 2021
Structure learning in polynomial time: Greedy algorithms, Bregman information, and exponential families
NIPS 2021