Omid Saremi
6 papers · 2024–2024 · 2 venues · across top CS/AI venues
Authors are matched by name, so researchers who share a name ("Yang Liu", "Wei Wang") appear as a single profile and their paper counts, venues and achievements are combined. How identity is handled.
Achievements
Jump to papers ↓+1 more ↓ Show less ↑
🌍 Conference Polyglot · Bronze 🌉 Interdisciplinary Bridge · Bronze 🗺️ Taxonomy Completionist (17) 🐣 Hot Topic Early Bird ❓ The Questioner (3)
⚡
Prolific Year · Productive
Venues
Top co-authors
Research topics
Machine Learning
(5)
Artificial Intelligence
(4)
Core AI
(4)
Deep Learning
(3)
Learning Types · Machine Learning
(3)
Optimization & Theory · Machine Learning
(3)
Architectures
(2)
Core Methods
(2)
Interdisciplinary
(1)
Learning Paradigms
(1)
Learning Types · Deep Learning
(1)
Linguistics
(1)
Optimization & Theory · Deep Learning
(1)
Keywords
out-of-distribution generalization
(2)
self-supervised learning
(2)
feature learning
(1)
policy gradient
(1)
in-context learning
(1)
compositional generalization
(1)
relational reasoning
(1)
gradient descent
(1)
linear discriminant analysis
(1)
reward function
(1)
implicit bia
(1)
algorithmic reasoning
(1)
data efficiency
(1)
linear probing
(1)
length generalization
(1)
masked autoencoder
(1)
joint embedding predictive architecture
(1)
latent space prediction
(1)
vanishing gradient
(1)
joint embedding
(1)
Papers
How Far Can Transformers Reason? The Globality Barrier and Inductive Scratchpad
NeurIPS 2024
How JEPA Avoids Noisy Features: The Implicit Bias of Deep Linear Self Distillation Networks
NeurIPS 2024
LiDAR: Sensing Linear Probing Performance in Joint Embedding SSL Architectures
ICLR 2024
Vanishing Gradients in Reinforcement Finetuning of Language Models
ICLR 2024
What Algorithms can Transformers Learn? A Study in Length Generalization
ICLR 2024
When can transformers reason with abstract symbols?
ICLR 2024