📁 Papers

Section folder view.

Subfolders

World Models Reading List

Notes

A Compression Perspective on Simplicity Bias

Beyond Neural Scaling Law - Data Pruning improve power law

Boosting Data-Driven Mirror Descent with Randomization,Equivariance, and Acceleration

Cram Less to Fit More

Crowd-Aware Robot Navigation with Switching BetweenLearning-Based and Rule-Based Methods Using Normalizing Flows

Deep Learning on a Data Diet

Diffusion-LM

Information Theory for Repr Learning

Invariant Risk Minimization

Large-Scale Pruning using Dyna Uncertainty

Learned feature representations are biased by complexity, learning order, position, and more

Learning Stable Normalizing-Flow Control for Robotic Manipulation

Learning to Theorise

Learning2Optimize

MACO-Feature Visualization via Magnitude Constrained Optimization

MaxEnRL via Energy-based Normalising Flows 1

Mechanistic Mode Connectivity

Neural Ordinary Differential Equations

RecoveryChaining-Learning Local Recovery Policies for Robust Manipulation

RESPO (Reachability Estimation for Safe Policy Optimization)

Scaling for Data Filtering

TIC-CLIP - continual training of clip models

Understanding BlackBox Predictions - Using Influence Functions

Understanding Visual Feature Reliance through the Lens of Complexity

When Do Curricula Work?