Papers

Papers with Notes:

• Intention Conditioned Value Functions (ICVF)
• Neural Acceleration of Graph Based Utility Functions for Sparse Matrices
• Neural Acceleration for Incomplete Factorization Preconditioning
• Grokking - Generalization Beyond Overfitting on Small Algorithmic Datasets
• Grokfast - Accelerated Grokking by Ampilfying Slow Gradients
• Evaluation of Pre-training LLMs on Leadership Class Supercomputers
• Every Model Learned by Gradient Descent Is Approximately a Kernel Machine
• Converting Neural Networks to Symbolic Models
• Iterative alpha-(de)blending
• Learning to Reach Goals via Iterated Supervised Learning
• Goal-Conditioned Supervised Learning
• Reward Conditioned Policies
• Spectral Algorithms
• What if Neural Networks had SVDs?
• The Surprising Effectiveness of PPO in Cooperative Multi-Agent Games
• Theories of Artificial Intelligence
• HPAC & HPAC-ML - Presentation
• Different Conceptions of Learning
• Assumptions of Decision Making Models in AGI
• YuNet (Face Detector)
• K-SVD - Sparse Representation using Overcomplete Dictionaries
• HRL with Timed Subgoals
• Gradient Projection for Sparse Reconstruction
• Generative Modeling by Estimating Gradients of the Data Distribution
• Gradient Estimation on Quantum Hardware
• SO3 Action Representation for Deep RL
• Gradients Explode - Residual Explained
• Hierarchical Planning with Latent World Models

Minimal notes:

• Patches Are All You Need?
• Solving Offline Reinforcement Learning with Decision Tree Regression
• Early Stopping of Episodes improves RL
• Horizon Reduction is needed to scale RL
• Encoder can improve RL performance without Algorithmic changes
• Agents are World Model
• Divide and Conquer Reinforcement Learning
• Synergistic Action Representation
• SimBa - RL
• Embedding Logical Queries on Knowledge Graphs

Other articles:

• Principles of Beautiful Figures for Research Papers
• How To Find, Read, And Master Research Papers