With how promising Evolution Strategies is as an RL alternative, we just completed our own research to compare and evaluate it against GRPO! Our major finding lines up with the ES paper: Evolution Strategies can beat GRPO even when you have only a little fine-tuning data.
@askalphaxiv
-
LeWorldModel: JEPA Training Breakthrough by LeCun
By
–
Yann LeCun and his team can't stop cooking "LeWorldModel: Stable End-to-End Joint-Embedding Predictive Architecture from Pixels" One of the biggest bottlenecks of JEPA is they are hard to train, and this new research changes that. They propose LeWorldModel, which shows that a
-
GradMem: Gradient Descent Memory for Context Compression
By
–
"GradMem: Learning to Write Context into Memory with Test-Time Gradient Descent" Instead of giving LLMs huge KV caches or one-shot compressed summaries, GradMem shows that you can let a frozen model take some test-time gradient steps to write a long context into a small memory.
-
Exclusive Self-Attention: Two-Line Transformer Improvement
By
–
"Exclusive Self Attention" This paper proposed Exclusive Self-Attention (XSA), which is a tiny two-line change that stops attention from looking at itself. This forces it to focus on the rest of the sequence, and can make transformers more effective! This improves the
-
Schrödinger Bridges: Unified Foundation for Generative Models
By
–
"Foundations of Schrödinger Bridges for Generative Modeling" This paper shows that diffusion models, score-based models, and flow matching are really just different views of the same core idea: a Schrödinger bridge that moves noise into data along the most efficient stochastic
-
AI Learning Scientific Taste Through Community Feedback
By
–
scientific “taste” isn’t some unique human instinct "AI Can Learn Scientific Taste" This paper shows that if you train on community feedback like citations, AI can learn to judge and generate research ideas with higher long-term impact. This moves AI scientists beyond just