Eran Ben Artzy·Sep 9“PoT-Diffusion”: General Solution For CSP NP-HARD based ProblemsHappy to share my new research paper called “PoT-Diffusion: Outer and Inner Thinking Cycles” ⚡A response icon1A response icon1
Abu Huzaifah Bin Haji Bidin·Sep 6Energy Based Models for AI (Part 4): Let’s Implement the SamplerAlright, sorry for taking a loooong break between articles. I’m busy writing my thesis and my research journal paper. I’m at the last phase…A response icon1A response icon1
Eleventh Hour Enthusiast·Aug 23An Energy Correction for Diffusion Language ModelsMasked diffusion predicts tokens independently, and an energy on the full sequence adds back what that misses
Eleventh Hour Enthusiast·Aug 21The Optimal Aligned Policy as an Energy-Based ModelDetailed balance, a condition from statistical physics, describes the equilibrium structure of an idealized model’s answers
Eleventh Hour Enthusiast·Aug 20When Latent Optimization and the Decoder DisagreeWhy lower energy need not mean a better answer
Eleventh Hour Enthusiast·Aug 20Attention Is an OptimizerThe weights produced by softmax in standard attention have the form of a distribution from physics, and that fact organizes several…
Eleventh Hour Enthusiast·Aug 17Spilled Energy in Large Language ModelsHow the LLM contradicts itself, and why that contradiction is measurable
Eleventh Hour Enthusiast·Aug 15Autoregressive Language Models are Secretly Energy-Based ModelsWhat the bijection proves, and what it asks the architecture to carry
J.L·Jul 15HarmonicEBM: A Trigonometric Energy-Based Model with Analytically Tractable Partition FunctionThe accompanying code for this article is available at https://github.com/fooooola/harmonicebm
Eleventh Hour Enthusiast·May 31Energy-Based TransformersWhat the architecture inherits from classical energy models, and where that leaves it
