3 papers
cs.LG2026
Revisiting scaling laws for reward optimization
Ali Aouad, Aymane El Gadarri, Vivek F. Farias
Scaling laws for optimization against reward models in AI alignment have pinned down how performance depends on optimization effort---measured by a KL-divergence budget relative to…
cs.AI2026
OS-Pruner: Pruning Chains-of-Thought of Reasoning Models via Optimal Stopping
Mohammed Ehab, Aymane El Gadarri, Vivek F. Farias +2
Large Language Models (LLMs) have achieved remarkable success in complex reasoning tasks through Chain-of-Thought (CoT) prompting. However, these models often exhibit "computationa…
cs.LG2026
The Value of Covariance Matching in Gaussian DDPMs and the Lanczos Sampler
Md Sahil Akhtar, Aymane El Gadarri, Vivek F. Farias +1
A central error measure in Gaussian DDPMs is the path-space KL divergence between the exact reverse chain and the learned Gaussian reverse process. This quantity is especially rele…