5 papers
LiLiCorr: Lightweight Likelihood Correlation of Parallel Drafts for Speculative Decoding
Matan Rusanovsky, Yoav Miron, Roy Uziel +4
Speculative decoding accelerates language-model inference by drafting future tokens that the target model verifies in parallel. A diffusion-style block head such as DFlash is an at…
Learn from Your Mistakes: Self-Correcting Masked Diffusion Models
Yair Schiff, Omer Belhasin, Roy Uziel +6
Masked diffusion models (MDMs) have emerged as a promising alternative to autoregressive models, enabling parallel token generation while achieving competitive performance. Despite…
CRoCoDiL: Continuous and Robust Conditioned Diffusion for Language
Roy Uziel, Omer Belhasin, Itay Levy +4
Masked Diffusion Models (MDMs) provide an efficient non-causal alternative to autoregressive generation but often struggle with token dependencies and semantic incoherence due to t…
Clustering via Self-Supervised Diffusion
Roy Uziel, Irit Chelly, Oren Freifeld +1
Diffusion models, widely recognized for their success in generative tasks, have not yet been applied to clustering. We introduce Clustering via Diffusion (CLUDI), a self-supervised…
Consistent Amortized Clustering via Generative Flow Networks
Irit Chelly, Roy Uziel, Oren Freifeld +1
Neural models for amortized probabilistic clustering yield samples of cluster labels given a set-structured input, while avoiding lengthy Markov chain runs and the need for explici…