HC-DLM Couples Discrete Tokens With Continuous Latent Trajectories, Beating Diffusion Baselines on Sudoku and Countdown
UIUC-CS · hf · 2026-10-02
Hierarchical Continuous Diffusion Language Models (HC-DLM) address structural bottlenecks in both discrete diffusion (independent per-token sampling severs dependencies among parallel-decoded tokens) and continuous diffusion (the denoiser is untied from valid token configurations until final decoding). HC-DLM couples discrete token generation with a continuous latent trajectory in one principled denoising process, with a training objective derived from a variational bound. It improves puzzle accuracy on Sudoku/Countdown and generative perplexity on LM1B over matched-size baselines.
More from Research
- Audit: 23% of "wrong" cache hits in a semantic caching benchmark were identical prompts — Reasonable_Royal_621 · 2026-10-02
- Amazon Research Awards Fall 2026 opens with 9 research tracks, deadline Nov 4 — DocXavi · 2026-10-02
- Meta research: Knowledge distillation cuts training data memorization by over 50% vs fine-tuning — lasha_nlp · 2026-10-02
- Hugging Face publishes the ultimate guide to multi-harness RL — Saboo_Shubham_ · 2026-10-02
- From Transformers to MoE: The Research Papers Behind Every Major AI Concept — goyalshaliniuk · 2026-10-02
- MoE Paper Explained: Sparsely-Gated Experts Boost Model Capacity Over 1000x — goyalshaliniuk · 2026-10-02