Enabling Approximate Joint Sampling in Diffusion LMs

arXiv:2509.22738 · cs.CL, cs.LG · Submitted 2025-09-25 · Read on arXiv

cs.CL, cs.LG

Submitted: 2025-09-25

Updated: 2026-09-24

Code: https://github.com/DreamLM/Dream

Terminology

• Diffusion Model — 36× in this paper · explained in 8 episodes
• Embeddings — 13× in this paper · explained in 2 episodes
• Accuracy — 11× in this paper · explained in 1 episode
• Speculative Decoding — 10× in this paper · explained in 2 episodes
• Diffusion Models — 9× in this paper · explained in 6 episodes
• Transformer — 5× in this paper · explained in 2 episodes
• Downstream Tasks — 4× in this paper · explained in 2 episodes
• Copula — 3× in this paper · explained in 2 episodes
• Verification — 3× in this paper · explained in 2 episodes
• Perplexity — 2× in this paper · explained in 3 episodes
• Verifier — 2× in this paper · explained in 2 episodes
• Benchmark — 1× in this paper · explained in 2 episodes
• Generalization — 1× in this paper · explained in 13 episodes
• Generative Model — 1× in this paper · explained in 2 episodes
• Generative Models — 1× in this paper · explained in 2 episodes
• Quantization — 1× in this paper · explained in 12 episodes
• Robustness — 1× in this paper · explained in 5 episodes
• Sample Complexity — 1× in this paper · explained in 15 episodes

Sources

• Discrete Copula Diffusion
• DeepSeek-V3 Technical Report
• Encoder-Decoder Diffusion Language Models for Efficient Training and Inference
• Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution
• Self-Speculative Masked Diffusions
• Large Language Diffusion Models
• Reviving Any-Subset Autoregressive Models with Principled Parallel Sampling and Speculative Decoding
• Anchored Diffusion Language Model
• OpenCoder: The Open Cookbook for Top-Tier Code Large Language Models
• Accelerating Diffusion LLMs via Adaptive Parallel Decoding
• Your LLM Knows the Future: Uncovering Its Multi-Token Prediction Potential
• ParallelBench: Understanding the Trade-offs of Parallel Decoding in Diffusion LLMs
• Mercury: Ultra-Fast Language Models Based on Diffusion
• Seed Diffusion: A Large-Scale Diffusion Language Model with High-Speed Inference
• Energy-Based Diffusion Language Models for Text Generation
• EAGLE: Speculative Sampling Requires Rethinking Feature Uncertainty
• Dream 7B: Diffusion Large Language Models
• EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Test
• MetaMath: Bootstrap Your Own Mathematical Questions for Large Language Models
• Learning Harmonized Representations for Speculative Sampling

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering