New survey on continuous diffusion and flow language models caps CDLM comeback year
sedielem · x · 2026-11-09
sophiatang announces a new survey on Continuous Diffusion and Flow Language Models (CDLMs). She recounts her early-2025 work on continuous simplex flows: at the time, flow matching in continuous spaces was much harder to train than discrete diffusion, but 2026 flipped the picture — continuous LMs took off across pre-training, few-step generation, guidance, and scaling, and the survey (with @chandavidlee, @FEijkelboom and others) maps this fast-moving landscape. sedielem calls it a beautiful wrap-up of the comeback year for CDLMs.
Related event: New Survey Charts Rise of Continuous Diffusion and Flow Language Models(4 posts)→
More from Research
- Perplexity Frontier Research agenda: multi-agent orchestration, continual learning, next-gen architectures — andrewgwils · 2026-11-09
- Formulate a problem as alternating convex optimization and solvers follow naturally — caglar_ee · 2026-11-09
- GitHub study: engineer productivity rises 2.8% per 10% growth of local tech cluster — lemire · 2026-11-09
- 5 PEFT techniques explained: what LoRA, VeRA, Delta-LoRA actually do to model weights — techNmak · 2026-11-09
- Guava distills GPT-5.4 into a 4B robot agent with just 2,000 simulated trajectories — jbhuang0604 · 2026-11-09
- Dex-One2Many learns dexterous manipulation from one human video, 95%+ success across 4 hands — jbhuang0604 · 2026-11-09