Benjamin Rozonoyer

We will be presenting our poster today at 4-5pm at the #SPIGM workshop at #ICML2026. If you are working on diffusion for text, either using discrete or continuous space approaches, you will find our results interested. Come chat with us. With @rozonoyer96703 and Jacopo Minniti.

Every non-autoregressive LM we picked up came with its own training script, dataloader, and eval code. so when a metric moved we couldn't tell if it was the new idea or the plumbing. So we built xLM: one CLI and one harness for training, eval, and generation. code: github.com/dhruvdcoder/xl… pypi: pypi.org/project/xlm-co… docs: dhruveshp.com/xlm-core/dev demo paper: arxiv.org/abs/2512.17065 📍 Come say hi! We will be at "Bridging Research and Open Source," social at #ICML2026. COEX center, Seoul, rooms E1-E4 | 📅 Wed July 8, 19:00-21:00 🧵👇

@yuezhouhu Really interesting paper! I wanted to share our recent paper arxiv.org/pdf/2605.22967 (at SPIGM and FoGen) which tackles the same problem of carrying forward computation in MDLMs, but uses Loopholing and truncated BPTT.

Very excited to see that the core idea of DiffusionGemma directly stems from our work, Residual Context Diffusion (arXiv:2601.22954)! Code- and architecture-level comparisons are attached. 🏆 RCD is accepted to ICML 2026! See you in Seoul! #DiffusionLLM #LLM #Reasoning #GenAI

Variable-length masked diffusion models (FlexMDM and friends) generate by inserting mask tokens into any gap and unmasking them. But the insertion/unmasking schedule is fixed and data-independent. So the model has to learn to produce every sequence in every possible order. For structured data that's a huge waste of capacity. How do you learn data-dependent insertion and unmasking orders without breaking tractable training? We propose LoFlexMDM, which does exactly that. 🧵👇

What if diffusion models could think ahead instead of being greedy at every step?🤔 We introduce: Learned Relay Representations for Forward-Thinking Discrete Diffusion Models