Guy Van den Broeck

Professor of Computer Science at UCLA @UCLAComSci;
working on Artificial Intelligence

[Hacker News 20 @betterhn20]

A Canonical Generalization of OBDD https://t.co/HaL2j51faK (https://t.co/ueOt9bOrZV)

[Daniel Israel @danielmisrael]

Congratulations @itisalex3 on acceptance to #ACL2026 🎉 If you are interested in KV cache compression, I highly recommend reading this paper.

[Poorva Garg @PoorvaGarg11]

Very excited to be leading this research direction, employing probabilistic programming to improve LLM inference for code generation.

[Daniel Israel @danielmisrael]

When you prompt an LLM for code, you get one deterministic program. However, the LLM actually defines a distribution over many programs, and existing methods discard it‼️ PPoT uses this distribution to extract free performance and efficiency gains. 🧵👇

[Konstantinos Kallas @KonsKallas]

We just implemented trai (try + AI) (https://t.co/50VH8V9CgR), a Claude plugin that can help you isolate 🫷 changes done in your file system by tool calls (like pip install), and only commit them if they are intended ☺️. Try it out (pun intended) and share your feedback!

[Guy Van den Broeck @guyvdb]

@trunghlt @IanLi1118 new tokens depend on already unmasked tokens, but they do not depend on each other when unmasking multiple tokens in a single step

[Guy Van den Broeck @guyvdb]

We put probabilistic circuits into diffusion language models and got a big boost in reasoning performance!

[Ian Li @IanLi1118]

One of the biggest promises of Diffusion LLMs is parallel generation: predicting multiple tokens at once to bypass the sequential bottleneck of autoregressive models.

However, parallel generation comes with a price. For example:

Should the sentence “He is from [MASK] [MASK]” be filled with [New] [York] or [San] [Diego]?

If a diffusion model predicts both at the exact same time, it assumes independence and may produce... [San] [York]. 🤦‍♂️

We argue this arises from a structural misspecification: models are restricted to fully factorized outputs because parameterizing the full joint distribution would require a prohibitively massive output head.

This is the Factorization Barrier crippling parallel generation. Here is how we broke it with CoDD.

[Zilei Shao @zileishao]

Check out our most recent work on dLLM!

[Anji Liu @liu_anji]

Check out our recent work offering a principled way to perform parallel prediction (few-step generation) in Diffusion LLMs with minimal performance degradation!

[Ian Li @IanLi1118]

One of the biggest promises of Diffusion LLMs is parallel generation: predicting multiple tokens at once to bypass the sequential bottleneck of autoregressive models.

However, parallel generation comes with a price. For example:

Should the sentence “He is from [MASK] [MASK]” be filled with [New] [York] or [San] [Diego]?

If a diffusion model predicts both at the exact same time, it assumes independence and may produce... [San] [York]. 🤦‍♂️

We argue this arises from a structural misspecification: models are restricted to fully factorized outputs because parameterizing the full joint distribution would require a prohibitively massive output head.

This is the Factorization Barrier crippling parallel generation. Here is how we broke it with CoDD.

[Zhihan Yang @zhihanyang_]

Join our reading group next Monday!

Paper: Planned Diffusion
Presenters: Daniel Israel (@danielmisrael), Tian Jin (@jintian)

[Discrete Diffusion Reading Group @diffusion_llms]

📢Feb 2 (Mon): Planned Diffusion

🙅Diffusion language models are capable of parallelizing text generation but can struggle with coherence in low time-step regimes.

💡Planned Diffusion unlocks a new axis of parallelism: Token-level parallelism ➡️ semantic parallelism

✍️Planned diffusion first generates a structured plan, then diffuses semantically independent spans of text in parallel according to the plan.

This Monday, Daniel Israel (UCLA) (@danielmisrael) and Tian Jin (MIT) (@jintian) will discuss their exciting Planned Diffusion paper as joint first authors.

Collaborators: Ellie Cheng (https://t.co/qdy0TAVHCD), Guy Van den Broeck (@guyvdb), Aditya Grover (@adityagrover_), Suvinay Subramanian (@suvinay), Michael Carbin (@mcarbin)

Paper link: https://t.co/ptRpLFXHDh

[Luis Lamb @luislamb]

⁦@RealAAAI⁩ AAAI2021Conference - Neuro-Symbolic AI Panel, during the COVID-19 crisis. With ⁦@kerstingAIML⁩ ⁦@guyvdb⁩ ⁦@mattbotvinick⁩ Marta Kwiatkowska, Leslie Pack Kaelbling. Just a picture 🙂

[Guy Van den Broeck @guyvdb]

@YuanqiD Thanks, you can find the paper here: https://t.co/DmiI6zGJuG

[NeSy 2026 @nesyconf]

Recordings of the NeSy 2025 keynotes are now available! 🎥

Check out insightful talks from @guyvdb , @tkipf and @dlmcguinness on our new Youtube channel.

Topics include using symbolic reasoning for LLM, and object-centric representations
https://t.co/iIHdKTr432

[Guy Van den Broeck @guyvdb]

I gave a keynote at @nesyconf on
"Symbolic Reasoning in the Age of Large Language Models"
Check out the recording if you are curious about neurosymbolic generative AI:
https://t.co/VUUJD4vdYB

[Alex Chen @itisalex3]

What happens when we compress the KV cache of prompts with multiple instructions? 🤔

Existing compression methods can lead to some instructions being ignored. 🙀

We propose simple changes to KV cache eviction that fix this problem alongside other pitfalls to be aware of. 💯

[Tian Jin @jintian]

Plan autoregressively, denoise in parallel!

[Ellie Cheng @ellieyhc]

Diffusion 🤝 Autoregressive
Fast high-quality generation

[Daniel Israel @danielmisrael]

"An hour of planning can save you 10 hours of doing."

✨📝 Planned Diffusion 📝 ✨ makes a plan before parallel dLLM generation.
Planned Diffusion runs 1.2-1.8× faster than autoregressive and an order of magnitude faster than diffusion, while staying within 0.9–5% AR quality.