Guy Van den Broeck
Professor
Directs the StarAI lab on probabilistic circuits, tractable probabilistic reasoning and neurosymbolic learning.
Guy Van den Broeck's lab at UCLA works on probabilistic & causal inference, reasoning & neuro-symbolic AI and reasoning & agents. 70 papers since 2023, 49 at top venues, mostly NeurIPS and ICML. 7 PhD students. 23 alumni, most often to Meta. Applications close .
At a glance
- papers, 2023–2026
- 18/ year
- at T1 venues
- 70%
- PhD students
- 7
- MS students
- 0
- alumni listed
- 23
Papers by yearfrom the lab homepage, 2023–2026
- 23 papers2026 · T1 11 · collecting
- 21 papers2025 · T1 16
- 12 papers2024 · T1 9
- 14 papers2023 · T1 13
T1 venuesOtherVenues: NeurIPS, ICML, ICLR, AISTATS, AAAI
Research focuspaper titles that use each topic's words
- Probabilistic & causal inference27 papers
- probabilistic23 papers
- Reasoning & neuro-symbolic AI4 papers
- Reasoning & agents1 paper
- probabilistic reasoning1 paper
Paper list
From the lab homepage.
202623 papers
- SATA canonical generalization of OBDD
- ICDTAlgorithms for Optimizing Acyclic Queries
- ICML (T1)Breaking the Factorization Barrier in Diffusion Language Models
- NeurIPS (T1)CoTs as Probabilistic Programs: A Programmatic View of Thinking Step-by-Step in Language Models
- TPMCoTs as Tractable Probabilistic Programs
- EACLEnabling Autoregressive Models to Fill In Masked Tokens
- TPMEntropy-Guided LLM Decoding via Probabilistic Circuits
- HPCA (T1)ESTroM: Element-Flow Architecture For Processing Sparse Tractable Probabilistic Models
- PLOS GeneticsGPC: An expressive and tractable deep generative model for genetic variation data
- ICLR WorkshopGPC: Deep generative model of genetic variation data improves imputation accuracy in private populations
- PriGMHow Should We Train Tractable Proxies for Language Models?
- GCLRHow to Marginalize in Causal Structure Learning?
202521 papers
- NeurIPS (T1)Accelerating Diffusion LLMs via Adaptive Parallel Decoding
- ACL (T1)Adversarial Tokenization
- TPMBitblasting for Constrained Decorrelation in Tractable Image Modeling
- ICLR (T1)Controllable Generation via Locally Constrained Resampling
- ICLR (T1)Discrete Copula Diffusion
- DCCEnhancing and Evaluating Probabilistic Circuits for High-Resolution Lossless Image Compression
- ICLR (T1)Learning to Discretize Denoising Diffusion ODEs
- TMLR (T1)On the Challenges and Opportunities in Generative AI
- AAAI (T1)On the Relationship between Monotone and Squared Probabilistic Circuits
- NeurIPS (T1)Plug-and-Play Context Feature Reuse for Efficient Masked Generation
- AISTATS (T1)Prepacking: A Simple Method for Fast Prefilling and Increased Throughput in Large Language Models
- NeurIPS (T1)Rao-Blackwell Gradient Estimators for Equivariant Denoising Diffusion
202412 papers
- NeurIPS (T1)A Compositional Atlas for Algebraic Circuits
- NeurIPS (T1)A Tractable Inference Perspective of Offline RL
- NeurIPS (T1)Adaptable Logical Control for Large Language Models
- PACMPL (T1)Bit Blasting Probabilistic Programs
- ICLR (T1)Image Inpainting via Tractable Steering of Diffusion Models
- UAI (T1)Polynomial Semantics of Tractable Probabilistic Circuits
- TPMProbabilistic Circuits for Cumulative Distribution Functions
- ICLR (T1)Probabilistically Rewired Message-Passing Neural Networks
- Dagstuhl ReportsScalable Analysis of Probabilistic Models and Programs (Dagstuhl Seminar 23241)
- ICML (T1)Scaling Tractable Probabilistic Circuits: A Systems Perspective
- Chapter in Compendium of Neurosymbolic Artificial IntelligenceSemantic Loss Functions for Neuro-Symbolic Structured Prediction
- EMNLP (T1)Where is the signal in tokenization space?
202314 papers
- NeurIPS (T1)A Pseudo-Semantic Loss for Deep Generative Models with Logical Constraints
- NeurIPS (T1)A Unified Approach to Count-Based Weakly Supervised Learning
- AAAI (T1)Certifying Fairness of Probabilistic Circuits
- NeurIPS (T1)Collapsed Inference for Bayesian Deep Learning
- AISTATS (T1)Mixtures of All Trees
- IJCAI (T1)On the Paradox of Learning to Reason from Data
- AAAI (T1)Out-of-Distribution Generalization by Neural-Symbolic Joint Training
- ICML WorkshopProbabilistic Task-Adaptive Graph Rewiring
- UAI (T1)Scaling Integer Arithmetic in Probabilistic Programs
- ICLR (T1)Scaling Up Probabilistic Circuits by Latent Variable Distillation
- AISTATS (T1)Semantic Strengthening of Neuro-Symbolic Learning
- ICLR (T1)SIMPLE: A Gradient Estimator for k-subset sampling
Where alumni went23 listed, 10% to industry
Industry 1Academia 9
Most often to
- Meta
Also to
- Arizona State University
- Northeastern University
- NUS
Apply to UCLAFall 2027
Computer Science PhD
- Round
- Single round (Dec 15)
- Opens
- —
- GRE
- optional
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- GBImperial
Stefan Vlaski
- Learning theory & optimization
- USNYU
Rudy Morel
- Learning theory & optimization
- CNTsinghua
Hairong Lv
- Trustworthy & safe ML
- CNSJTU
Hongsheng Hu
- Trustworthy & safe ML
- USUCSD
Mikhail Belkin
- Learning theory & optimization
- CHETH Zurich
Celestine Mendler-Dünner
- Interpretability, evaluation & safety of LMs
Your notes
Save this lab to keep notes.
Data
Checked against the lab homepage.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
UCLAcloses