Publications

(*) denotes equal contribution, ($^\dagger$) denotes co-last authors

• Sigmoid Self-Attention is Better than Softmax Self-Attention: A Mixture-of-Experts Perspective
  ArXiv
  Fanqi Yan*, Huy Nguyen*, Pedram Akbarian, Nhat Ho $^\dagger$, Alessandro Rinaldo $^\dagger$

• Understanding Expert Structures on Minimax Parameter Estimation in Contaminated Mixture of Experts
  AISTATS 2025
  Fanqi Yan*, Huy Nguyen*, Dung Le*, Pedram Akbarian, Nhat Ho

• Statistical Perspective of Top-K Sparse Softmax Gating Mixture of Experts
  ICLR 2024
  Huy Nguyen, Pedram Akbarian, Fanqi Yan, Nhat Ho

• Is Parameter Learning via Weighted Model Integration Tractable?
  TPM Workshop at UAI 2021
  Zhe Zeng*, Paolo Morettin*, Fanqi Yan*, Antonio Vergari, Guy Van den Broeck

• Probabilistic Inference with Algebraic Constraints: Theoretical Limits and Practical Approximations
  NeurIPS 2020, Spotlight presentation, acceptance rate 280/9454 = 2.96%
  Zhe Zeng*, Paolo Morettin*, Fanqi Yan*, Antonio Vergari, Guy Van den Broeck

• Relax, Compensate and then Integrate
  DeCoDeML Workshop at ECML-PKDD 2020
  Zhe Zeng*, Paolo Morettin*, Fanqi Yan*, Antonio Vergari, Guy Van den Broeck

• Scaling up Hybrid Probabilistic Inference with Logical and Arithmetic Constraints via Message Passing
  ICML 2020
  Zhe Zeng*, Paolo Morettin*, Fanqi Yan*, Antonio Vergari, Guy Van den Broeck

• Hybrid Probabilistic Inference with Logical Constraints: Tractability and Message Passing
  KR2ML Workshop at NeurIPS 2019
  Zhe Zeng*, Fanqi Yan*, Paolo Morettin*, Antonio Vergari, Guy Van den Broeck

Conference Reviewing: ICML 2024, ICLR 2025