Four papers accepted by NeurIPS 2025 and two papers accepted by NeurIPS 2025 workshops!

From the OIST machine learning and data science (MLDS) unit, we have four papers accepted! Congratulations!

Main venue

• Sasan Sharifipour, Constantino Alvarez Casado, Mohammad Sabokrou, & Miguel Bordallo López. APML: Adaptive Probabilistic Matching Loss for Robust 3D Point Cloud Reconstruction. In NeurIPS, 2025.
• Mojtaba Nafez, Mobina Poulaei, Nikan Vasei, Bardia, Mohammad Sabokrou, & Mohammad Hossein Rohban. FrameShield: Adversarially Robust Video Anomaly Detection. In NeurIPS, 2025. 
• Han Bao, Shinsaku Sakaue, & Yuki Takezawa. Any-stepsize Gradient Descent for Separable Data under Fenchel–Young Losses. In NeurIPS, 2025.
• Kenta Niwa, Yuki Takezawa, Guoqiang Zhang, & W. Bastiaan Kleijn. Revisiting 1-peer Exponential Graph for Enhancing Decentralized Learning Efficiency. In NeurIPS, 2025.

Workshops

• Kaoru Otsuka, Yuki Takezawa, & Makoto Yamada. Delayed Momentum Aggregation: Communication-efficient Byzantine-robust Federated Learning with Partial Participation. In OPT workshop 2025 @ NeurIPS 2025
• Mohammed-Yassine Habibi, Klea Ziu, & Makoto Yamada, Neuroscience-Inspired Encoding and Learning: A Path to Robust Representation Learning. Women in Machine Learning Workshop @ NeurIPS 2025