Qibin Zhao
Qibin Zhao works on tensor methods, robust learning and interpretable machine learning. He directs RIKEN’s Tensor Learning Team, which researches factorization, tensor networks and quantum learning. His coauthored work includes a 2023 probabilistic tensor-decomposition model and the 2024 tnGPS framework for discovering tensor-network search algorithms with language models. The latter uses iterative prompting to generate and refine candidate algorithms. His team’s research also covers representation learning and deep generative models.
Timeline
-
2023
Probabilistic tensor models
Zhao, Zerui Tao and Toshihisa Tanaka coauthored an undirected probabilistic model for tensor decomposition.
-
February 2024
Algorithm discovery with language models
Zhao coauthored tnGPS with Junhua Zeng, Chao Li, Zhun Sun and Guoxu Zhou. The original abstract describes iterative prompting to generate tensor-network structure-search algorithms. Its reported gains concern the tested benchmarks.
-
2024
Interpretable and robust methods
The team has published work on multi-output regression under changing combinations, adversarially robust clustering and tensor-network structure search. These are concrete methods in the team’s wider effort to make learned models more efficient and reliable.
-
As of October 2026
Tensor Learning Team
Zhao directs the Tensor Learning Team. His research covers tensor methods, robust and interpretable learning, and quantum learning. He is also interested in collaboration on brain interfaces and medical analysis.
Sources and credits
Photo: www.riken.jp; Qibin Zhao profile photograph; creator not stated. License: Not stated by source. Source: riken.jp. Cropped to a square and resized. Ask us to remove or replace this photo.