Chiming Liu

Chiming Liu

Auto-generated profile, based on publicly available information

Chiming Liu

AgiBot, Shanghai, China

Research Focus & Metrics

Academic Career Length

20Years (2007-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
6 June 2024 in Nature Methods
Verified
Journal Article
Closed access
PDF file
MHMinsheng Hao
Minsheng Hao
JGJing Gong
Jing Gong
XZXin Zeng
Xin Zeng
CLChiming Liu
Chiming Liu
YGYucheng Guo
Yucheng Guo
LSLe Song
Le Song

3 April 2025 in Nature Methods
Verified
Journal Article
Closed access
BCBo Chen
Bo Chen
XCXingyi Cheng
Xingyi Cheng
PLPan Li
Pan Li
YGYangli-Ao Geng
Yangli-Ao Geng
JGJing Gong
Jing Gong
LSLe Song
Le Song

1 January 2007
Verified
Conference Paper
Closed access
CLChi-Huang Lu
Chi-Huang Lu
CTChing-Chih Tsai
Ching-Chih Tsai
CLChi-Ming Liu
Chi-Ming Liu
YCYuan-Hai Charng
Yuan-Hai Charng
This paper presents a predictive control based on recurrent neural network (RNN) for a class of nonlinear systems and investigates its application to temperature control of plastic injection molding processes. The RNN is used as a model identifier for approximating the nonlinear discrete-time systems and the multivariable predictive control based on the RNN is derived from a generalized predictive performance criterion. The adaptive learning rates of the RNN model and the controller are investigated via the discrete Lyapunov stability theorem, which are respectively used to guarantee the convergences of both the RNN model and the predictive controller. Finally, numerical simulations and experimental results are provided to demonstrate the effectiveness of the proposed control strategy under setpoint changes and bounded disturbances.

21 April 2026 in IEEE Transactions on Robotics
Verified
Journal Article
Open access
MSModi Shi
Modi Shi
LCLi Chen
Li Chen
JCJin Chen
Jin Chen
YLYuxiang Lu
Yuxiang Lu
CLChiming Liu
Chiming Liu
HLHongyang Li
Hongyang Li
Data scaling has driven remarkable success in foundation models for Natural Language Processing (NLP) and Computer Vision (CV), yet the principles of effective data scaling in robotic manipulation remain insufficiently understood. In this work, we investigate the nuanced role of data diversity in robot learning by examining three critical dimensions-task (what to do), embodiment (which robot to use), and expert (who demonstrates)-challenging the conventional intuition of “more diverse is better”. Throughout extensive experiments on various robot platforms, we reveal that (1) task diversity proves more critical than per-task demonstration quantity, with scene diversity playing a more important role than skill diversity for robustness and generalization under distribution shifts; (2) multi-embodiment pre-training data is non-essential for cross-embodiment transfer-models trained on high-quality single-embodiment data can efficiently transfer to different platforms, showing desirable scaling property during fine-tuning and its potential of replacing large-scale multi-embodiment pre-training; and (3) expert diversity, arising from individual operational preferences and stochastic variations in human demonstrations, can be confounding to policy learning, with action rate multimodality emerging as a key contributing factor. Based on this insight, we propose a distribution debiasing method to mitigate action rate ambiguity, the yielding GO-1-Pro achieves substantial performance gains of 15%, equivalent to using 2.5× pre-training data. Collectively, these findings provide new perspectives and offer practical guidance on how to scale robotic manipulation datasets effectively. The code will be released.