Ziwei Wang
Ziwei Wang is currently a postdoc fellow in Robotics Institute, Carnegie Mellon University, supervised by Prof. Changliu Liu. He received the Ph.D and the B.S degrees from the Department of Automation, Tsinghua University in 2023 and the Department of Physics, Tsinghua University in 2018 respectively.
His research interests include tiny machine learning and embodied visual perception. He has published over 20 scientific papers in the IEEE Transactions on Pattern Analysis and Machine Intelligence (6 papers in TPAMI), International Journal of Computer Vision, IEEE Robotics and Automation Letters, CVPR, ICCV, ECCV, NeurIPS, IROS and ICRA.
He serves as a regular reviewer member for the IEEE Transactions on Pattern Analysis and Machine Intelligence, IEEE Transactions on Image Processing, IEEE Transactions on Circuits and Systems for Video Technology, IEEE Robotics and Automation Letters, Pattern Recognition, CVPR, ICCV, ECCV, NeurIPS, ICML, ICLR and ICRA.
Email  / 
CV  / 
Google Scholar  / 
Github
|
|
News
2024-03: Our paper on general robotic manipulation with dynamic Gaussian Splatting is pre-printed on Arxiv. [PDF][Website][Code]
2024-02: Three papers are accepted to CVPR 2024.
2024-02: One paper is accepted to TIP.
2023-10: One paper is accepted to TPAMI.
2023-09: One paper is accepted to NeurIPS 2023.
2023-07: Our paper on embodied task planning with large language models is pre-printed on Arxiv. [PDF][Website][Code][Demo][Model]
2023-03: One paper is accepted to IJCV.
|
Preprint
Guanxing Lu, Shiyi Zhang, Ziwei Wang, Changliu Liu, Jiwen Lu, Yansong Tang. ManiGaussian: Dynamic Gaussian Splatting for Multi-task Robotic Manipulation. [PDF][Website][Code]
Zhenyu Wu, Ziwei Wang, Xiuwei Xu, Jie Zhou, Jiwen Lu. Embodied Task Planning with Large Language Models. [PDF][Website][Code][Demo][Model]
|
* means equal contribution
|
Towards Accurate Data-free Quantization for Diffusion Models
Changyuan Wang, Ziwei Wang, Xiuwei Xu, Yansong Tang, Jie Zhou, Jiwen Lu
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024.
[(coming soon)]
we propose a post-training quantization framework to compress diffusion models, which performs group-wise quantization to minimize rounding errors across time steps and selects generated contents in the optimal time steps for calibration.
|
|
MCUFormer: Deploying Vision Transformers on Microcontrollers with Limited Memory
Yinan Liang, Ziwei Wang, Xiuwei Xu, Yansong Tang, Jie Zhou, Jiwen Lu
Thirty-seventh Conference on Neural Information Processing Systems (NeurIPS), 2023.
[PDF]
[Supp]
[Code]
we propose a hardware-algorithm co-optimizations method called MCUFormer to deploy vision transformers on microcontrollers with extremely limited memory, where we jointly design transformer architecture and construct the inference operator library to fit the memory resource constraint.
|
|
Learning Accurate Performance Predictors for Ultrafast Automated Model Compression
Ziwei Wang, Jiwen Lu, Han Xiao, Shengyu Liu, Jie Zhou
International Journal of Computer Vision (IJCV), 2023.
[PDF]
[Code]
We propose an ultrafast auto- mated model compression framework for flexible network deployment, where we can obtain the optimal compression policy within several seconds.
|
|
Category-level Shape Estimation for Densely Cluttered Objects
Zhenyu Wu, Ziwei Wang, Jiwen Lu, Haibin Yan
IEEE International Conference on Robotics and Automation (ICRA), 2023.
[PDF]
[Code]
We propose a category-level shape estimation method for densely cluttered objects, which addresses the challenges of large object segmentation errors and inaccurate shape recovery on unseen instances.
|
|
Quantformer: Learning Extremely Low-precision Vision Transformers
Ziwei Wang, Changyuan Wang, Xiuwei Xu, Jie Zhou, Jiwen Lu
IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), 2022.
[PDF]
[Supp]
[Code]
We propose the extremely low-precision vision transformers in 2-4 bits, where the self-attention rank consistency and group-wise quantization are presented for quantization error minimization.
|
|
Planning Irregular Object Packing via Hierarchical Reinforcement Learning
Sichao Huang, Ziwei Wang, Jie Zhou, Jiwen Lu
IEEE Robotics and Automation Letters (RAL), 2022
[PDF]
[Robot Demo]
[Simulation Demo]
[Code]
we develop a packing planning method for general objects including the packing sequence, locations and orientations to maximize the space utilization ratio.
|
|
Smart Explorer: Recognizing Objects in Dense Clutter via Interactive Exploration
Zhenyu Wu*, Ziwei Wang*, Zibu Wei, Yi Wei, Haibin Yan
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2022.
[PDF]
[Demo]
[Code]
We propose an interactive exploration framework called Smart Explorer for recognizing all objects in dense clutters.
|
|
GE-Grasp: Efficeint Target Oriented Grasping in Dense Clutter
Zhan Liu, Ziwei Wang, Sichao Huang, Jie Zhou, Jiwen Lu
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2022.
[PDF]
[Demo]
[Code]
we present a generic framework for robotic motion planning in dense clutter with diverse action primitives and generator-evaluator architectures.
|
|
Shapley-NAS: Discovering Operation Contribution for Neural Architecture Search
Han Xiao, Ziwei Wang, Zheng Zhu, Jie Zhou, Jiwen Lu
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022.
[PDF]
[Supplement]
[Code]
We propose a Shapley value based method to evaluate operation contribution (Shapley-NAS) for neural architecture search.
|
|
Generalizable Mixed-Precision Quantization via Attribution Rank Preservation
Ziwei Wang, Han Xiao, Jiwen Lu, Jie Zhou
IEEE International Conference on Computer Vision (ICCV), 2021.
[PDF]
[Supplement]
[Code]
We propose a generalizable mixed-precision quantization (GMPQ) method for efficient inference.
|
|
Learning Efficient Binarized Object Detectors with Information Compression
Ziwei Wang, Jiwen Lu, Ziyi Wu, Jie Zhou
IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI), 2022.
[PDF]
[Supplement]
[Code]
We propose apresent binary neural networks with automatic information compression (AutoBiDet) to automatically adjust the IB trade-off according to the input complexity.
|
|
BiDet: An Efficient Binarized Object Detector
Ziwei Wang, Ziyi Wu, Jiwen Lu, Jie Zhou
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2020
[PDF] [Code]
We propose an efficient binarized object detector that fully utilizes the representational capacity of the binary neural networks by redundancy removal.
|
|
Learning Channel-wise Interactions for Binary Convolutional Neural Networks
Ziwei Wang, Jiwen Lu, Chenxin Tao, Jie Zhou
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019
[PDF] [Code]
We propose a channel-wise interaction based binary convolutional neural network learning method (CI-BCNN) for efficient inference with minimal information loss.
|
|
GraphBit: Bitwise Interaction Mining via Deep Reinforcement Learning
Yueqi Duan, Ziwei Wang, Jiwen Lu, Xudong Lin, Jie Zhou
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2018
[PDF] [Code]
We propose a GraphBit method to learn deep binary descriptors with enhanced reliability in a directed acyclic graph unsupervisedly.
|
|
Learning Deep Binary Descriptor with Multi-Quantization
Yueqi Duan, Jiwen Lu, Ziwei Wang, Jianjiang Feng, Jie Zhou
IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2017
[PDF]
We present an unsupervised feature learning method called deep binary descriptor with multiquantization (DBD-MQ) for visual matching.
|
Honors and Awards
2023 Outstanding Doctoral Dissertation of Tsinghua University
2022 National Scholarship
2020 National Scholarship
2018 Chi-Sun Yeh Scholarship
2016 Qualcomm Scholarship
|
Academic Services
Conference Reviewer: CVPR 2020/2021/2022/2023, ICCV 2021/2023, ECCV 2022, NeurIPS 2020/2021/2022/2023, ICML2021/2022/2023, ICLR 2021/2022/2023, IJCAI 2022, ICRA 2023
Journal Reviewer: T-IP, T-CSVT, R-AL, T-BIOM, Pattern Recognition, JVIC
|
|