no code implementations • 15 Oct 2024 • Zhiyuan Zhang, Dongdong Chen, Jing Liao
Scene graphs offer a structured, hierarchical representation of images, with nodes and edges symbolizing objects and the relationships among them.
1 code implementation • 12 Aug 2024 • Zhiyuan Zhang, Licheng Yang, Zhiyu Xiang
We achieve an overall accuracy of 96. 0% (+4. 7%) on ModelNet40, 93. 1% (+12. 8%) on ScanObjectNN, and class accuracies of 91. 5% (+3. 6%), 82. 7% (+5. 1%), and 78. 5% (+9. 2%) on the three categories of the FG3D dataset for the fine-grained classification task.
1 code implementation • 6 Jun 2024 • Xiaosong Jia, Zhenjie Yang, QiFeng Li, Zhiyuan Zhang, Junchi Yan
In an era marked by the rapid scaling of foundation models, autonomous driving technologies are approaching a transformative threshold where end-to-end autonomous driving (E2E-AD) emerges due to its potential of scaling up in the data-driven manner.
1 code implementation • 18 May 2024 • Zeyu Zhang, Yiran Wang, Biao Wu, Shuo Chen, Zhiyuan Zhang, Shiya Huang, Wenbo Zhang, Meng Fang, Ling Chen, Yang Zhao
Firstly, we proposed a novel agent-based approach named Motion Avatar, which allows for the automatic generation of high-quality customizable human and animal avatars with motions through text queries.
no code implementations • 16 Apr 2024 • Peiyuan Zhi, Zhiyuan Zhang, Muzhi Han, Zeyu Zhang, Zhitian Li, Ziyuan Jiao, Baoxiong Jia, Siyuan Huang
Autonomous robot navigation and manipulation in open environments require reasoning and replanning with closed-loop feedback.
no code implementations • CVPR 2024 • Nan Jiang, Zhiyuan Zhang, Hongjie Li, Xiaoxuan Ma, Zan Wang, Yixin Chen, Tengyu Liu, Yixin Zhu, Siyuan Huang
Confronting the challenges of data scarcity and advanced motion synthesis in human-scene interaction modeling, we introduce the TRUMANS dataset alongside a novel HSI motion synthesis method.
1 code implementation • 25 Dec 2023 • Xiangyuan Xue, Kailing Wang, Jiazi Bu, Qirui Li, Zhiyuan Zhang
In this work, we propose MetaScript, a novel Chinese content generation system designed to address the diminishing presence of personal handwriting styles in the digital representation of Chinese characters.
no code implementations • 22 Nov 2023 • Tuan-Anh Vu, Srinjay Sarkar, Zhiyuan Zhang, Binh-Son Hua, Sai-Kit Yeung
We are inspired by the recent revolution of learning implicit representation and point cloud upsampling, which can produce high-quality 3D surface reconstruction and proximity-to-surface, respectively.
2 code implementations • 5 Oct 2023 • Omar Khattab, Arnav Singhvi, Paridhi Maheshwari, Zhiyuan Zhang, Keshav Santhanam, Sri Vardhamanan, Saiful Haq, Ashutosh Sharma, Thomas T. Joshi, Hanna Moazam, Heather Miller, Matei Zaharia, Christopher Potts
The ML community is rapidly exploring techniques for prompting language models (LMs) and for stacking them into pipelines that solve complex tasks.
no code implementations • 13 Sep 2023 • Zhiyuan Zhang, Hua Yang, Zhouping Yin
However, these methods need to be more precise for accurately measuring the displacement of markers during large elastic deformation of the gel, as this can significantly impact the accuracy of downstream tasks.
1 code implementation • 11 Sep 2023 • Ruibo Chen, Zhiyuan Zhang, Yi Liu, Ruihan Bao, Keiko Harimoto, Xu sun
Existing multimodal works that train models from scratch face the problem of lacking universal knowledge when modeling financial news.
no code implementations • 22 Jun 2023 • Zhiyuan Zhang, Zhitong Huang, Jing Liao
However, none of these methods have been able to edit the layout of single existing images.
2 code implementations • 31 May 2023 • Xue Xia, Pong Eksombatchai, Nikil Pancha, Dhruvil Deven Badani, Po-Wei Wang, Neng Gu, Saurabh Vishwas Joshi, Nazanin Farahpour, Zhiyuan Zhang, Andrew Zhai
This paper (1) presents Pinterest's ranking architecture for Homefeed, our personalized recommendation product and the largest engagement surface; (2) proposes TransAct, a sequential model that extracts users' short-term preferences from their realtime activities; (3) describes our hybrid approach to ranking, which combines end-to-end sequential modeling via TransAct with batch-generated user embeddings.
no code implementations • 8 May 2023 • Zhiyuan Zhang, Deli Chen, Hao Zhou, Fandong Meng, Jie zhou, Xu sun
To settle this issue, we propose the Fine-purifying approach, which utilizes the diffusion theory to study the dynamic process of fine-tuning for finding potentially poisonous dimensions.
1 code implementation • ICCV 2023 • Nan Jiang, Tengyu Liu, Zhexuan Cao, Jieming Cui, Zhiyuan Zhang, Yixin Chen, He Wang, Yixin Zhu, Siyuan Huang
By learning the geometrical relationships in HOI, we devise the very first model that leverage human pose estimation to tackle the estimation of articulated object poses and shapes during whole-body interactions.
no code implementations • 26 Oct 2022 • Zhiyuan Zhang, Yuchao Dai, Bin Fan, Jiadai Sun, Mingyi He
In this paper, we propose to learn a robust task-specific feature descriptor to consistently describe the correct point correspondence under interference.
no code implementations • 26 Oct 2022 • Zhiyuan Zhang, Jiadai Sun, Yuchao Dai, Bin Fan, Qi Liu
In response, this paper presents a novel end-to-end learning-based method to estimate the dense correspondence of 3D point clouds, in which the problem of point matching is formulated as a zero-one assignment problem to achieve a permutation matching matrix to implement the one-to-one principle fundamentally.
1 code implementation • 18 Oct 2022 • Zhiyuan Zhang, Lingjuan Lyu, Xingjun Ma, Chenguang Wang, Xu sun
In this work, we take the first step to exploit the pre-trained (unfine-tuned) weights to mitigate backdoors in fine-tuned language models.
1 code implementation • 14 Oct 2022 • Sishuo Chen, Wenkai Yang, Zhiyuan Zhang, Xiaohan Bi, Xu sun
In this work, we take the first step to investigate the unconcealment of textual poisoned samples at the intermediate-feature level and propose a feature-based efficient online defense method.
no code implementations • 13 Oct 2022 • Zhiyuan Zhang, Qi Su, Xu sun
NLP attacks tend to have small relative backdoor strengths, which may result in the failure of robust federated aggregation methods for NLP attacks.
no code implementations • 13 Oct 2022 • Zhiyuan Zhang, Ruixuan Luo, Qi Su, Xu sun
It demonstrates that flat minima tend to imply better generalization abilities.
1 code implementation • 11 Oct 2022 • Ruibo Chen, Wei Li, Zhiyuan Zhang, Ruihan Bao, Keiko Harimoto, Xu sun
Our method can model the common pattern behind different stocks with a meta-learner, while modeling the specific pattern for each stock across time spans with stock-dependent parameters.
1 code implementation • 4 Aug 2022 • Lei LI, Zhiyuan Zhang, Ruihan Bao, Keiko Harimoto, Xu sun
Traditional knowledge distillation in classification problems transfers the knowledge via class correlations in the soft label produced by teacher models, which are not available in regression problems like stock trading volume prediction.
1 code implementation • CVPR 2022 • Bin Fan, Yuchao Dai, Zhiyuan Zhang, Qi Liu, Mingyi He
Then, a refinement scheme is proposed to guide the GS frame synthesis along with bilateral occlusion masks to produce high-fidelity GS video frames at arbitrary times.
no code implementations • 24 Mar 2022 • Zhiyuan Zhang, Jiadai Sun, Yuchao Dai, Bin Fan, Mingyi He
3D point cloud registration is fragile to outliers, which are labeled as the points without corresponding points.
no code implementations • 24 Mar 2022 • Zhiyuan Zhang, Jiadai Sun, Yuchao Dai, Dingfu Zhou, Xibin Song, Mingyi He
Existing correspondences-free methods generally learn the holistic representation of the entire point cloud, which is fragile for partial and noisy point clouds.
no code implementations • 13 Mar 2022 • Jiaqi Gu, Zhiyu Xiang, Pan Zhao, Tingming Bai, Lingxuan Wang, Xijun Zhao, Zhiyuan Zhang
In recent years 3D object detection from LiDAR point clouds has made great progress thanks to the development of deep learning technologies.
1 code implementation • 26 Feb 2022 • Zhiyuan Zhang, Binh-Son Hua, Sai-Kit Yeung
3D point clouds deep learning is a promising field of research that allows a neural network to learn features of point clouds directly, making it a robust tool for solving 3D scene understanding tasks.
no code implementations • 28 Oct 2021 • Zhiyuan Zhang, Jiadai Sun, Yuchao Dai, Dingfu Zhou, Xibin Song, Mingyi He
Even though considerable progress has been made in deep learning-based 3D point cloud processing, how to obtain accurate correspondences for robust registration remains a major challenge because existing hard assignment methods cannot deal with outliers naturally.
3 code implementations • EMNLP 2021 • Runxin Xu, Fuli Luo, Zhiyuan Zhang, Chuanqi Tan, Baobao Chang, Songfang Huang, Fei Huang
Recent pretrained language models extend from millions to billions of parameters.
no code implementations • 7 Sep 2021 • Zhiyuan Zhang, Ruixuan Luo, Xuancheng Ren, Qi Su, Liangyou Li, Xu sun
To enhance neural networks, we propose the adversarial parameter defense algorithm that minimizes the average risk of multiple adversarial parameter corruptions.
no code implementations • ICLR 2022 • Zhiyuan Zhang, Lingjuan Lyu, Weiqiang Wang, Lichao Sun, Xu sun
In this work, we observe an interesting phenomenon that the variations of parameters are always AWPs when tuning the trained clean model to inject backdoors.
no code implementations • 20 Aug 2021 • Zhiyuan Zhang, Wei Li, Ruihan Bao, Keiko Harimoto, Yunfang Wu, Xu sun
Besides the security concerns of potential adversarial examples, adversarial training can also improve the generalization ability of neural networks, train robust neural networks, and provide interpretability for neural networks.
no code implementations • NAACL 2021 • Zhiyuan Zhang, Xuancheng Ren, Qi Su, Xu sun, Bin He
Motivated by neuroscientific evidence and theoretical results, we demonstrate that side effects can be controlled by the number of changed parameters and thus, we propose to conduct \textit{neural network surgery} by only modifying a limited number of parameters.
no code implementations • 15 May 2021 • Fenglin Liu, Xuancheng Ren, Zhiyuan Zhang, Xu sun, Yuexian Zou
In this work, we investigate how the scale factors in the effectiveness of the skip connection and reveal that a trivial adjustment of the scale will lead to spurious gradient exploding or vanishing in line with the deepness of the models, which could be addressed by normalization, in particular, layer normalization, which induces consistent improvements over the plain skip connection.
1 code implementation • NAACL 2021 • Wenkai Yang, Lei LI, Zhiyuan Zhang, Xuancheng Ren, Xu sun, Bin He
However, in this paper, we find that it is possible to hack the model in a data-free way by modifying one single word embedding vector, with almost no accuracy sacrificed on clean samples.
no code implementations • 19 Feb 2021 • Yuting Luo, Sum Wai Chiang, Lei Tang, Zhiyuan Zhang, Fengning Yang, Qiangmin Yu, Baofu Ding, Bilu Liu
Understanding the mechanism and developing strategies toward efficient electrocatalysis at gas-liquidsolid interfaces are important yet challenging.
Applied Physics Materials Science
1 code implementation • 25 Dec 2020 • Ruixuan Luo, Wei Li, Zhiyuan Zhang, Ruihan Bao, Keiko Harimoto, Xu sun
Recent deep learning based methods focus on learning clustering oriented representations.
no code implementations • COLING 2020 • Fenglin Liu, Xuancheng Ren, Zhiyuan Zhang, Xu sun, Yuexian Zou
In this work, we investigate how the scale factors in the effectiveness of the skip connection and reveal that a trivial adjustment of the scale will lead to spurious gradient exploding or vanishing in line with the deepness of the models, which could by addressed by normalization, in particular, layer normalization, which induces consistent improvements over the plain skip connection.
no code implementations • 25 Nov 2020 • Linhao Luo, Liqi Yang, Ju Xin, Yixiang Fang, Xiaofeng Zhang, Xiaofei Yang, Kai Chen, Zhiyuan Zhang, Kai Liu
In particular, we technically propose a novel random CNN component that can randomly convolute non-adjacent features to capture their interaction information and learn feature embeddings of key attributes to make the final recommendation.
no code implementations • Findings of the Association for Computational Linguistics 2020 • Zhiyuan Zhang, Xiaoqian Liu, Yi Zhang, Qi Su, Xu sun, Bin He
Conventional knowledge graph embedding (KGE) often suffers from limited knowledge representation, leading to performance degradation especially on the low-resource problem.
no code implementations • 7 Aug 2020 • Zhiyuan Zhang, Binh-Son Hua, Wei Chen, Yibin Tian, Sai-Kit Yeung
We found that a key reason is that compared to point coordinates, rotation-invariant features consumed by point cloud convolution are not as distinctive.
no code implementations • 18 Jun 2020 • Raymond Shiau, Hao-Yu Wu, Eric Kim, Yue Li Du, Anqi Guo, Zhiyuan Zhang, Eileen Li, Kunlong Gu, Charles Rosenberg, Andrew Zhai
As online content becomes ever more visual, the demand for searching by visual queries grows correspondingly stronger.
1 code implementation • 10 Jun 2020 • Xu Sun, Zhiyuan Zhang, Xuancheng Ren, Ruixuan Luo, Liangyou Li
We argue that the vulnerability of model parameters is of crucial value to the study of model robustness and generalization but little research has been devoted to understanding this matter.
no code implementations • 10 Jun 2020 • Lifu Wang, Bo Shen, Ning Zhao, Zhiyuan Zhang
In this paper, we follow this line to study the topology (sub-level sets) of the loss landscape of deep ReLU neural networks with a skip connection and theoretically prove that the skip connection network inherits the good properties of the two-layer network and skip connections can help to control the connectedness of the sub-level sets, such that any local minima worse than the global minima of some two-layer ReLU network will be very ``shallow".
2 code implementations • 25 Dec 2019 • Guangxiang Zhao, Junyang Lin, Zhiyuan Zhang, Xuancheng Ren, Qi Su, Xu sun
Self-attention based Transformer has demonstrated the state-of-the-art performances in a number of natural language processing tasks.
no code implementations • 1 Dec 2019 • Zhiyuan Zhang, Xiaoqian Liu, Yi Zhang, Qi Su, Xu sun, Bin He
Learning knowledge graph embeddings (KGEs) is an efficient approach to knowledge graph completion.
2 code implementations • 17 Nov 2019 • Guangxiang Zhao, Xu sun, Jingjing Xu, Zhiyuan Zhang, Liangchen Luo
In this work, we explore parallel multi-scale representation learning on sequence data, striving to capture both long-range and short-range language structures.
Ranked #8 on Machine Translation on WMT2014 English-French
2 code implementations • NeurIPS 2019 • Jingjing Xu, Xu sun, Zhiyuan Zhang, Guangxiang Zhao, Junyang Lin
Unlike them, we find that the derivatives of the mean and variance are more important than forward normalization by re-centering and re-scaling backward gradients.
Ranked #5 on Machine Translation on IWSLT2015 English-Vietnamese
no code implementations • 25 Sep 2019 • Guangxiang Zhao, Junyang Lin, Zhiyuan Zhang, Xuancheng Ren, Xu sun
Extensive experimental results on a series of natural language processing tasks, including neural machine translation, image captioning, and language modeling, all demonstrate the advantages of Sparse Transformer in model performance.
1 code implementation • ICCV 2019 • Zhiyuan Zhang, Binh-Son Hua, Sai-Kit Yeung
Deep learning with 3D data has progressed significantly since the introduction of convolutional neural networks that can handle point order ambiguity in point cloud data.
Ranked #9 on 3D Semantic Segmentation on DALES
1 code implementation • 17 Aug 2019 • Zhiyuan Zhang, Binh-Son Hua, David W. Rosen, Sai-Kit Yeung
Our core idea is to use low-level rotation invariant geometric features such as distances and angles to design a convolution operator for point cloud learning.
4 code implementations • 27 Jun 2019 • Ruixuan Luo, Jingjing Xu, Yi Zhang, Zhiyuan Zhang, Xuancheng Ren, Xu sun
Through this method, we generate synthetic data using a large amount of unlabeled data in the target domain and then obtain a word segmentation model for the target domain.
no code implementations • 24 May 2019 • Zhiyuan Zhang, Pengcheng Yang, Xuancheng Ren, Qi Su, Xu sun
Neural network learning is usually time-consuming since backpropagation needs to compute full gradients and backpropagate them across multiple layers.
no code implementations • 14 Aug 2018 • Zhiyuan Zhang, Wei Li, Jingjing Xu, Xu sun
We define the primal meaning of an expression to be a frequently used sense of that expression from which its other frequent senses can be deduced.
no code implementations • 5 Mar 2018 • Zhiyuan Zhang, Wei Li, Qi Su
In this paper, we propose to build an end-to-end neural model to automatically translate between ancient and contemporary Chinese.
1 code implementation • LREC 2018 • Xuancheng Ren, Xu sun, Ji Wen, Bingzhen Wei, Weidong Zhan, Zhiyuan Zhang
Web 2. 0 has brought with it numerous user-produced data revealing one's thoughts, experiences, and knowledge, which are a great source for many tasks, such as information extraction, and knowledge base construction.