Recent Papers

For full paper list, please refer to Prof. Xu’s Google Scholar page.

Are Transformers Effective for Time Series Forecasting?

Ailing Zeng, Muxi Chen, Lei Zhang, Qiang Xu
AAAI 2023 (Oral)
This work questions the effectiveness of emerging favored Transformer-based solutions for the long- term time series forecasting problem. We propose a simple but effective linear model LTSF-Linear as a baseline to verify our claim.
[paper] [code]

DeepSAT:An EDA-Driven Learning Framework for SAT.

Min Li, Zhengyuan Shi, Qiuxia Lai, Sadaf Khan, Shaowei Cai, Qiang Xu
Arxiv preprint
DeepSAT is an end-to-end learning framework for the Boolean satisfiability problem that uses advanced logic synthesis algorithms and a DAG-based GNN to approximate the SAT solving procedure.
[paper] [code]

SCINet:Time Series Modeling and Forecasting with Sample Convolution and Interaction

Liu, M., Zeng, A., Lai, Q., & Xu, Q.
NeurIPS 2022
We propose a novel and general CNN-based backbone for the time series forecasting problem, achieving SOTAs on several real-world datasets.

SmoothNet:A Plug-and-Play Network for Refining Human Poses in Videos.

Ailing Zeng, Lei Yang, Xuan Ju, Jiefeng Li, Jianyi Wang, Qiang Xu
ECCV 2022
SmoothNet is a refinement network, which can be attached to any existing pose estimators to improve their temporal smoothness and enhance the per-frame precision [[Website](].
[paper] [code]

DeciWatch:A Simple Baseline for 10× Efficient 2D and 3D Pose Estimation.

Ailing Zeng, Xuan Ju, Lei Yang, Ruiyuan Gao, Xizhou Zhu, Bo Dai, Qiang Xu
ECCV 2022
DeciWatch is a simple baseline framework with the sample-denoise-recover scheme for video-based 2D/3D human pose estimation that can achieve 10 times efficiency improvement over existing works without any performance degradation [[Website](].
[paper] [code]

Out-of-Distribution Detection with Semantic Mismatch under Masking.

Yijun Yang, Ruiyuan Gao, Qiang Xu
ECCV 2022
MoodCat is an out-of-distribution (OOD) detection framework, which identify OOD sample by calculating the semantic difference between the original image and its synthesized version conditioned on a random mask and the predicted label.
[paper] [code]

Be Your Own Neighborhood:Detecting Adversarial Example by the Neighborhood Relations Built on Self-Supervised Learning.

Zhiyuan He, Yijun Yang, Pin-Yu Chen, Qiang Xu, Tsung-Yi Ho
ECCV 2022 (WorkShop)
We take advantage of SSL model’s robust representation capacity to identify AE by referring to its neighbours.

DeepTPI:Test Point Insertion with Deep Reinforcement Learning.

Zhengyuan Shi, Min Li, Sadaf Khan, Liuzheng Wang, Naixing Wang, Yu Huang, Qiang Xu
IEEE International Test Conference (ITC 2022)
DeepTPI is a RL-based test point insertion approach for VLSI testing, which equips the pre-trained DeepGate and improves test coverage.
[paper] [code]

DeepGate:Learning Neural Representations of Logic Gates.

Min Li, Sadaf Khan, Zhengyuan Shi, Naixing Wang, Huang Yu, Qiang Xu
DAC 2022
DeepGate is a representation learning model for electronic design automation (EDA) that effectively embeds rich information of each gate in circuit and is suitable for many downstream tasks.
[paper] [code]

T-WaveNet:Tree-Structured Wavelet Neural Network for Sensor-Based Time Series Analysis.

Minhao Liu, Ailing Zeng, Qiuxia Lai, Qiang Xu
ICLR 2022
We introduce a tree-structured wavelet deep neural network to effectively extract more discriminative and expressive feature representations in time series signals.
[paper] [code]

What You See in Not What the Network Infers:Detecting Adversarial Examples Based on Semantic Contradiction.

Yijun Yang, Ruiyuan Gao, Yu Li, Qiuxia Lai, Qiang Xu
NDSS 2022
We figure out the contradiction capsuled in AEs, i.e. their semantic information is inconsistent with the discriminative features extracted by the victim DNN model. By leveraging above contradiction, we propose a generative model based AE detection framework, namely, ContraNet, which outperforms existing methods by a large margin, especially under adaptive attacks.
[paper] [code]

Learning Skeletal Graph Neural Networks for Hard 3D Pose Estimation

Ailing Zeng, Xiao Sun, Lei Yang, Nanxuan Zhao, Minhao Liu, Qiang Xu
2021 International Conference on Computer Vision (ICCV’21)
[paper] [code]

Hop-Aware Dimension Optimization for Graph Neural Networks

Ailing Zeng, Minhao Liu, Zhiwei Liu, Ruiyuan Gao, Qiang Xu

AppealNet:An Efficient and Highly-Accurate Edge/Cloud Collaborative Architecture for DNN Inference

Min Li, Yu Li, Ye Tian, Li Jiang, Qiang Xu
ACM/IEEE Design Automation Conference (DAC), 2021

Skimming and Scanning for Untrimmed Video Action Recognition

Yunyan Hong*, Ailing Zeng*, Min Li, Cewu Lu, Li Jiang, Qiang Xu
2021 14th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)

MixDefense:A Defense-in-Depth Framework for Adversarial Example Detection Based on Statistical and Semantic Analysis

Yijun Yang, Ruiyuan Gao, Yu Li, Qiuxia Lai, Qiang Xu

DeepDyve:Dynamic Verification for Deep Neural Networks

Yu Li*, Min Li*, Bo Luo, Ye Tian, Qiang Xu (co-first authors)
2020 ACM SIGSAC Conference on Computer and Communications Security (CCS’20)

Srnet:Improving generalization in 3d human pose estimation with a split-and-recombine approach

Ailing Zeng, Xiao Sun, Fuyang Huang, Minhao Liu, Qiang Xu, Stephen Lin
2020 European Conference on Computer Vision (ECCV’20)
[paper] [code]

D2NN:a fine-grained dual modular redundancy framework for deep neural networks

Yu Li, Yannan Liu, Min Li, Ye Tian, Bo Luo, Qiang Xu
35th Annual Computer Security Applications Conference (ACSAC’19)