Author name cluster

Ping Wei

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

15 papers

2 author rows

IS Journal 2026 Journal Article

Learning Motion-Scene Disentanglement for Trajectory Prediction From Videos

Haowen Tang
Ping Wei
Ziyang Ren
Huan Li

Agent trajectory prediction plays a significant role in various intelligent systems, as it entails the accurate anticipation of the future trajectory based on the historical data. Conventional approaches often rely on ready-made trajectory coordinates as inputs, which remain inapplicable in video-based scenarios. While two-stage trajectory prediction methods based on tracking-prediction paradigms have made progress in predicting trajectories, they still suffer from the information degradation and error accumulation due to the independence stages. In this article, we propose an end-to-end model (MSDN) to directly predict future trajectories from videos. We design novel disentanglement structures to explicitly learn the motion-aware and scene-aware representations from videos to avoid information degradation. To alleviate error accumulation, we propose the temporal consistency learning structure. Extensive experiments on the ETH-UCY dataset and the Stanford Drone Dataset demonstrate that the proposed model not only outperforms the previous approaches, but also significantly enhances the inference speed.

Details DOI

IS Journal 2026 Journal Article

Physical Relation Reasoning for 3-D Object Detection

Jialu Qin
Ping Wei

3-D object detection is an important problem in many intelligent system applications. Based on powerful spatial information provided by point clouds, existing methods focus primarily on the intrinsic geometric properties of objects while neglecting the physical relationships and interactions among the objects. This may lead to physically unreasonable predictions, such as floating objects or object volume overlaps. In this article, we propose a novel 3-D object detection method from the perspective of physical relation reasoning. Specifically, we introduce two aspects of physical relations, including stability and volume exclusion. In addition, we introduce room layouts to assist in 3-D object detection and formulate two physical constraints on the basic of volume exclusion and stability to ensure that all objects conform to real-world physics constraints. We validate our proposed model on ScanNetV2 and SUN RGB-D datasets, and the results demonstrate the effectiveness.

Details DOI

ICLR Conference 2025 Conference Paper

Alchemy: Amplifying Theorem-Proving Capability Through Symbolic Mutation

Shaonan Wu
Shuai Lu
Yeyun Gong
Nan Duan
Ping Wei

Formal proofs are challenging to write even for experienced experts. Recent progress in Neural Theorem Proving (NTP) shows promise in expediting this process. However, the formal corpora available on the Internet are limited compared to the general text, posing a significant data scarcity challenge for NTP. To address this issue, this work proposes Alchemy, a general framework for data synthesis that constructs formal theorems through symbolic mutation. Specifically, for each candidate theorem in Mathlib, we identify all invocable theorems that can be used to rewrite or apply to it. Subsequently, we mutate the candidate theorem by replacing the corresponding term in the statement with its equivalent form or antecedent. As a result, our method increases the number of theorems in Mathlib by an order of magnitude, from 110k to 6M. Furthermore, we perform continual pretraining and supervised finetuning on this augmented corpus for large language models. Experimental results demonstrate the effectiveness of our approach, achieving a 4.70% absolute performance improvement on Leandojo benchmark. Additionally, our approach achieves a 2.47% absolute performance gain on the out-of-distribution miniF2F benchmark based on the synthetic data. To provide further insights, we conduct a comprehensive analysis of synthetic data composition and the training paradigm, offering valuable guidance for developing a strong theorem prover.

Details

IJCAI Conference 2025 Conference Paper

CFDONEval: A Comprehensive Evaluation of Operator-Learning Neural Network Models for Computational Fluid Dynamics

Menghan Liu
Jianhuan Cen
Ziyang Zhou
Haolong Fan
Hongji Li
Ping Wei
Guohang Peng
Changye He

In this paper, we introduce CFDONEval, a comprehensive evaluation of 12 operator-learning-based neural network (ON) models to simulate 7 benchmark fluid dynamics problems. These problems cover a range of 2D scenarios, including Darcy flow, two-phase flow, Taylor-Green vortex, lid-driven cavity flow, tube flow, circular cylinder flow, and 3D periodic hill flow. For a rigorous evaluation, we establish 22 fluid dynamics datasets for these benchmark problems, 18 of which are newly generated using traditional numerical methods, such as the finite element method. Our evaluation tackles 5 key challenges: multiscale phenomena, convection dominance, long-term predictions, multiphase flows, and unstructured meshes over complex geometries. We assess computational accuracy, efficiency, and flow field visualization, offering valuable insights into the application of ON models in fluid dynamics research. Our findings show that attention-based models perform well in handling almost all challenges; models with a U-shaped structure excel in handling multiscale problems; and the NU-FNO model demonstrates the smallest relative error in L2 norm when processing nonuniform grid data. The related code, dataset, and appendix are publicly available at: https: //github. com/Sysuzqs/CFDNNEval.

PDF Details DOI

ECAI Conference 2025 Conference Paper

GCQ-ViT: Group-Aware Collaborative Post-Training Quantization for Vision Transformers

Pan Peng 0006
Wenbin Guo
Ping Wei
Wei Zhou 0011

Post-training quantization (PTQ) is widely utilized in Vision Transformers (ViTs) for its computational efficiency and retraining elimination. However, the unique architecture of ViTs introduces significant quantization challenges. Dynamic fluctuations in channel activations, particularly post-LayerNorm, result in distributional mismatches. Additionally, the heavy-tailed nature of post-Softmax activations compromises the accurate representation of critical attention regions, vital for ViT performance. Moreover, weight quantization at low bit-widths leads to a loss of structural information, degrading global feature representation. To address these challenges, we introduce the Group-aware Collaborative Quantization framework (GCQ-ViT), which significantly improves both the accuracy and efficiency of ViT quantization. The GCQ-ViT framework integrates a novel dynamic perception grouping quantization mechanism to ensure distributional consistency within groups, thus reducing hardware expense. It also utilizes a self-adaptive displaced uniform log2 quantizer, optimizing shift factors and nonlinear intervals to enhance representation in high-density regions of post-Softmax activations. Additionally, we propose a dynamic dimension-aware error compensation method to correct quantization errors across channel dimensions using a residual mean compensation skill, ensuring robust feature preservation. Extensive experiments on image classification, object detection, and instance segmentation tasks demonstrate that GCQ-ViT outperforms the current leading PTQ methods, setting a new benchmark for ViT quantization.

Details

AAAI Conference 2025 Conference Paper

OTIAS: OcTree Implicit Adaptive Sampling for Multispectral and Hyperspectral Image Fusion

Shangqi Deng
Jun Ma
Liang-Jian Deng
Ping Wei

Implicit Neural Representation (INR) methods have demonstrated great potential in arbitrary-scale super-resolution tasks. This success is primarily due to their ability to continuously represent images using coordinates. In the task of remote sensing image fusion, INR methods have also shown promising applications. However, the previous INR methods neglect channel-wise modeling, while sharing a single kernel across all channels at each position, resulting in a lack of sensitivity to data specificity. To address these issues, we propose the OcTree Implicit Adaptive Sampling (OTIAS) method, which innovatively applies the octree structure to restore data from both horizontal and vertical directions, effectively incorporating spatial and spectral information from hyperspectral data. Additionally, we introduce a novel method to adaptively generate interpolation kernels based on coordinates. This approach efficiently produces customized interpolation kernel parameters for octree nodes, tailored to different spectral information. Overall, our method achieves state-of-the-art performance on the CAVE and Harvard datasets with 4× and 8× scaling factors, outperforming existing approaches.

PDF Details DOI

AAAI Conference 2025 Conference Paper

Unveiling Multi-View Anomaly Detection: Intra-view Decoupling and Inter-view Fusion

Kai Mao
Yiyang Lian
Yangyang Wang
Meiqin Liu
Nanning Zheng
Ping Wei

Anomaly detection has garnered significant attention for its extensive industrial application value. Most existing methods focus on single-view scenarios and fail to detect anomalies hidden in blind spots, leaving a gap in addressing the demands of multi-view detection in practical applications. Ensemble of multiple single-view models is a typical way to tackle the multi-view situation, but it overlooks the correlations between different views. In this paper, we propose a novel multi-view anomaly detection framework, Intra-view Decoupling and Inter-view Fusion (IDIF), to explore correlations among views. Our method contains three key components: 1) a proposed Consistency Bottleneck module extracting the common features of different views through information compression and mutual information maximization; 2) an Implicit Voxel Construction module fusing features of different views with prior knowledge represented in the form of voxels; and 3) a View-wise Dropout training strategy enabling the model to learn how to cope with missing views during test. The proposed IDIF achieves state-of-the-art performance on three datasets. Extensive ablation studies also demonstrate the superiority of our methods.

PDF Details DOI

IJCAI Conference 2024 Conference Paper

PDENNEval: A Comprehensive Evaluation of Neural Network Methods for Solving PDEs

Ping Wei
Menghan Liu
Jianhuan Cen
Ziyang Zhou
Liao Chen
Qingsong Zou

The rapid development of neural network (NN) methods for solving partial differential equations (PDEs) has created an urgent need for evaluation and comparison of these methods. In this study, we propose PDENNEval, a comprehensive and systematic evaluation of 12 NN methods for PDEs. These methods are classified into function learning type and operator learning type based on their different mathematical foundations. The evaluation is implemented using a diverse dataset comprising 19 distinct PDE problems selected from various scientific fields such as fluid, materials, finance, and electromagnetic. Several evaluation results are reported, aiming to provide guidance for further research in this field. Our code and data are publicly available at https: //github. com/zhouzy36/PDENNEval.

PDF Details DOI

YNIMG Journal 2022 Journal Article

Expectations of immediate and delayed reward differentially affect cognitive task performance

Yachao Rong
Ningxuan Chen
Jiarui Dong
Qi Li
Xiaodong Yue
Li Hu
Ping Wei

The current study used a modified Monetary Incentive Delay task to examine the neural mechanisms underlying anticipating and receiving an immediate or delayed reward and examined the influence of pursuing these rewards on cognitive task performance. A pre-cue indicating the potential of gaining a monetary reward (immediate-, delayed-, vs. no-reward) was followed by a target stimulus requiring a fast and accurate response. Then, response-contingent feedback was presented indicating whether or not the participant would receive the corresponding reward. Linear mixed-effect models revealed the fastest behavioural responses and the strongest neural activity, as reflected in event-related-potentials and event-related-spectral-perturbation responses, for immediate reward, followed by delayed reward, with the slowest behavioural responses and the weakest neural activities observed in the no-reward condition. Expectations related to the cue-P3 component and the cue-delta activities predicted behavioural performance, especially in the immediate reward condition. Moreover, exploratory analyses revealed that depression moderated the relationship between target-locked neural activity and behavioural performance in the delayed reward condition, with lower neural activity being related to worse behavioural performance amongst participants scoring high on depression. These results indicate that differential value representations formed through delay discounting directly affect neural responses in reward processing and directly influence the effort invested in the current task, which is reflected by behavioural responses and is in agreement with the expected value of control theory.

Details DOI

YNICL Journal 2021 Journal Article

Neural substrates of deficient cognitive control in individuals with severe internet gaming disorder

Lingxiao Wang
Guochun Yang
Ya Zheng
Zhenghan Li
Ping Wei
Qi Li
Kesong Hu
Xun Liu

BACKGROUND AND AIMS: Internet gaming disorder (IGD) is rapidly becoming a worldwide health concern. The prefrontal-subcortical model of self-regulation emphasizes that an impaired prefrontal cognitive control system and an overwhelming subcortical reward-seeking system are both crucial factors in health problems, including addiction. This study focused on the cognitive control system of IGD, aiming to investigate whether cognitive control is altered and the underlying neural correlates in college students with IGD. METHODS: Thirty college students with IGD and twenty-five matched healthy controls were asked to complete a stop-signal task that measures cognitive control while being monitored by functional magnetic resonance imaging (fMRI). RESULTS: Compared to the controls, only the college students with severe IGD, rather than those with mild IGD, had deficient brain activity involved in inhibitory control and response execution (specifically, the inferior frontal gyrus, anterior cingulate cortex and primary motor cortex); this result implies that cognitive control deficits are closely linked to addiction severity in individuals with IGD. Regarding performance monitoring function, college students with IGD exhibited unabated behavioral and brain activity, as did the control group. CONCLUSIONS: Combined with our previous finding that the subcortical reward system was enhanced in individuals with IGD, the present findings extend the prefrontal-subcortical model of self-regulation from the perspective of IGD in a college student population and thus provide useful insight for the effective prevention and treatment of IGD.

Details DOI

AAAI Conference 2021 Conference Paper

Semantic Consistency Networks for 3D Object Detection

Wenwen Wei
Ping Wei
Nanning Zheng

Detecting 3D objects from point clouds is a significant yet challenging issue in many applications. While most existing approaches seek to leverage geometric information of point clouds, few studies accommodate the inherent semantic characteristics of each point and the consistency between the geometric and semantic cues. In this work, we propose a novel semantic consistency network (SCNet) driven by a natural principle: the class of a predicted 3D bounding box should be consistent with the classes of all the points inside this box. Specifically, our SCNet consists of a feature extraction structure, a detection decision structure, and a semantic segmentation structure. In inference, the feature extraction and the detection decision structures are used to detect 3D objects. In training, the semantic segmentation structure is jointly trained with the other two structures to produce more robust and applicative model parameters. A novel semantic consistency loss is proposed to regulate the output 3D object boxes and the segmented points to boost the performance. Our model is evaluated on two challenging datasets and achieves comparable results to the state-of-the-art methods.

PDF Details

AAAI Conference 2021 Conference Paper

Static-Dynamic Interaction Networks for Offline Signature Verification

Huan Li
Ping Wei
Ping Hu

Offline signature verification is a challenging issue that is widely used in various fields. Previous approaches model this task as a static feature matching or distance metric problem of two images. In this paper, we propose a novel Static- Dynamic Interaction Network (SDINet) model which introduces sequential representation into static signature images. A static signature image is converted to sequences by assuming pseudo dynamic processes in the static image. A static representation extracting deep features from signature images describes the global information of signatures. A dynamic representation extracting sequential features with LSTM networks characterizes the local information of signatures. A dynamic-to-static attention is learned from the sequences to refine the static features. Through the static-to-dynamic conversion and the dynamic-to-static attention, the static representation and dynamic representation are unified into a compact framework. The proposed method was evaluated on four popular datasets of different languages. The extensive experimental results manifest the strength of our model.

PDF Details

IJCAI Conference 2017 Conference Paper

Inferring Human Attention by Learning Latent Intentions

Ping Wei
Dan Xie
Nanning Zheng
Song-Chun Zhu

This paper addresses the problem of inferring 3D human attention in RGB-D videos at scene scale. 3D human attention describes where a human is looking in 3D scenes. We propose a probabilistic method to jointly model attention, intentions, and their interactions. Latent intentions guide human attention which conversely reveals the intention features. This mutual interaction makes attention inference a joint optimization with latent intentions. An EM-based approach is adopted to learn the latent intentions and model parameters. Given an RGB-D video with 3D human skeletons, a joint-state dynamic programming algorithm is utilized to jointly infer the latent intentions, the 3D attention directions, and the attention voxels in scene point clouds. Experiments on a new 3D human attention dataset prove the strength of our method.

PDF Details

YNIMG Journal 2011 Journal Article

Neural correlates of binding features within- or cross-dimensions in visual conjunction search: An fMRI study

Ping Wei
Hermann J. Müller
Stefan Pollmann
Xiaolin Zhou

The fMRI technique was used to investigate the functional neuroanatomy of binding features within- or cross-dimension during visual conjunction search. Participants were asked to perform feature search (FS; e. g. , search for a vertical bar among tilted bars), within-dimension search (WS; e. g. , search for an upright T among non-target oriented Ts and Ls), cross-dimension search (CS; e. g. , search for an orange vertical bar among blue vertical bars and orange tilted bars), and complex search combining within- and cross-dimension features (WCS; e. g. , search for an orange upright T among orange leftward Ts and blue Ls). Reaction times (RTs) taken to decide whether a target was present or absent were faster in the FS than in the WS, CS, and WCS conditions, but did not differ between the latter three conditions. Neuroimaging results revealed a set of fronto-parietal regions, including frontal eye field and intraparietal sulcus, to be consistently activated in conjunction search (WS, CS, and WCS) relative to feature search, suggesting that these regions play a more prominent role in matching visual input against the target template in conjunction search. Furthermore, left occipito-temporal cortex was more activated in within-dimension conjunction search, and bilateral intraparietal sulci were more activated in cross-dimension conjunction search. This suggests that features from the same dimension are ‘bound’ at a higher stage of the ventral pathway by conjoining the inputs from lower-level neurons, whereas neurons along the intraparietal sulcus appear to be necessary for discerning the presence of cross-dimensional conjunctions.

Details DOI

YNIMG Journal 2009 Journal Article

Neural basis of interaction between target presence and display homogeneity in visual search: An fMRI study

Ping Wei
Hermann J. Müller
Stefan Pollmann
Xiaolin Zhou

The functional magnetic resonance imaging (fMRI) technique was used to investigate the functional neuroanatomy of the attention mechanisms employed in visual search with homogeneous or heterogeneous displays. Participants were asked to search for a vertically oriented bar among distractor bars with the same or different orientations, with half of the trials being target-present and the other half being target-absent. Behaviorally, RTs were slower for target-absent than for -present trials when the distractors were heterogeneous, but were faster for target-absent than for -present trials when the distractors were homogeneous. At the neural level, a widely distributed brain network was involved in this interaction. The bilateral frontal eye field, intraparietal sulcus, precentral gyrus and supplementary eye field may play a role in representing the target against distractors and further in detecting and responding to the presence of the target. The right superior frontal gyrus and the bilateral temporal–parietal junction may play a role in filtering distracting information in the search process.

Details DOI