Arrow Research search

Author name cluster

Lei Xu

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.

54 papers
2 author rows

Possible papers

54

EAAI Journal 2026 Journal Article

A lightweight steel surface defect detection model for automotive manufacturing production lines

  • Lu Zheng
  • Chengao Zhu
  • Yingzhi Li
  • Lei Xu
  • Mengting Ruan
  • Mian Wu
  • ChongWei Ruan
  • Jun Tie

Due to the limitations of traditional inspection methods, which are highly subjective and low in accuracy, ensuring steel surface quality is essential for product reliability and operational safety. To address these limitations, this study proposes the Advanced Convolution with Multi-Scale Structure and Pruning You Only Look Once (ACMSP-YOLO) model, a lightweight object-detection model specifically designed for steel-surface defect detection in automotive manufacturing. The model directly employs a Roboflow-based data augmentation strategy to alleviate class imbalance. An average pooling downsampling (ADown) module is integrated into the You Only Look Once version 8 (YOLOv8) framework to compress feature maps while preserving fine defect details. The advanced convolution with multi-scale structure reduces computation while maintaining multiscale feature representation, and the detail-enhanced, shared convolution detection head improves detection robustness with fewer parameters. A lightweight spatial kernel attention mechanism enhances sensitivity to small defects, while the model further uses the Growing Regularization (Growing_Reg) pruning strategy to compress the model size without compromising accuracy. On the Northeastern University Surface Defect Detection (NEU-DET) dataset, ACMSP-YOLO outperformed YOLOv8 by 4. 9%, while reducing parameter count, floating-point operations per second, and model size by 46. 7%, 46. 3%, and 35. 6%, respectively. On the Jetson Orin Nano edge platform, the model achieved 83. 3 frames per second (FPS) under TensorRT configuration, demonstrating its capability for real-time detection in resource-constrained environments. ACMSP-YOLO achieves an effective balance between detection accuracy and computational efficiency, rendering it highly suitable for industrial deployment.

Details DOI

EAAI Journal 2026 Journal Article

Consistency and consensus-based decision-making for probabilistic linguistic information reliability

  • Yan Chen
  • Dan Li
  • Lin Liu
  • Xiao Wang
  • Lei Xu

Probabilistic linguistic information is a new decision-making tool that can effectively handle uncertain information in the decision-making process. In the group decision-making process based on probabilistic linguistic information, the PageRank algorithm based on social network expert weight calculation, the reliability of individual experts and the reliability of the group decision-making are rarely considered. This research mainly addresses this series of questions. Four processes are included: the operation laws for linguistic term sets, the consistency analysis process of the probabilistic linguistic preference matrix, expert weights based on the improved PageRank algorithm and the process of expert reliability analysis. In the first process, a new algorithm is proposed in this paper, which solves the problem of closure and transferability of operations between linguistic term sets. In the second process, the concept of the derived matrix is defined, and combined with the concept of the reachability matrix, the method of judging the complete consistency and satisfactory consistency of the probabilistic linguistic preference matrix is given. In the third process, the concept of dual trust propagation is defined and combined with the trust relationship between experts and the PageRank algorithm to calculate the expert weights. The concept of two-level consensus and a consensus threshold are introduced in the meanwhile, which ensures the quality of decision-making. According to the properties of the derived matrix, a ranking method of alternatives is given by solving the eigenvector of the maximum eigenvalue. In the last process, the concept of decision-making reliability is presented, and the calculation method of expert reliability and group decision-making reliability are given. The effectiveness and practicality of the method given in this paper are illustrated by numerical examples.

Details DOI

EAAI Journal 2026 Journal Article

Damage assessment of thermal-humidity-mechanical coupling field of early-age concrete based on adaptive physics informed neural network

  • Shiqi Wang
  • Yue Chen
  • Jinlong Liu
  • Fangzhou Lin
  • Lei Xu

The crack-damage resistance of early-age concrete is affected by multiple factors such as hydration, self-drying, temperature and humidity diffusion, and material properties, which are difficult to be accurately evaluated by traditional theories and numerical models. This paper proposed an adaptive physics-informed back propagation neural network (BPINN) to accurately evaluate the damage of early-age concrete under multi-physics field coupling. The temperature and humidity diffusion and shrinkage models are used as physics loss functions to guide the model in learning the physics laws. Furthermore, time-dependent factor weights are constructed for both the physics and boundary equations to enhance the model's ability to learn the spatiotemporal feature distribution of the sampling points. BPINN effectively simulates the influence of concrete strength grade and boundary conditions on temperature and humidity diffusion, with the average error less than 5 %. The LOSS differences of traditional physics informed neural network (PINN) and BPINN in time step, activation function, hidden layer and neuron number are quantified. Compared with the traditional PINN, the LOSS of BPINN is reduced by 62. 4 %. On this basis, the predictive performance of BPINN and four types of data-driven models is compared to verify the influence of physics constraint, as BPINN has the smallest statistical loss and data discreteness. The model proposed in this paper enhances the learning ability of spatial-temporal features by balancing the weight between boundary and physics equations, providing new insights for the thermo-hygro-mechanical coupling field in early-age concrete.

Details DOI

AAAI Conference 2026 Conference Paper

DEPO: Dual-Efficiency Preference Optimization for LLM Agents

  • Sirui Chen
  • Mengshi Zhao
  • Lei Xu
  • Yuying Zhao
  • Beier Zhu
  • Hanwang Zhang
  • Shengjie Zhao
  • Chaochao Lu

Recent advances in large language models (LLMs) have greatly improved their reasoning and decision-making abilities when deployed as agents. Richer reasoning, however, often comes at the cost of longer chain of thought (CoT), hampering interaction efficiency in real-world scenarios. Nevertheless, there still lacks systematic definition of LLM‑Agent efficiency, hindering targeted improvements. To this end, we introduce dual‑efficiency, comprising (i) step-level efficiency, which minimizes tokens per step, and (ii) trajectory-level efficiency, which minimizes the number of steps to complete a task. Building on this definition, we propose DEPO, a dual-efficiency preference‑based optimization method that jointly rewards succinct responses and fewer action steps. Experiments on WebShop and BabyAI show that DEPO cuts token usage by up to 60.9% and steps by up to 26.9%, while achieving up to a 29.3% improvement in task performance. DEPO also generalizes to three out-of-domain math benchmarks and retains its efficiency gains when trained on only 25% of the data.

PDF Details DOI

EAAI Journal 2026 Journal Article

Explainable multiscale representation learning for anticancer peptide prediction

  • Yongqing Zhang
  • Xinyu Mao
  • Yuhang Liu
  • Wang Li
  • Zhigan Zhou
  • Yugui Xu
  • Jin Wu
  • Quan Zou

Anticancer peptides (ACPs) exhibit significant pharmacological potential in studying cancer and other diseases. Although anticancer peptides have been extensively investigated, many remain undiscovered and hold promise as future therapeutics. Computational methods offer efficient solutions for identifying novel anticancer peptides, but current approaches often rely heavily on local amino acid sequence information, overlooking global interactions, which limits their performance and generalizability. Here, we present an explainable multiscale representation learning framework that integrates sequence information from bidirectional encoder and handcrafted features to enhance anticancer peptides predictions. Our model learns sequence representations at two scales: it captures local representations through handcrafted feature methods at residue scale and global representations through a bidirectional encoder at protein scale. In addition, it employs a cross-attention mechanism to fuse these two types of representations automatically. Compared to state-of-the-art techniques, our model demonstrates superior performance with a significant improvement of 2. 3%, 1. 3%, 2. 5% and 3. 2% in prediction accuracy. To further demonstrate the superiority of our model, it can achieve optimal performance even under conditions where negative samples are randomly generated in the alternate dataset and other datasets, resulting in a significant improvement in accuracy over current state-of-the-art techniques. Moreover, model provides “white-box” prediction, revealing representation shifts during inference and effectively identifying important subregions of sequences, thereby uncovering sequence motifs. Shapley additive explanation value analysis was performed on the fused features, revealing that certain features contribute significantly to the model’s predictions. In general, our model is a powerful tool for advancing artificial intelligence-driven drug discovery.

Details DOI

AAAI Conference 2026 Conference Paper

Full-Atom Peptide Design via Riemannian–Euclidean Bayesian Flow Networks

  • Hao Qian
  • Shikui Tu
  • Lei Xu

Diffusion and flow matching models have recently emerged as promising approaches for peptide binder design. Despite their progress, these models still face two major challenges. First, categorical sampling of discrete residue types collapses their continuous parameters into one-hot assignments, while continuous variables (e.g., atom positions) evolve smoothly throughout the generation process. This mismatch disrupts the update dynamics and results in suboptimal performance. Second, current models assume unimodal distributions for side-chain torsion angles, which conflicts with the inherently multimodal nature of side-chain rotameric states and limits prediction accuracy. To address these limitations, we introduce PepBFN, the first Bayesian flow network for full-atom peptide design that directly models parameter distributions in fully continuous space. Specifically, PepBFN models discrete residue types by learning their continuous parameter distributions, enabling joint and smooth Bayesian updates with other continuous structural parameters. It further employs a novel Gaussian mixture-based Bayesian flow to capture the multimodal side-chain rotameric states and a Matrix Fisher-based Riemannian flow to directly model residue orientations on the SO(3) manifold. Together, these parameter distributions are progressively refined via Bayesian updates, yielding smooth and coherent peptide generation. Experiments on side-chain packing, reverse folding, and binder design tasks demonstrate the strong potential of PepBFN in computational peptide design.

PDF Details DOI

AAAI Conference 2026 Conference Paper

SEBSFormer: A Spectral-Enhanced Bi-Stream Transformer for Robust EEG Decoding

  • Lin Zhang
  • Shikui Tu
  • Lei Xu

Electroencephalography (EEG) plays a vital role in clinical and cognitive applications such as epilepsy diagnosis and emotion recognition. However, the low signal-to-noise ratio, inter-subject variability, and inherent non-stationarity of EEG signals present substantial modeling challenges. While recent Transformer-based models offer promising long-range modeling capabilities, their self-attention mechanism behaves as a low-pass filter, suppressing high-frequency neural patterns critical for decoding transient events. In this work, we provide the first formal analysis demonstrating this low-pass behavior in self-attention mechanisms when applied to EEG signals, revealing a fundamental limitation of deep attention-based EEG models. To address this, we propose SEBSFormer, a spectral-enhanced bi-Stream Transformer that jointly models temporal dependencies and spectral structures. SEBSFormer integrates three key modules: a spectral compensation module that restores high-frequency components via residual correction in the Fourier domain; a multi-scale temporal attention module for saliency-guided temporal compression; and a graph-guided dynamic fusion module for adaptive spatial aggregation across electrodes. Extensive experiments on three benchmark datasets—TUAB, TUEV, and SEED—demonstrate that SEBSFormer consistently outperforms existing state-of-the-art models across both clinical and affective tasks. Our findings establish a new paradigm for frequency-aware EEG modeling.

PDF Details DOI

EAAI Journal 2025 Journal Article

Assessment of hybrid kernel function in extreme support vector regression model for streamflow time series forecasting based on a bayesian estimator decomposition algorithm

  • Peng Shi
  • Lei Xu
  • Simin Qu
  • Hongshi Wu
  • Qiongfang Li
  • Yiqun Sun
  • Xiaoqiang Yang
  • Wei Gao

Diverse decomposition algorithms have been widely employed to streamflow time series forecasting. Their applications, however, are hindered by the plausible high accuracy in the overall decomposition-based framework. This paper firstly introduces a novel decomposition algorithm named Bayesian estimator of abrupt change, seasonality and trend (BEAST) into streamflow forecasting to alleviate the boundary effect. Practical samples are generated under the modified two-stage decomposition prediction (TSDP) framework. A hybrid kernel function, which benefits from two different standalone ones, is designed for kernel extreme support vector regression and the HKESVR model is trained on the samples using 10-fold cross-validation strategy. Comparative experiments are conducted on three monthly streamflow series from basins with diverse hydroclimatic conditions. The results in different lead times (1-, 3-, and 5-month-ahead) show that the BEAST algorithm imposes an average improvement of 5. 14% and 12. 25% for the root-mean-square error and Nash-Sutcliffe efficiency coefficient respectively on the standalone models and shares a comprehensive similar performance on the mean absolute percentage error. And the nonparametric test results reveal that the BEAST method shows a significant improvement on the comprehensive performance compared with a conventional decomposition method. By contrast, the differences between machine learning models are much smaller. The hybrid kernel function works well in some specific cases in which the standalone kernel function fails. The hybrid BEAST-HKESVR is reliable enough to rank the second place among the fifteen tested models. Finally, the effects of hyperparameters in the BEAST algorithm are discussed and relevant suggestions on them are provided.

Details DOI

ICLR Conference 2025 Conference Paper

Beyond Surface Structure: A Causal Assessment of LLMs' Comprehension ability

  • Yujin Han
  • Lei Xu
  • Sirui Chen
  • Difan Zou
  • Chaochao Lu

Large language models (LLMs) have shown remarkable capability in natural language tasks, yet debate persists on whether they truly comprehend deep structure (i.e., core semantics) or merely rely on surface structure (e.g., presentation format). Prior studies observe that LLMs' performance declines when intervening on surface structure, arguing their success relies on surface structure recognition. However, surface structure sensitivity does not prevent deep structure comprehension. Rigorously evaluating LLMs' capability requires analyzing both, yet deep structure is often overlooked. To this end, we assess LLMs' comprehension ability using causal mediation analysis, aiming to fully discover the capability of using both deep and surface structures. Specifically, we formulate the comprehension of deep structure as direct causal effect (DCE) and that of surface structure as indirect causal effect (ICE), respectively. To address the non-estimability of original DCE and ICE --- stemming from the infeasibility of isolating mutual influences of deep and surface structures, we develop the corresponding quantifiable surrogates, including approximated DCE (ADCE) and approximated ICE (AICE). We further apply the ADCE to evaluate a series of mainstream LLMs (and the one with random weights), showing that most of them exhibit deep structure comprehension ability, which grows along with the prediction accuracy. Comparing ADCE and AICE demonstrates closed-source LLMs (e.g., GPT) rely more on deep structure, while open-source LLMs (e.g., Llama) are more surface-sensitive, which decreases with model scale. Theoretically, ADCE is a bidirectional evaluation, which measures both the sufficiency and necessity of deep structure changes in causing output variations, thus offering a more comprehensive assessment than accuracy, a common evaluation in LLMs. Our work provides new insights into LLMs' deep structure comprehension and offers novel methods for LLMs evaluation. The code for our project is available at [ADCE Project](https://github.com/OpenCausaLab/ADCE).

Details

TCS Journal 2025 Journal Article

Bribery in elections with randomly selected voters: Hardness and algorithm

  • Liangde Tao
  • Lin Chen
  • Lei Xu
  • Weidong Shi
  • Md Mahabub Uz Zaman
  • Ahmed Sunny

Many research works in computational social choice assume a fixed set of voters in an election and study the resistance of different voting rules against electoral manipulation. In recent years, however, a new technique known as random sample voting has been adopted in many multi-agent systems. One of the most prominent examples is blockchain. Many proof-of-stake based blockchain systems like Algorand will randomly select a subset of participants of the system to form a committee, and only the committee members will be involved in the decision of some important system parameters. This can be viewed as running an election where the voter committee (i. e. , the voters whose votes will be counted) is randomly selected. It is generally expected that the introduction of such randomness should make the election more resistant to electoral manipulation, despite the lack of theoretical analysis. In this paper, we present a systematic study on the resistance of an election with a randomly selected voter committee against bribery. Since the committee is randomly generated, by bribing any fixed subset of voters, the designated candidate may or may not win. Consequently, we consider the problem of finding a feasible solution that maximizes the winning probability of the designated candidate. We show that for most voting rules, this problem becomes extremely difficult for the briber as even finding any non-trivial solution with non-zero objective value becomes NP-hard. However, for plurality and veto, there exists a polynomial time approximation scheme that computes a near-optimal solution efficiently. The algorithm builds upon a novel integer programming formulation together with techniques from n-fold integer programming, which may be of separate interest.

Details DOI

AAAI Conference 2025 Conference Paper

CAD-GPT: Synthesising CAD Construction Sequence with Spatial Reasoning-Enhanced Multimodal LLMs

  • Siyu Wang
  • Cailian Chen
  • Xinyi Le
  • Qimin Xu
  • Lei Xu
  • Yanzhou Zhang
  • Jie Yang

Computer-aided design (CAD) significantly enhances the efficiency, accuracy, and innovation of design processes by enabling precise 2D and 3D modeling, extensive analysis, and optimization. Existing methods for creating CAD models rely on latent vectors or point clouds, which are difficult to obtain, and storage costs are substantial. Recent advances in Multimodal Large Language Models (MLLMs) have inspired researchers to use natural language instructions and images for CAD model construction. However, these models still struggle with inferring accurate 3D spatial location and orientation, leading to inaccuracies in determining the spatial 3D starting points and extrusion directions for constructing geometries. This work introduces CAD-GPT, a CAD synthesis method with spatial reasoning-enhanced MLLM that takes either a single image or a textual description as input. To achieve precise spatial inference, our approach introduces a 3D Modeling Spatial Mechanism. This method maps 3D spatial positions and 3D sketch plane rotation angles into a 1D linguistic feature space using a specialized spatial unfolding mechanism, while discretizing 2D sketch coordinates into an appropriate planar space to enable precise determination of spatial starting position, sketch orientation, and 2D sketch coordinate translations. Extensive experiments demonstrate that CAD-GPT consistently outperforms existing state-of-the-art methods in CAD model synthesis, both quantitatively and qualitatively.

PDF Details DOI

AAAI Conference 2025 Conference Paper

CriSPO: Multi-Aspect Critique-Suggestion-guided Automatic Prompt Optimization for Text Generation

  • Han He
  • Qianchu Liu
  • Lei Xu
  • Chaitanya Shivade
  • Yi Zhang
  • Sundararajan Srinivasan
  • Katrin Kirchhoff

Existing automatic prompt engineering methods are typically designed for discriminative tasks, where new task prompts are iteratively refined with limited feedback from a single metric reflecting a single aspect. However, these approaches are suboptimal for generative tasks, which require more nuanced guidance beyond a single numeric metric to improve the prompt and optimize multiple aspects of the generated text. To address these challenges, we propose a novel multi-aspect Critique-Suggestion-guided automatic Prompt Optimization (CriSPO) approach. CriSPO introduces a critique-suggestion module as its core component. This module spontaneously discovers aspects, and compares generated and reference texts across these aspects, providing specific suggestions for prompt modification. These clear critiques and actionable suggestions guide a receptive optimizer module to make more substantial changes, exploring a broader and more effective search space. To further improve CriSPO with multi-metric optimization, we introduce an Automatic Suffix Tuning (AST) extension to enhance the performance of task prompts across multiple metrics. We evaluate CriSPO on 4 state-of-the-art Large Language Models (LLMs) across 4 summarization and 5 Question Answering (QA) datasets. Extensive experiments show 3-4% ROUGE score improvement on summarization and substantial improvement of various metrics on QA.

PDF Details DOI

JBHI Journal 2025 Journal Article

Decoding SSVEP Via Calibration-Free TFA-Net: A Novel Network Using Time-Frequency Features

  • Lei Xu
  • Xinyi Jiang
  • Ruimin Wang
  • Pan Lin
  • Yuankui Yang
  • Yue Leng
  • Wenming Zheng
  • Sheng Ge

Brain-computer interfaces (BCIs) based on steady-state visual evoked potential (SSVEP) signals offer high information transfer rates and non-invasive brain-to-device connectivity, making them highly practical. In recent years, deep learning techniques, particularly convolutional neural network (CNN) architectures, have gained prominence in EEG (e. g. , SSVEP) decoding because of their nonlinear modeling capabilities and autonomy from manual feature extraction. However, most studies using CNNs employ temporal signals as the input and cannot directly mine the implicit frequency information, which may cause crucial frequency details to be lost and challenges in decoding. By contrast, the prevailing supervised recognition algorithms rely on a lengthy calibration phase to enhance algorithm performance, which could impede the popularization of SSVEP based BCIs. To address these problems, this study proposes the Time-Frequency Attention Network (TFA-Net), a novel CNN model tailored for SSVEP signal decoding without the calibration phase. Additionally, we introduce the Frequency Attention and Channel Recombination modules to enhance ability of TFA-Net to infer finer frequency-wise attention and extract features efficiently from SSVEP in the time-frequency domain. Classification results on a public dataset demonstrated that the proposed TFA-Net outperforms all the compared models, achieving an accuracy of 79. 00% $\pm$ 0. 27% and information transfer rate of 138. 82 $\pm$ 0. 78 bits/min with a 1-s data length. TFA-Net represents a novel approach to SSVEP identification as well as time-frequency signal analysis, offering a calibration-free solution that enhances the generalizability and practicality of SSVEP based BCIs.

Details DOI

NeurIPS Conference 2025 Conference Paper

Dual Data Alignment Makes AI-Generated Image Detector Easier Generalizable

  • Ruoxin Chen
  • Junwei Xi
  • Zhiyuan Yan
  • Ke-Yue Zhang
  • Shuang Wu
  • Jingyi Xie
  • Xu Chen
  • Lei Xu

The rapid increase in AI-generated images (AIGIs) underscores the need for detection methods. Existing detectors are often trained on biased datasets, leading to overfitting on spurious correlations between non-causal image attributes and real/synthetic labels. While these biased features enhance performance on the training data, they result in substantial performance degradation when tested on unbiased datasets. A common solution is to perform data alignment through generative reconstruction, matching the content between real and synthetic images. However, we find that pixel-level alignment alone is inadequate, as the reconstructed images still suffer from frequency-level misalignment, perpetuating spurious correlations. To illustrate, we observe that reconstruction models restore the high-frequency details lost in real images, inadvertently creating a frequency-level misalignment, where synthetic images appear to have richer high-frequency content than real ones. This misalignment leads to models associating high-frequency features with synthetic labels, further reinforcing biased cues. To resolve this, we propose Dual Data Alignment (DDA), which aligns both the pixel and frequency domains. DDA generates synthetic images that closely resemble real ones by fusing real and synthetic image pairs in both domains, enhancing the detector's ability to identify forgeries without relying on biased features. Moreover, we introduce two new test sets: DDA-COCO, containing DDA-aligned synthetic images, and EvalGEN, featuring the latest generative models. Our extensive evaluations demonstrate that a detector trained exclusively on DDA-aligned MSCOCO improves across diverse benchmarks. Code is available at https: //github. com/roy-ch/Dual-Data-Alignment.

PDF Details

NeurIPS Conference 2025 Conference Paper

KeeA*: Epistemic Exploratory A* Search via Knowledge Calibration

  • Dengwei Zhao
  • Shikui Tu
  • Yanan Sun
  • Lei Xu

In recent years, neural network-guided heuristic search algorithms, such as Monte-Carlo tree search and A$^\*$ search, have achieved significant advancements across diverse practical applications. Due to the challenges stemming from high state-space complexity, sparse training datasets, and incomplete environmental modeling, heuristic estimations manifest uncontrolled inherent biases towards the actual expected evaluations, thereby compromising the decision-making quality of search algorithms. Sampling exploration enhanced A$^\*$ (SeeA$^\*$) was proposed to improve the efficiency of A$^\*$ search by constructing an dynamic candidate subset through random sampling, from which the expanded node was selected. However, uniform sampling strategy utilized by SeeA$^\*$ facilitates exploration exclusively through the injection of randomness, which completely neglects the heuristic knowledge relevant to open nodes. Moreover, the theoretical support of cluster sampling remains ambiguous. Despite the existence of potential biases, heuristic estimations still encapsulate certain valuable information. In this paper, epistemic exploratory A$^\*$ search (KeeA$^\*$) is proposed to integrate heuristic knowledge for calibrating the sampling process. We first theoretically demonstrate that SeeA$^\*$ with cluster sampling outperforms uniform sampling due to the distribution-aware selection with higher variance. Building on this insight, cluster scouting and path-aware sampling are introduced in KeeA$^\*$ to further exploit heuristic knowledge to increase the sampling mean and variance, respectively, thereby generating higher-quality extreme candidates and enhancing overall decision-making performance. Finally, empirical results on retrosynthetic planning and logic synthesis demonstrate superior performance of KeeA$^*$ compared to state-of-the-art heuristic search algorithms.

PDF Details

AAAI Conference 2025 Conference Paper

Personalized Label Inference Attack in Federated Transfer Learning via Contrastive Meta Learning

  • Hanyu Zhao
  • Zijie Pan
  • Yajie Wang
  • Zuobin Ying
  • Lei Xu
  • Yu-an Tan

Federated Transfer Learning (FTL) is a popular approach to solve the problem of heterogeneous feature space and label distribution. Among the mainstream strategies for FTL, parameter decoupling, which balance the impact of a single global model and multiple personalized models under data heterogeneity, has attracted the attention of many researchers. However, few attacks have been proposed to evaluate the privacy risk of FTL. We find that the fine-tuned structures and the gradient update mechanisms of parameter decoupling would be more likely to leak personalized information for the server to infer private labels. Based on our findings, we propose the label inference attack that combines meta classifier with contrastive learning in FTL. Our experiments show that the proposed attack has ability to extract local personalized information from the differences before and after fine-tuning to improve the accuracy of the attack in the absence of a downstream model. Our research can reveal potential privacy risks in FTL and motivate more research on private and secure FTL.

PDF Details DOI

NeurIPS Conference 2025 Conference Paper

Prior-Guided Flow Matching for Target-Aware Molecule Design with Learnable Atom Number

  • Jingyuan Zhou
  • Hao Qian
  • Shikui Tu
  • Lei Xu

Structure-based drug design (SBDD), aiming to generate 3D molecules with high binding affinity toward target proteins, is a vital approach in novel drug discovery. Although recent generative models have shown great potential, they suffer from unstable probability dynamics and mismatch between generated molecule size and the protein pockets geometry, resulting in inconsistent quality and off-target effects. We propose PAFlow, a novel target-aware molecular generation model featuring prior interaction guidance and a learnable atom number predictor. PAFlow adopts the efficient flow matching framework to model the generation process and constructs a new form of conditional flow matching for discrete atom types. A protein–ligand interaction predictor is incorporated to guide the vector field toward higher-affinity regions during generation, while an atom number predictor based on protein pocket information is designed to better align generated molecule size with target geometry. Extensive experiments on the CrossDocked2020 benchmark show that PAFlow achieves a new state-of-the-art in binding affinity (up to -8. 31 Avg. Vina Score), simultaneously maintains favorable molecular properties.

PDF Details

NeurIPS Conference 2025 Conference Paper

Text to Sketch Generation with Multi-Styles

  • Tengjie Li
  • Shikui Tu
  • Lei Xu

Recent advances in vision-language models have facilitated progress in sketch generation. However, existing specialized methods primarily focus on generic synthesis and lack mechanisms for precise control over sketch styles. In this work, we propose a training-free framework based on diffusion models that enables explicit style guidance via textual prompts and referenced style sketches. Unlike previous style transfer methods that overwrite key and value matrices in self-attention, we incorporate the reference features as auxiliary information with linear smoothing and leverage a style-content guidance mechanism. This design effectively reduces content leakage from reference sketches and enhances synthesis quality, especially in cases with low structural similarity between reference and target sketches. Furthermore, we extend our framework to support controllable multi-style generation by integrating features from multiple reference sketches, coordinated via a joint AdaIN module. Extensive experiments demonstrate that our approach achieves high-quality sketch generation with accurate style alignment and improved flexibility in style control. The official implementation of M3S is available at https: //github. com/CMACH508/M3S.

PDF Details

IJCAI Conference 2024 Conference Paper

Boosting Efficiency in Task-Agnostic Exploration through Causal Knowledge

  • Yupei Yang
  • Biwei Huang
  • Shikui Tu
  • Lei Xu

The effectiveness of model training heavily relies on the quality of available training resources. However, budget constraints often impose limitations on data collection efforts. To tackle this challenge, we introduce causal exploration in this paper, a strategy that leverages the underlying causal knowledge for both data collection and model training. We, in particular, focus on enhancing the sample efficiency and reliability of the world model learning within the domain of task-agnostic reinforcement learning. During the exploration phase, the agent actively selects actions expected to yield causal insights most beneficial for world model training. Concurrently, the causal knowledge is acquired and incrementally refined with the ongoing collection of data. We demonstrate that causal exploration aids in learning accurate world models using fewer data and provide theoretical guarantees for its convergence. Empirical experiments, on both synthetic data and real-world applications, further validate the benefits of causal exploration. The source code is available at https: //github. com/CMACH508/CausalExploration.

PDF Details DOI

EAAI Journal 2024 Journal Article

Inverse parameter identifications and forward strip temperature simulations of the continuous annealing line with physics-informed neural network and operation big data

  • Kai Chen
  • Mingyang Dai
  • Lei Xu
  • Songjiang Xu
  • Xin Xie
  • Xiaoguang Hu
  • Feng Huang
  • Heming Zhang

The continuous annealing line is a key equipment in industrial metal heat treatment. It is a large-thermal-inertia cascade thermal system, and has minute-level time hysteresis between adjustable control parameters and outlet strip temperature, which is challenging for real operation. Large strip temperature deviation from targets will greatly damage the mechanical properties and the surface coating quality of the strip. Traditional methods such as energy balance method, computational fluid dynamics and end-to-end data-driven model are hard to solve the problem due to complex parameter settings, large computing costs and low interpretability, respectively. Here, a novel way seamlessly coupling physical model and data based on physics-informed neural network is used to solve the problem. The mathematical-physical model of the continuous annealing line is developed and a physics-informed neural network model is built to solve inverse problem to identify the heat transfer coefficient of the continuous annealing line. The strip temperature is numerically simulated both with physics-informed neural network model and computational fluid dynamics model. The simulated results are compared with the measured data. The simulated outlet strip temperatures of the two models agree well with the measured data and the accuracy of the identified heat transfer coefficient is verified, where the computational fluid dynamics model has higher accuracy. The physics-informed neural network models developed here will benefit intelligent feedforward cascade control of the continuous annealing line and improve the strip quality in real productions.

Details DOI

AAAI Conference 2024 Conference Paper

Multilevel Attention Network with Semi-supervised Domain Adaptation for Drug-Target Prediction

  • Zhousan Xie
  • Shikui Tu
  • Lei Xu

Prediction of drug-target interactions (DTIs) is a crucial step in drug discovery, and deep learning methods have shown great promise on various DTI datasets. However, existing approaches still face several challenges, including limited labeled data, hidden bias issue, and a lack of generalization ability to out-of-domain data. These challenges hinder the model's capacity to learn truly informative interaction features, leading to shortcut learning and inferior predictive performance on novel drug-target pairs. To address these issues, we propose MlanDTI, a semi-supervised domain adaptive multilevel attention network (Mlan) for DTI prediction. We utilize two pre-trained BERT models to acquire bidirectional representations enriched with information from unlabeled data. Then, we introduce a multilevel attention mechanism, enabling the model to learn domain-invariant DTIs at different hierarchical levels. Moreover, we present a simple yet effective semi-supervised pseudo-labeling method to further enhance our model's predictive ability in cross-domain scenarios. Experiments on four datasets show that MlanDTI achieves state-of-the-art performances over other methods under intra-domain settings and outperforms all other approaches under cross-domain settings. The source code is available at https://github.com/CMACH508/MlanDTI.

PDF Details DOI

EAAI Journal 2024 Journal Article

Response prediction and probabilistic analysis of the vehicle-ballasted track system considering track irregularity based on long-short term memory neural network

  • Hubing Liu
  • Li Song
  • Lei Xu
  • Zhiwu Yu

It is necessary to efficiently and accurately predict the dynamic behavior of the train-track system subjected to massive random track geometric excitations for system evaluation and maintenance. This paper proposed an engineering practical approach for the train-track system behavior prediction by integrating the deep learning surrogate model-based long-short term memory (LSTM) and physical model of the train-ballasted track interaction (TBTI) system, and then the generalized probability density evolution method (GPDEM) is introduced to achieve the probabilistic assessment of the TBTI system under limited track irregularities, where the track irregularity probability model (TIPM) coupled with generalized-F discrepancy points sets is employed to compress the track irregularity sets to reduce the number of system responses (as training target) but without loss of the responses completeness, therefore, it can promote the efficiency and accuracy of training model. The prediction performance of the surrogate model is illustrated by comparing the time-frequency domain information, mean and standard deviation of the responses obtained by the physical model and surrogate model. The results indicated the proposed method can accurately predict the system response, especially the low and medium frequency response below 50 Hz, and improve the efficiency of stochastic analysis for the TBTI system by approximately 20 times. Furthermore, the applicability of the proposed method is further examined by investigating the influence of probability levels of track irregularity and train speeds on train performance. Its extended model is also implemented to predict the response of systems induced by any measured track irregularity in real-time, which promotes the generalization capability of the proposed surrogate model.

Details DOI

NeurIPS Conference 2024 Conference Paper

SeeA*: Efficient Exploration-Enhanced A* Search by Selective Sampling

  • Dengwei Zhao
  • Shikui Tu
  • Lei Xu

Monte-Carlo tree search (MCTS) and reinforcement learning contributed crucially to the success of AlphaGo and AlphaZero, and A$^*$ is a tree search algorithm among the most well-known ones in the classical AI literature. MCTS and A$^*$ both perform heuristic search and are mutually beneficial. Efforts have been made to the renaissance of A$^*$ from three possible aspects, two of which have been confirmed by studies in recent years, while the third is about the OPEN list that consists of open nodes of A$^*$ search, but still lacks deep investigation. This paper aims at the third, i. e. , developing the Sampling-exploration enhanced A$^*$ (SeeA$^*$) search by constructing a dynamic subset of OPEN through a selective sampling process, such that the node with the best heuristic value in this subset instead of in the OPEN is expanded. Nodes with the best heuristic values in OPEN are most probably picked into this subset, but sometimes may not be included, which enables SeeA$^*$ to explore other promising branches. Three sampling techniques are presented for comparative investigations. Moreover, under the assumption about the distribution of prediction errors, we have theoretically shown the superior efficiency of SeeA$^*$ over A$^*$ search, particularly when the accuracy of the guiding heuristic function is insufficient. Experimental results on retrosynthetic planning in organic chemistry, logic synthesis in integrated circuit design, and the classical Sokoban game empirically demonstrate the efficiency of SeeA$^*$, in comparison with the state-of-the-art heuristic search algorithms.

PDF Details DOI

IJCAI Conference 2024 Conference Paper

Self-Supervised Learning for Enhancing Spatial Awareness in Free-Hand Sketches

  • Xin Wang
  • Tengjie Li
  • Sicong Zang
  • Shikui Tu
  • Lei Xu

Free-hand sketch, as a versatile medium of communication, can be viewed as a collection of strokes arranged in a spatial layout to convey a concept. Due to the abstract nature of the sketches, changes in stroke position may make them difficult to recognize. Recently, Graphic sketch representations are effective in representing sketches. However, existing methods overlook the significance of the spatial layout of strokes and the phenomenon of strokes being drawn in the wrong positions is common. Therefore, we developed a self-supervised task to correct stroke placement and investigate the impact of spatial layout on learning sketch representations. For this task, we propose a spatially aware method, named SketchGloc, utilizing multiple graphs for graphic sketch representations. This method utilizes grids for each stroke to describe the spatial layout with other strokes, allowing for the construction of multiple graphs. Unlike other methods that rely on a single graph, this design conveys more detailed spatial layout information and alleviates the impact of misplaced strokes. The experimental results demonstrate that our model outperforms existing methods in both our proposed task and the traditional controllable sketch synthesis task. Additionally, we found that SketchGloc can learn more robust representations under our proposed task setting. The source code is available at https: //github. com/CMACH508/SketchGloc.

PDF Details DOI

IJCAI Conference 2024 Conference Paper

SketchEdit: Editing Freehand Sketches at the Stroke-Level

  • Tengjie Li
  • Shikui Tu
  • Lei Xu

Recent sketch synthesis methods have demonstrated the capability of generating lifelike outcomes. However, these methods directly encode the entire sketches making it challenging to decouple the strokes from the sketches and have difficulty in controlling local sketch synthesis, e. g. , stroke editing. Besides, the sketch editing task encounters the issue of accurately positioning the edited strokes, because users may not be able to draw on the exact position and the same stroke may appear in various locations in different sketches. We propose SketchEdit to realize flexible editing of sketches at the stroke-level for the first time. To tackle the challenge of decoupling strokes, SketchEdit divides a drawing sequence of a sketch into a series of strokes based on the pen state, aligns the stroke segments to have the same starting position, and learns the embeddings of every stroke by a proposed stroke encoder. Moreover, we overcome the problem of stroke placement via a diffusion process, which progressively generates the locations for the strokes to be synthesized, using the stroke features as the guiding condition. Experiments demonstrate that SketchEdit is effective for stroke-level sketch editing and sketch reconstruction. The source code is publicly available at https: //github. com/CMACH508/SketchEdit/.

PDF Details DOI

AAAI Conference 2023 Conference Paper

Efficient Top-K Feature Selection Using Coordinate Descent Method

  • Lei Xu
  • Rong Wang
  • Feiping Nie
  • Xuelong Li

Sparse learning based feature selection has been widely investigated in recent years. In this study, we focus on the l2,0-norm based feature selection, which is effective for exact top-k feature selection but challenging to optimize. To solve the general l2,0-norm constrained problems, we novelly develop a parameter-free optimization framework based on the coordinate descend (CD) method, termed CD-LSR. Specifically, we devise a skillful conversion from the original problem to solving one continuous matrix and one discrete selection matrix. Then the nontrivial l2,0-norm constraint can be solved efficiently by solving the selection matrix with CD method. We impose the l2,0-norm on a vanilla least square regression (LSR) model for feature selection and optimize it with CD-LSR. Extensive experiments exhibit the efficiency of CD-LSR, as well as the discrimination ability of l2,0-norm to identify informative features. More importantly, the versatility of CD-LSR facilitates the applications of the l2,0-norm in more sophisticated models. Based on the competitive performance of l2,0-norm on the baseline LSR model, the satisfactory performance of its applications is reasonably expected. The source MATLAB code are available at: https://github.com/solerxl/Code_For_AAAI_2023.

PDF Details DOI

NeurIPS Conference 2023 Conference Paper

Generalized Weighted Path Consistency for Mastering Atari Games

  • Dengwei Zhao
  • Shikui Tu
  • Lei Xu

Reinforcement learning with the help of neural-guided search consumes huge computational resources to achieve remarkable performance. Path consistency (PC), i. e. , $f$ values on one optimal path should be identical, was previously imposed on MCTS by PCZero to improve the learning efficiency of AlphaZero. Not only PCZero still lacks a theoretical support but also considers merely board games. In this paper, PCZero is generalized into GW-PCZero for real applications with non-zero immediate reward. A weighting mechanism is introduced to reduce the variance caused by scouting's uncertainty on the $f$ value estimation. For the first time, it is theoretically proved that neural-guided MCTS is guaranteed to find the optimal solution under the constraint of PC. Experiments are conducted on the Atari $100$k benchmark with $26$ games and GW-PCZero achieves $198\%$ mean human performance, higher than the state-of-the-art EfficientZero's $194\\%$, while consuming only $25\\%$ of the computational resources consumed by EfficientZero.

PDF Details

IJCAI Conference 2023 Conference Paper

GLPocket: A Multi-Scale Representation Learning Approach for Protein Binding Site Prediction

  • Peiying Li
  • Yongchang Liu
  • Shikui Tu
  • Lei Xu

Protein binding site prediction is an important prerequisite for the discovery of new drugs. Usually, natural 3D U-Net is adopted as the standard site prediction framework to do per-voxel binary mask classification. However, this scheme only performs feature extraction for single-scale samples, which may bring the loss of global or local information, resulting in incomplete, artifacted or even missed predictions. To tackle this issue, we propose a network called GLPocket, which is based on the Lmser (Least mean square error reconstruction) network and utilizes multi-scale representation to predict binding sites. Firstly, GLPocket uses Target Cropping Block (TCB) for targeted prediction. TCB selects the local interested feature from the global representations to perform concentrated prediction, and reduces the volume of feature maps to be calculated by 82% without adding additional parameters. It integrates global distribution information into local regions, making prediction more concentrated on decoding stage. Secondly, GLPocket establishes long-range relationship of patches within the local region with Transformer Block (TB), to enrich local context semantic information. Experiments show that GLPocket improves by 0. 5%-4% on DCA Top-n prediction compared with previous state-of-the-art methods on four datasets. Our code has been released in https: //github. com/CMACH508/GLPocket.

PDF Details DOI

JBHI Journal 2023 Journal Article

HDL: Hybrid Deep Learning for the Synthesis of Myocardial Velocity Maps in Digital Twins for Cardiac Analysis

  • Xiaodan Xing
  • Javier Del Ser
  • Yinzhe Wu
  • Yang Li
  • Jun Xia
  • Lei Xu
  • David Firmin
  • Peter Gatehouse

Synthetic digital twins based on medical data accelerate the acquisition, labelling and decision making procedure in digital healthcare. A core part of digital healthcare twins is model-based data synthesis, which permits the generation of realistic medical signals without requiring to cope with the modelling complexity of anatomical and biochemical phenomena producing them in reality. Unfortunately, algorithms for cardiac data synthesis have been so far scarcely studied in the literature. An important imaging modality in the cardiac examination is three-directional CINE multi-slice myocardial velocity mapping (3Dir MVM), which provides a quantitative assessment of cardiac motion in three orthogonal directions of the left ventricle. The long acquisition time and complex acquisition produce make it more urgent to produce synthetic digital twins of this imaging modality. In this study, we propose a hybrid deep learning (HDL) network, especially for synthetic 3Dir MVM data. Our algorithm is featured by a hybrid UNet and a Generative Adversarial Network with a foreground-background generation scheme. The experimental results show that from temporally down-sampled magnitude CINE images (six times), our proposed algorithm can still successfully synthesise high temporal resolution 3Dir MVM CMR data (PSNR=42. 32) with precise left ventricle segmentation (DICE=0. 92). These performance scores indicate that our proposed HDL algorithm can be implemented in real-world digital twins for myocardial velocity mapping data simulation. To the best of our knowledge, this work is the first one investigating digital twins of the 3Dir MVM CMR, which has shown great potential for improving the efficiency of clinical studies via synthesised cardiac data.

Details DOI

NeurIPS Conference 2023 Conference Paper

Joint Feature and Differentiable $ k $-NN Graph Learning using Dirichlet Energy

  • Lei Xu
  • Lei Chen
  • Rong Wang
  • Feiping Nie
  • Xuelong Li

Feature selection (FS) plays an important role in machine learning, which extracts important features and accelerates the learning process. In this paper, we propose a deep FS method that simultaneously conducts feature selection and differentiable $ k $-NN graph learning based on the Dirichlet Energy. The Dirichlet Energy identifies important features by measuring their smoothness on the graph structure, and facilitates the learning of a new graph that reflects the inherent structure in new feature subspace. We employ Optimal Transport theory to address the non-differentiability issue of learning $ k $-NN graphs in neural networks, which theoretically makes our method applicable to other graph neural networks for dynamic graph learning. Furthermore, the proposed framework is interpretable, since all modules are designed algorithmically. We validate the effectiveness of our model with extensive experiments on both synthetic and real-world datasets.

PDF Details

AAAI Conference 2023 Conference Paper

Linking Sketch Patches by Learning Synonymous Proximity for Graphic Sketch Representation

  • Sicong Zang
  • Shikui Tu
  • Lei Xu

Graphic sketch representations are effective for representing sketches. Existing methods take the patches cropped from sketches as the graph nodes, and construct the edges based on sketch's drawing order or Euclidean distances on the canvas. However, the drawing order of a sketch may not be unique, while the patches from semantically related parts of a sketch may be far away from each other on the canvas. In this paper, we propose an order-invariant, semantics-aware method for graphic sketch representations. The cropped sketch patches are linked according to their global semantics or local geometric shapes, namely the synonymous proximity, by computing the cosine similarity between the captured patch embeddings. Such constructed edges are learnable to adapt to the variation of sketch drawings, which enable the message passing among synonymous patches. Aggregating the messages from synonymous patches by graph convolutional networks plays a role of denoising, which is beneficial to produce robust patch embeddings and accurate sketch representations. Furthermore, we enforce a clustering constraint over the embeddings jointly with the network learning. The synonymous patches are self-organized as compact clusters, and their embeddings are guided to move towards their assigned cluster centroids. It raises the accuracy of the computed synonymous proximity. Experimental results show that our method significantly improves the performance on both controllable sketch synthesis and sketch healing.

PDF Details DOI

JBHI Journal 2023 Journal Article

Multiple Adversarial Learning Based Angiography Reconstruction for Ultra-Low-Dose Contrast Medium CT

  • Weiwei Zhang
  • Zhen Zhou
  • Zhifan Gao
  • Guang Yang
  • Lei Xu
  • Weiwen Wu
  • Heye Zhang

Iodinated contrast medium (ICM) dose reduction is beneficial for decreasing potential health risk to renal-insufficiency patients in CT scanning. Due to the low-intensity vessel in ultra-low-dose-ICM CT angiography, it cannot provide clinical diagnosis of vascular diseases. Angiography reconstruction for ultra-low-dose-ICM CT can enhance vascular intensity for directly vascular diseases diagnosis. However, the angiography reconstruction is challenging since patient individual differences and vascular disease diversity. In this paper, we propose a Multiple Adversarial Learning based Angiography Reconstruction (i. e. , MALAR) framework to enhance vascular intensity. Specifically, a bilateral learning mechanism is developed for mapping a relationship between source and target domains rather than the image-to-image mapping. Then, a dual correlation constraint is introduced to characterize both distribution uniformity from across-domain features and sample inconsistency within domain simultaneously. Finally, an adaptive fusion module by combining multi-scale information and long-range interactive dependency is explored to alleviate the interference of high-noise metal. Experiments are performed on CT sequences with different ICM doses. Quantitative results based on multiple metrics demonstrate the effectiveness of our MALAR on angiography reconstruction. Qualitative assessments by radiographers confirm the potential of our MALAR for the clinical diagnosis of vascular diseases.

Details DOI

AAAI Conference 2023 Conference Paper

Self-Supervised Bidirectional Learning for Graph Matching

  • Wenqi Guo
  • Lin Zhang
  • Shikui Tu
  • Lei Xu

Deep learning methods have demonstrated promising performance on the NP-hard Graph Matching (GM) problems. However, the state-of-the-art methods usually require the ground-truth labels, which may take extensive human efforts or be impractical to collect. In this paper, we present a robust self-supervised bidirectional learning method (IA-SSGM) to tackle GM in an unsupervised manner. It involves an affinity learning component and a classic GM solver. Specifically, we adopt the Hungarian solver to generate pseudo correspondence labels for the simple probabilistic relaxation of the affinity matrix. In addition, a bidirectional recycling consistency module is proposed to generate pseudo samples by recycling the pseudo correspondence back to permute the input. It imposes a consistency constraint between the pseudo affinity and the original one, which is theoretically supported to help reduce the matching error. Our method further develops a graph contrastive learning jointly with the affinity learning to enhance its robustness against the noise and outliers in real applications. Experiments deliver superior performance over the previous state-of-the-arts on five real-world benchmarks, especially under the more difficult outlier scenarios, demon- strating the effectiveness of our method.

PDF Details DOI

TCS Journal 2022 Journal Article

Approximation algorithms for the min-max clustered k-traveling salesmen problems

  • Xiaoguang Bao
  • Lei Xu
  • Wei Yu
  • Wei Song

Given a complete undirected graph G = ( V, E ), where V is the vertex set partitioned into K clusters V 1, V 2, …, V K and E is the edge set with edge weights satisfying triangle inequality, and a positive integer k, the min-max clustered k-traveling salesmen problem (min-max Ck-TSP) asks to find a set of k tours to visit all vertices, such that each cluster is visited by exactly one tour and the vertices of each cluster are visited consecutively. The objective is to minimize the weight of the maximum weight tour. The problem is known to be NP-hard even when k = 1 and K = 1. In this paper, we consider two variants of the problem. The first one is all the k tours have a common predefined starting vertex, and the other one is no starting vertex of any tour is specified. For both the variants we propose the first constant-factor approximation algorithms with ratios 5. 5 and 16, respectively.

Details DOI

AAMAS Conference 2022 Conference Paper

How Hard is Bribery in Elections with Randomly Selected Voters

  • Liangde Tao
  • Lin Chen
  • Lei Xu
  • Weidong Shi
  • Ahmed Sunny
  • Md Mahabub Uz Zaman

Many research works in computational social choice assume a fixed set of voters in an election and study the resistance of different voting rules against electoral manipulation. In recent years, however, a new technique known as random sample voting has been adopted in many multi-agent systems. One of the most prominent examples is blockchain. Many proof-of-stake based blockchain systems like Algorand will randomly select a subset of participants of the system to form a committee, and only the committee members will be involved in the decision of some important system parameters. This can be viewed as running an election where the voter committee (i. e. , the voters whose votes will be counted) is randomly selected. It is generally expected that the introduction of such randomness should make the election more resistant to electoral manipulation, despite the lack of theoretical analysis. In this paper, we present a systematic study on the resistance of an election with a randomly selected voter committee against bribery. Since the committee is randomly generated, by bribing any fixed subset of voters, the designated candidate may or may not win. Consequently, we consider the problem of finding a feasible solution that maximizes the winning probability of the designated candidate. We show that for most voting rules, this problem becomes extremely difficult for the briber as even finding any non-trivial solution with non-zero objective value becomes NP-hard. However, for plurality and veto, there exists a polynomial time approximation scheme that computes a near-optimal solution efficiently. The algorithm builds upon a novel integer programming formulation together with techniques from 𝑛-fold integer programming, which may be of a separate interest.

PDF

IJCAI Conference 2022 Conference Paper

Local Differential Privacy Meets Computational Social Choice - Resilience under Voter Deletion

  • Liangde Tao
  • Lin Chen
  • Lei Xu
  • Weidong Shi

The resilience of a voting system has been a central topic in computational social choice. Many voting rules, like plurality, are shown to be vulnerable as the attacker can target specific voters to manipulate the result. What if a local differential privacy (LDP) mechanism is adopted such that the true preference of a voter is never revealed in pre-election polls? In this case, the attacker can only infer stochastic information about a voter's true preference, and this may cause the manipulation of the electoral result significantly harder. The goal of this paper is to provide a quantitative study on the effect of adopting LDP mechanisms on a voting system. We introduce the metric PoLDP (power of LDP) that quantitatively measures the difference between the attacker's manipulation cost under LDP mechanisms and that without LDP mechanisms. The larger PoLDP is, the more robustness LDP mechanisms can add to a voting system. We give a full characterization of PoLDP for the voting system with plurality rule and provide general guidance towards the application of LDP mechanisms.

PDF Details DOI

YNIMG Journal 2022 Journal Article

The mirror neuron system compensates for amygdala dysfunction - associated social deficits in individuals with higher autistic traits

  • Lei Xu
  • Xiaoxiao Zheng
  • Shuxia Yao
  • Jialin Li
  • Meina Fu
  • Keshuang Li
  • Weihua Zhao
  • Hong Li

The amygdala is a core node in the social brain which exhibits structural and functional abnormalities in Autism spectrum disorder and there is evidence that the mirror neuron system (MNS) can functionally compensate for impaired emotion processing following amygdala lesions. In the current study, we employed an fMRI paradigm in 241 subjects investigating MNS and amygdala responses to observation, imagination and imitation of dynamic facial expressions and whether these differed in individuals with higher (n = 77) as opposed to lower (n = 79) autistic traits. Results indicated that individuals with higher compared to lower autistic traits showed worse recognition memory for fearful faces, smaller real-life social networks, and decreased left basolateral amygdala (BLA) responses to imitation. Additionally, functional connectivity between the left BLA and the left inferior frontal gyrus (IFG) as well as some other MNS regions was increased in individuals with higher autistic traits, especially during imitation of fearful expressions. The left BLA-IFG connectivity significantly moderated the autistic group differences on recognition memory for fearful faces, indicating that increased amygdala-MNS connectivity could diminish the social behavioral differences between higher and lower autistic trait groups. Overall, findings demonstrate decreased imitation-related amygdala activity in individuals with higher autistic traits in the context of increased amygdala-MNS connectivity which may functionally compensate for amygdala dysfunction and social deficits. Training targeting the MNS may capitalize on this compensatory mechanism for therapeutic benefits in Autism spectrum disorder.

Details DOI

AAMAS Conference 2021 Conference Paper

A Game Theoretical Analysis of Non-Linear Blockchain System

  • Lin Chen
  • Lei Xu
  • Zhimin Gao
  • Ahmed Imtiaz Sunny
  • Keshav Kasichainula
  • Weidong Shi

Recent advances in the blockchain research have been made in two important directions. One is refined resilience analysis utilizing game theory to study the consequences of selfish behavior of users (miners), and the other is the extension from a linear (chain) structure to a non-linear (graphical) structure for performance improvements, such as IOTA and Graphcoin. The first question that comes to mind is what improvements that a blockchain system would see by leveraging these new advances. In this paper, we consider three major properties for a blockchain system: 𝛼-partial verification, scalability, and finality-duration. We establish a formal framework and prove that no blockchain system can achieve 𝛼-partial verification for any fixed constant 𝛼, high scalability, and low finality-duration simultaneously. We observe that classical blockchain systems like Bitcoin achieves full verification (𝛼 = 1) and low finality-duration, Ethereum 2. 0 Sharding achieves low finality-duration and high scalability. We are interested in whether it is possible to partially satisfy the three properties.

PDF

TCS Journal 2021 Journal Article

Computational complexity characterization of protecting elections from bribery

  • Lin Chen
  • Ahmed Imtiaz Sunny
  • Lei Xu
  • Shouhuai Xu
  • Zhimin Gao
  • Yang Lu
  • Weidong Shi
  • Nolan Shah

The bribery problem in election has received considerable attention in the literature, upon which various algorithmic and complexity results have been obtained. In this setting, it is natural to ask whether we can protect an election from potential bribery attacks. We consider a scenario where the protector (or defender) can protect a voter at some cost such that a protected voter cannot be bribed (e. g. , by isolating the voter from potential bribers). This leads to the following bi-level decision problem: Is it possible for the protector to protect a proper subset of voters such that no briber with a fixed budget for bribery can alter the election result? The goal of this paper is to give a full characterization of the complexity of the associated protection problems. We conduct an extensive study on the protection problem and provide algorithmic and complexity results. When compared with the bribery problems that have been studied in the literature, we observe that the protection problem we study is significantly harder in general. Indeed, it becomes Σ 2 p -complete even for very restricted special cases, while most bribery problems lie in NP. However, it is not necessarily the case that the protection problem is always harder. Some of the protection problems can still be solved in polynomial time, while some of them remain as hard as the bribery problem with the same setting.

Details DOI

AAAI Conference 2021 Conference Paper

DeepTrader: A Deep Reinforcement Learning Approach for Risk-Return Balanced Portfolio Management with Market Conditions Embedding

  • Zhicheng Wang
  • Biwei Huang
  • Shikui Tu
  • Kun Zhang
  • Lei Xu

Most existing reinforcement learning (RL)-based portfolio management models do not take into account the market conditions, which limits their performance in risk-return balancing. In this paper, we propose Deep- Trader, a deep RL method to optimize the investment policy. In particular, to tackle the risk-return balancing problem, our model embeds macro market conditions as an indicator to dynamically adjust the proportion between long and short funds, to lower the risk of market fluctuations, with the negative maximum drawdown as the reward function. Additionally, the model involves a unit to evaluate individual assets, which learns dynamic patterns from historical data with the price rising rate as the reward function. Both temporal and spatial dependencies between assets are captured hierarchically by a specific type of graph structure. Particularly, we find that the estimated causal structure best captures the interrelationships between assets, compared to industry classification and correlation. The two units are complementary and integrated to generate a suitable portfolio which fits the market trend well and strikes a balance between return and risk effectively. Experiments on three well-known stock indexes demonstrate the superiority of DeepTrader in terms of risk-gain criteria.

PDF Details

AAAI Conference 2021 Conference Paper

IA-GM: A Deep Bidirectional Learning Method for Graph Matching

  • Kaixuan Zhao
  • Shikui Tu
  • Lei Xu

Existing deep learning methods for graph matching (GM) problems usually considered affinity learning to assist combinatorial optimization in a feedforward pipeline, and parameter learning is executed by backpropagating the gradients of the matching loss. Such a pipeline pays little attention to the possible complementary benefit from the optimization layer to the learning component. In this paper, we overcome the above limitation under a deep bidirectional learning framework. Our method circulates the output of the GM optimization layer to fuse with the input for affinity learning. Such direct feedback enhances the input by a feature enrichment and fusion technique, which exploits and integrates the global matching patterns from the deviation of the similarity permuted by the current matching estimate. As a result, the circulation enables the learning component to benefit from the optimization process, taking advantage of both global feature and the embedding result which is calculated by local propagation through node-neighbors. Moreover, circulation consistency induces an unsupervised loss that can be implemented individually or jointly to regularize the supervised loss. Experiments on challenging datasets demonstrate the effectiveness of our methods for both supervised learning and unsupervised learning.

PDF Details

JBHI Journal 2021 Journal Article

rBPDL:Predicting RNA-Binding Proteins Using Deep Learning

  • Mengting Niu
  • Jin Wu
  • Quan Zou
  • Zhendong Liu
  • Lei Xu

RNA-binding protein (RBP) is a powerful and wide-ranging regulator that plays an important role in cell development, differentiation, metabolism, health and disease. The prediction of RBPs provides valuable guidance for biologists. Although experimental methods have made great progress in predicting RBP, they are time-consuming and not flexible. Therefore, we developed a network model, rBPDL, by combining a convolutional neural network and long short-term memory for multilabel classification of RBPs. Moreover, to achieve better prediction results, we used a voting algorithm for ensemble learning of the model. We compared rBPDL with state-of-the-art methods and found that rBPDL significantly improved identification performance for the RBP68 dataset, with a macro-Area Under Curve (AUC), micro-AUC, and weighted AUC of 0. 936, 0. 962, and 0. 946, respectively. Furthermore, through AUC statistical analysis of the RBP domain, we analyzed the performance of rBPDL and found that the RBP identification performance in the same domain was similar. In addition, we analyzed the performance preferences and physicochemical properties of the binding protein amino acids and explored the characteristics that affect the binding by using the RBP86 dataset.

Details DOI

YNIMG Journal 2021 Journal Article

Segregating domain-general from emotional context-specific inhibitory control systems - ventral striatum and orbitofrontal cortex serve as emotion-cognition integration hubs

  • Qian Zhuang
  • Lei Xu
  • Feng Zhou
  • Shuxia Yao
  • Xiaoxiao Zheng
  • Xinqi Zhou
  • Jialin Li
  • Xiaolei Xu

Inhibitory control hierarchically regulates cognitive and emotional systems in the service of adaptive goal-directed behavior across changing task demands and environments. While previous studies convergently determined the contribution of prefrontal-striatal systems to general inhibitory control, findings on the specific circuits that mediate emotional context-specific impact on inhibitory control remained inconclusive. Against this background we combined an evaluated emotional Go/No Go task with fMRI in a large cohort of subjects (N=250) to segregate brain systems and circuits that mediate domain-general from emotion-specific inhibitory control. Particularly during a positive emotional context, behavioral results showed a lower accuracy for No Go trials and a faster response time for Go trials. While the dorsal striatum and lateral frontal regions were involved in inhibitory control irrespective of emotional context, activity in the ventral striatum (VS) and medial orbitofrontal cortex (mOFC) varied as a function of emotional context. On the voxel-wise whole-brain network level, limbic and striatal systems generally exhibited highest changes in global brain connectivity during inhibitory control, while global brain connectivity of the left mOFC was less decreased during emotional contexts. Functional connectivity analyses moreover revealed that negative coupling between the VS with inferior frontal gyrus (IFG)/insula and mOFC varied as a function of emotional context. Together these findings indicate separable domain- general as well as emotional context-specific inhibitory brain systems which specifically encompass the VS and its connections with frontal regions.

Details DOI

IJCAI Conference 2020 Conference Paper

Discrete Biorthogonal Wavelet Transform Based Convolutional Neural Network for Atrial Fibrillation Diagnosis from Electrocardiogram

  • Qingsong Xie
  • Shikui Tu
  • Guoxing Wang
  • Yong Lian
  • Lei Xu

For the problem of early detection of atrial fibrillation (AF) from electrocardiogram (ECG), it is difficult to capture subject-invariant discriminative features from ECG signals, due to the high variation in ECG morphology across subjects and the noise in ECG. In this paper, we propose an Discrete Biorthogonal Wavelet Transform (DBWT) Based Convolutional Neural Network (CNN) for AF detection, shortly called DBWT-AFNet. In DBWT-AFNet, rather than directly feeding ECG into CNN, DBWT is used to separate sub-signals in frequency band of heart beat from ECG, whose output is fed to CNN for AF diagnosis. Such sub-signals are better than the raw ECG for subject-invariant CNN representation learning because noisy information irrelevant to human beat has been largely filtered out. To strengthen the generalization ability of CNN to discover subject-invariant pattern in ECG, skip connection is exploited to propagate information well in neural network and channel attention is designed to adaptively highlight informative channel-wise features. Experiments show that the proposed DBWT-AFNet outperforms the state-of- the-art methods, especially for ECG segments classification across different subjects, where no data from testing subjects have been used in training.

PDF Details DOI

IJCAI Conference 2019 Conference Paper

Election with Bribe-Effect Uncertainty: A Dichotomy Result

  • Lin Chen
  • Lei Xu
  • Shouhuai Xu
  • Zhimin Gao
  • Weidong Shi

We consider the electoral bribery problem in computational social choice. In this context, extensive studies have been carried out to analyze the computational vulnerability of various voting (or election) rules. However, essentially all prior studies assume a deterministic model where each voter has an associated threshold value, which is used as follows. A voter will take a bribe and vote according to the attacker's (i. e. , briber's) preference when the amount of the bribe is above the threshold, and a voter will not take a bribe when the amount of the bribe is not above the threshold (in this case, the voter will vote according to its own preference, rather than the attacker's). In this paper, we initiate the study of a more realistic model where each voter is associated with a willingness function, rather than a fixed threshold value. The willingness function characterizes the likelihood a bribed voter would vote according to the attacker's preference; we call this bribe-effect uncertainty. We characterize the computational complexity of the electoral bribery problem in this new model. In particular, we discover a dichotomy result: a certain mathematical property of the willingness function dictates whether or not the computational hardness can serve as a deterrence to bribery attackers.

PDF Details

AAAI Conference 2019 Conference Paper

Election with Bribed Voter Uncertainty: Hardness and Approximation Algorithm

  • Lin Chen
  • Lei Xu
  • Shouhuai Xu
  • Zhimin Gao
  • Weidong Shi

Bribery in election (or computational social choice in general) is an important problem that has received a considerable amount of attention. In the classic bribery problem, the briber (or attacker) bribes some voters in attempting to make the briber’s designated candidate win an election. In this paper, we introduce a novel variant of the bribery problem, “Election with Bribed Voter Uncertainty” or BVU for short, accommodating the uncertainty that the vote of a bribed voter may or may not be counted. This uncertainty occurs either because a bribed voter may not cast its vote in fear of being caught, or because a bribed voter is indeed caught and therefore its vote is discarded. As a first step towards ultimately understanding and addressing this important problem, we show that it does not admit any multiplicative O(1)-approximation algorithm modulo standard complexity assumptions. We further show that there is an approximation algorithm that returns a solution with an additive-ε error in FPT time for any fixed ε.

PDF Details

NeurIPS Conference 2019 Conference Paper

Modeling Tabular data using Conditional GAN

  • Lei Xu
  • Maria Skoularidou
  • Alfredo Cuesta-Infante
  • Kalyan Veeramachaneni

Modeling the probability distribution of rows in tabular data and generating realistic synthetic data is a non-trivial task. Tabular data usually contains a mix of discrete and continuous columns. Continuous columns may have multiple modes whereas discrete columns are sometimes imbalanced making the modeling difficult. Existing statistical and deep neural network models fail to properly model this type of data. We design CTGAN, which uses a conditional generative adversarial network to address these challenges. To aid in a fair and thorough comparison, we design a benchmark with 7 simulated and 8 real datasets and several Bayesian network baselines. CTGAN outperforms Bayesian methods on most of the real datasets whereas other deep learning methods could not.

PDF Details

AAMAS Conference 2018 Conference Paper

Protecting Election from Bribery: New Approach and Computational Complexity Characterization

  • Lin Chen
  • Lei Xu
  • Shouhuai Xu
  • Zhimin Gao
  • Nolan Shah
  • Yang Lu
  • Weidong Shi

The bribery problem in elections has received a considerable amount of attention. In this paper, we initiate the study of a related, but new problem, the protection problem, namely protecting elections from bribery. In this problem, there is a defender who is given a defense budget and can use the budget to award some of the voters such that they cannot be bribed anymore. This naturally leads to the following bi-level decision problem: Is it possible for the defender with a given defense budget to protect an election from being manipulated by the attacker with a given attack budget for bribing voters? We characterize the computational complexity of the protection problem. We show that it is in general significantly harder than the bribery problem. However, the protection problem can be solved, under certain circumstances, in polynomial time.

PDF

YNIMG Journal 2018 Journal Article

Sex- and context-dependent effects of oxytocin on social sharing

  • Xiaole Ma
  • Weihua Zhao
  • Ruixue Luo
  • Feng Zhou
  • Yayuan Geng
  • Lei Xu
  • Zhao Gao
  • Xiaoxiao Zheng

We interact socially and form bonds with others because such experiences are rewarding. However, an insecure attachment style or social anxiety can reduce these rewarding effects. The neuropeptide oxytocin (OXT) may facilitate social interactions either by increasing their rewarding experience or by attenuating anxiety, although effects can be sex- and attachment-style dependent. In this study, 128 pairs of same-sex friends completed a social sharing paradigm in a double-blind, placebo-controlled, between-subject design with one friend inside an MRI scanner and the other in a remote behavioral testing room. In this way we could examine whether intranasal-OXT differentially modulated the emotional impact of social sharing and associated neural processing. Additionally, we investigated if OXT effects were modulated by sex and attachment style. Results showed that in women, but not men, OXT increased ratings for sharing stimuli with their friend but not with a stranger, particularly in the friend in the scanner. Corresponding neuroimaging results showed that OXT decreased both amygdala and insula activity as well as their functional connectivity in women when they shared with friends but had the opposite effect in men. On the other hand, OXT did not enhance responses in brain reward circuitry. In the PLC treated group amygdala responses in women when they shared pictures with their friend were positively associated with attachment anxiety and OXT uncoupled this. Our findings demonstrate that OXT facilitates the impact of sharing positive experiences with others in women, but not men, and that this is associated with differential effects on the amygdala and insula and their functional connections. Furthermore, OXT particularly reduced increased amygdala responses during sharing in individuals with higher attachment anxiety. Thus, OXT effects in this context may be due more to reduced anxiety when sharing with a friend than to enhanced social reward.

Details DOI

IJCAI Conference 2015 Conference Paper

Joint Learning of Character and Word Embeddings

  • Xinxiong Chen
  • Lei Xu
  • Zhiyuan Liu
  • Maosong Sun
  • Huanbo Luan

Most word embedding methods take a word as a basic unit and learn embeddings according to words’ external contexts, ignoring the internal structures of words. However, in some languages such as Chinese, a word is usually composed of several characters and contains rich internal information. The semantic meaning of a word is also related to the meanings of its composing characters. Hence, we take Chinese for example, and present a characterenhanced word embedding model (CWE). In order to address the issues of character ambiguity and non-compositional words, we propose multipleprototype character embeddings and an effective word selection method. We evaluate the effectiveness of CWE on word relatedness computation and analogical reasoning. The results show that CWE outperforms other baseline methods which ignore internal character information. The codes and data can be accessed from https: //github. com/ Leonard-Xu/CWE.

PDF Details

NeurIPS Conference 1995 Conference Paper

A Unified Learning Scheme: Bayesian-Kullback Ying-Yang Machine

  • Lei Xu

A Bayesian-Kullback learning scheme, called Ying-Yang Machine, is proposed based on the two complement but equivalent Bayesian representations for joint density and their Kullback divergence. Not only the scheme unifies existing major supervised and unsu(cid: 173) pervised learnings, including the classical maximum likelihood or least square learning, the maximum information preservation, the EM & em algorithm and information geometry, the recent popular Helmholtz machine, as well as other learning methods with new variants and new results; but also the scheme provides a number of new learning models.

PDF Details

NeurIPS Conference 1994 Conference Paper

An Alternative Model for Mixtures of Experts

  • Lei Xu
  • Michael Jordan
  • Geoffrey Hinton

We propose an alternative model for mixtures of experts which uses a different parametric form for the gating network. The modified model is trained by the EM algorithm. In comparison with earlier models-trained by either EM or gradient ascent-there is no need to select a learning stepsize. We report simulation experiments which show that the new architecture yields faster convergence. We also apply the new model to two problem domains: piecewise nonlinear function approximation and the combination of multiple previously trained classifiers.

PDF Details

NeurIPS Conference 1993 Conference Paper

Bayesian Self-Organization

  • Alan Yuille
  • Stelios Smirnakis
  • Lei Xu

Recent work by Becker and Hinton (Becker and Hinton, 1992) shows a promising mechanism, based on maximizing mutual in(cid: 173) formation assuming spatial coherence, by which a system can self(cid: 173) organize itself to learn visual abilities such as binocular stereo. We introduce a more general criterion, based on Bayesian probability theory, and thereby demonstrate a connection to Bayesian theo(cid: 173) ries of visual perception and to other organization principles for early vision (Atick and Redlich, 1990). Methods for implementa(cid: 173) tion using variants of stochastic learning are described and, for the special case of linear filtering, we derive an analytic expression for the output.

PDF Details

NeurIPS Conference 1992 Conference Paper

Self-Organizing Rules for Robust Principal Component Analysis

  • Lei Xu
  • Alan Yuille

In the presence of outliers, the existing self-organizing rules for Principal Component Analysis (PCA) perform poorly. Using sta(cid: 173) tistical physics techniques including the Gibbs distribution, binary decision fields and effective energies, we propose self-organizing PCA rules which are capable of resisting outliers while fulfilling various PCA-related tasks such as obtaining the first principal com(cid: 173) ponent vector, the first k principal component vectors, and directly finding the subspace spanned by the first k vector principal com(cid: 173) ponent vectors without solving for each vector individually. Com(cid: 173) parative experiments have shown that the proposed robust rules improve the performances of the existing PCA algorithms signifi(cid: 173) cantly when outliers are present.

PDF Details
v2026.09.13