论文检索

Yongxin Xu, Xinke Jiang, Xu Chu, Rihong Qiu, Yujie Feng, Hongxin Ding, Junfeng Zhao, Yasha Wang, Bing Xie

Exploring the correlations between medical features is essential for extracting patient health patterns from electronic health records (EHR) data, and strengthening medical predictions and decision-making. To constrain the hypothesis space of pure data-driven deep learning in the context of limited annotated data, a common trend is to incorporate external knowledge, especially knowledge priors related to personalized health contexts, to optimize model training. However, most existing methods lack flexibility and are constrained by the uncertainties brought about by fixed feature correlation priors. In addition, in utilizing knowledge, these methods overlook the knowledge informative for personalized healthcare. To this end, we propose DearLLM, a novel and effective framework that leverages feature correlations deduced by large language models (LLMs) to enhance personalized healthcare. Concretely, DearLLM captures and learns quantitative correlations between medical features by calculating the conditional perplexity of LLMs’ deduction based on personalized patient backgrounds. Then, DearLLM enhances healthcare predictions by emphasizing knowledge that carries unique patient information through a feature-frequency-aware graph pooling method. Extensive experiments on two real-world benchmark datasets show significant performance gains brought by DearLLM. Furthermore, the discovered findings align well with medical literature, offering meaningful clinical interpretations.

Xovee Xu, Yifan Zhang, Fan Zhou, Jingkuan Song

Understanding and predicting the popularity of online User-Generated Content (UGC) is critical for various social and recommendation systems. Existing efforts have focused on extracting predictive features and using pre-trained deep models to learn and fuse multimodal UGC representations. However, the dissemination of social UGCs is not an isolated process in social network; rather, it is influenced by contextual relevant UGCs and various exogenous factors, including social ties, trends, user interests, and platform algorithms. In this work, we propose a retrieval-based framework to enhance the popularity prediction of multimodal UGCs. Our framework extends beyond a simple semantic retrieval, incorporating a meta retrieval strategy that queries a diverse set of relevant UGCs by considering multimodal content semantics, and metadata from user and post. Moreover, to eliminate irrelevant and noisy UGCs in retrieval, we introduce a new measure called Relative Retrieval Contribution to Prediction (RRCP), which selectively refines the retrieved UGCs. We then aggregate the contextual UGC knowledge using vision-language graph neural networks, and fuse them with an RRCP-Attention-based prediction network. Extensive experiments on three large-scale social media datasets demonstrate significant improvements ranging from 26.68% to 48.19% across all metrics compared to strong baselines.

Di Xiong, Shuoyuan Wang, Lei Zhang, Wenbo Huang, Chaolei Han

Human Activity Recognition (HAR) aims to recognize activities by training models on massive sensor data. In real-world deployment, a crucial aspect of HAR that has been largely overlooked is that the test sets may have different distributions from training sets due to inter-subject variability including age, gender, behavioral habits, etc., which leads to poor generalization performance. One promising solution is to learn domain-invariant representations to enable a model to generalize on an unseen distribution. However, most existing methods only consider the feature-invariance of the penultimate layer for domain-invariant learning, which leads to suboptimal results. In this paper, we propose a Categorical Concept Invariant Learning (CCIL) framework for generalizable activity recognition, which introduces a concept matrix to regularize the model in the training stage by simultaneously concertrating on feature-invariance and logit-invariance. Our key idea is that the concept matrix for samples belonging to the same activity category should be similar. Extensive experiments on four public HAR benchmarks demonstrate that our CCIL substantially outperforms the state-of-the-art approaches under cross-person, cross-dataset, cross-position, and one-person-to-another settings.

Zeke Xia, Ming Hu, Dengke Yan, Ruixuan Liu, Anran Li, Xiaofei Xie, Mingsong Chen

Although Split Federated Learning (SFL) effectively enables knowledge sharing among resource-constrained clients, it suffers from low training performance due to the neglect of data heterogeneity and catastrophic forgetting problems. To address these issues, we propose a novel SFL approach named MultiSFL, which adopts i) an effective multi-model aggregation mechanism to alleviate gradient divergence caused by heterogeneous data and ii) a novel knowledge replay strategy to deal with the catastrophic forgetting problem. MultiSFL adopts two servers (i.e., the fed server and main server) to maintain multiple branch models for local training and an aggregated master model for knowledge sharing among branch models. To mitigate catastrophic forgetting, the main server of MultiSFL selects multiple assistant devices for knowledge replay according to the training data distribution of each full branch model. Experimental results obtained from various non-IID and IID scenarios demonstrate that MultiSFL significantly outperforms conventional SFL methods by up to a 23.25% test accuracy improvement.

Zhihao Wu, Yushi Cheng, Tianyang Sun, Xiaoyu Ji, Wenyuan Xu

Personalized text-to-image synthesis models, such as DreamBooth, have demonstrated significant potential in creating lifelike images tailored to a specific individual by fine-tuning from a limited set of face images and simple prompts. However, if misused, these model could pose a serious risk of privacy infringement by generating harmful images containing violent or pornographic content. To tackle this issue, this paper introduces MYOPIA, a method that renders facial images unlearnable by incorporating error-minimizing perturbations. These meticulously designed perturbations enables the model to quickly overfit to them, resulting in a swift reduction in loss and the cessation of model fine-tuning, effectively preventing the model from capturing genuine facial features. Moreover, to ensure the imperceptibility and robustness of the perturbations, we utilize the Just-Noticeable-Difference and Expectation-of-Transformation techniques to regulate both their location and intensity. Evaluation on two face dataset, i.e., VGGFace2 and CelebA-HQ, with various model versions illustrates the effectiveness of our approach in preserving personal privacy. Furthermore, our method showcases robust transferability across diverse model versions and demonstrates resilience against various image pre-processing techniques.

Lirong Wu, Haitao Lin, Yufei Huang, Zhangyang Gao, Cheng Tan, Yunfan Liu, Tailin Wu, Stan Z. Li

Antibodies are Y-shaped proteins that protect the host by binding to specific antigens, and their binding is mainly determined by the Complementary Determining Regions (CDRs) in the antibody. Despite the great progress made in CDR design, existing computational methods still encounter several challenges: 1) poor capability of modeling complex CDRs with long sequences due to insufficient contextual information; 2) conditioned on pre-given antigenic epitopes and their static interaction with the target antibody; 3) neglect of specificity during antibody optimization leads to non-specific antibodies. In this paper, we take into account a variety of node features, edge features, and edge relations to include more contextual and geometric information. We propose a novel Relation-Aware Antibody Design (RAAD) framework, which dynamically models antigen-antibody interactions for co-designing the sequences and structures of antigen-specific CDRs. Furthermore, we propose a new evaluation metric to better measure antibody specificity and develop a contrasting specificity-enhancing constraint to optimize the specificity of antibodies. Extensive experiments have demonstrated the superior capability of RAAD in terms of antibody modeling, generation, and optimization across different CDR types, sequence lengths, pre-training strategies, and input contexts.

Juntao Wu, Ziyu Song, Xiaoyu Zhang, Shujun Xie, Longxin Lin, Ke Wang

For an extensive period, Vision Transformers (ViTs) have been deemed unsuitable for attaining robust performance on small-scale datasets, with WideResNet models maintaining dominance in this domain. While WideResNet models have persistently set the state-of-the-art (SOTA) benchmarks for robust accuracy on datasets such as CIFAR-10 and CIFAR-100, this paper challenges the prevailing belief that only WideResNet can excel in this context. We pose the critical question of whether ViTs can surpass the robust accuracy of WideResNet models. Our results provide a resounding affirmative answer. By employing ViT, enhanced with data generated by a diffusion model for adversarial training, we demonstrate that ViTs can indeed outshine WideResNet in terms of robust accuracy. Specifically, under the Infty-norm threat model with epsilon = 8/255, our approach achieves robust accuracies of 74.97% on CIFAR-10 and 44.07% on CIFAR-100, representing improvements of +3.9% and +1.4%, respectively, over the previous SOTA models. Notably, our ViT-B/2 model, with 3 times fewer parameters, surpasses the previously best-performing WRN-70-16. Our achievement opens a new avenue, suggesting that future models employing ViTs or other novel efficient architectures could eventually replace the long-dominant WRN models.

Fang Wu, Bozhen Hu, Stan Z. Li

Molecular dynamics (MD) has long been the de facto choice for simulating intricate physical systems from first principles. Recent efforts utilize the implicit neural representation (INR) to directly learn surface point clouds' signed distance function (SDF) with promising outcomes. However, INR's temporal generalization to unexplored molecular systems remains limited, which poses a significant barrier to applying INR to a broader range of real-world scenarios. This study introduces MoE-DSR, an enhanced version of dynamic surface representations (DSR) that effectively integrates the mixture-of-experts (MoE) strategy. Specifically, the router employs a novel geometric surface cloud network to extract the structural information from the initial static protein conformation as the prior knowledge. Meanwhile, experts compromising a team of equivariant implicit neural networks (E-INNs), each responsible for distinct protein families, ensure precise SDF estimation across varied protein data landscapes. We showcase the ability of MoE-DSR to model dynamic protein surface shapes using ensembles from ATLAS, the largest available protein MD simulations database. Extensive experiments validate its effectiveness in analyzing complex molecular systems across continuous space and time domains.

Yueqing Wang, Peng Zhang, Yushuang Liu, Jianing Zhao, Jie Lin, Yi Chen

Aerodynamic coefficient prediction is pivotal in aircraft and vehicles' design, performance evaluation, and motion control. Integrating artificial neural networks into aerodynamic coefficient prediction offers a promising alternative to traditional numerical methods burdened by extensive computations and high costs. Nevertheless, this data-driven approach faces several critical challenges, which limit its further performance enhancement: i) The current research lacks a profound understanding of the complex interplay between the shape of an object and its aerodynamic characteristics. ii) The scarcity of high-quality aerodynamic data poses a significant barrier. The models trained on limited datasets lack generalization ability, struggling to accurately predict and adapt to diverse aerodynamic performance under new shapes or conditions. To overcome these challenges, we introduce an innovative framework that employs cross-attention to capture the intimate interplay between shape and flow conditions and allows for the direct utilization of pre-trained models on general shape datasets to mitigate the scarcity of aerodynamic data. Furthermore, to bolster the inference capabilities of this data-driven approach, we integrate physical information constraints into the model, leveraging them as guiding principles to enhance the model's predictive power under unknown conditions. Experimental validation demonstrates that our proposed method performs excellently in multiple aerodynamic prediction tasks. This achievement brings a new technological breakthrough to the field of aerodynamic prediction and provides robust support for the design optimization of complex systems such as aircraft and vehicles.

Xiaozheng Wang, Yong Yang, Shuying Huang, Hangyuan Lu, Weiguo Wan, Aoqi Zhao

Currently, most Hyperspectral (HS) pansharpening methods have two problems, namely the lack of consideration the spatial variations of HS images and inaccurate feature reconstruction in multi-channel complex mapping relationships, leading to spectral and spatial distortions in the fusion results. To address these issues, we propose a dynamic network based on feature modulation and probability mask (FMPM-DNet) for HS pansharpening, including two stages of spectral-spatial feature modulation and feature reconstruction. In the first stage, to increase the feature representation ability of the model, a wave function is defined based on complex transformation to convert spatial features into wave-like features. On this basis, considering the spatial variations of HS images, a dynamic feature modulation unit (DFMU) is constructed to achieve adaptive modulation and coarse fusion of features by dynamically generating spectral-spatial correction matrix. In the second stage, a feature probability mask unit (FPMU) is designed to realize global feature embedding at different depths and local feature embedding at the same depth to obtain refined fused features. Extensive experiments on three widely used datasets demonstrate that the proposed FMPM-Net achieves significant improvements in both spatial and spectral quality metrics compared to some state-of-the-art (SOTA) methods.

Ruoqi Wang, Haitao Wang, Qiong Luo, Feng Wang, Hejun Wu

Radio telescopes produce visibility data about celestial objects, but these data are sparse and noisy. As a result, images created on raw visibility data are of low quality. Recent studies have used deep learning models to reconstruct visibility data to get cleaner images. However, these methods rely on a substantial amount of labeled training data, which requires significant labeling effort from radio astronomers. Addressing this challenge, we propose VisRec, a model-agnostic semi-supervised learning approach to visibility data reconstruction in radio astronomy. Specifically, VisRec consists of both a supervised learning module and an unsupervised learning module. In the supervised learning module, we introduce a set of data augmentation functions to produce diverse visibility examples. In comparison, the unsupervised learning module in VisRec augments unlabeled data and uses reconstructions from non-augmented visibility as pseudo-labels for training. This hybrid approach allows VisRec to effectively leverage both labeled and unlabeled data. This way, VisRec performs well even when labeled data is scarce. Our evaluation results show that VisRec is applicable to various models, and outperforms all baseline methods in terms of reconstruction quality, robustness, and generalizability.

Lingzhi Wang, Xingshan Zeng, Jinsong Guo, Kam-Fai Wong, Georg Gottlob

This paper explores Machine Unlearning (MU), an emerging field that is gaining increased attention due to concerns about neural models unintentionally remembering personal or sensitive information. We present SeUL, a novel method that enables selective and fine-grained unlearning for language models. Unlike previous work that employs a fully reversed training objective in unlearning, SeUL minimizes the negative impact on the capability of language models, particularly in terms of generation. Furthermore, we introduce two innovative evaluation metrics, sensitive extraction likelihood (S-EL) and sensitive memorization accuracy (S-MA), specifically designed to assess the effectiveness of forgetting sensitive information. In support of the unlearning framework, we propose efficient automatic online and offline sensitive span annotation methods. The online selection method, based on language probability scores, ensures computational efficiency, while the offline annotation involves a two-stage LLM-based process for robust verification. In summary, this paper contributes a novel selective unlearning method (SeUL), introduces specialized evaluation metrics (S-EL and S-MA) for assessing sensitive information forgetting, and proposes automatic online and offline sensitive span annotation methods to support the overall unlearning framework and evaluation.

Jingyuan Wang, Yujing Lin, Yudong Li

Trajectory data mining is crucial for smart city management. However, collecting large-scale trajectory datasets is challenging due to factors such as commercial conflicts and privacy regulations. Therefore, we urgently need trajectory generation techniques to address this issue. Existing trajectory generation methods rely on the global road network structure of cities. When the road network structure changes, these methods are often not transferable to other cities. In fact, there exist invariant mobility patterns between different cities: 1) People prefer paths with the minimal travel cost; 2) The travel cost of roads has an invariant relationship with the topological features of the road network. Based on the above insight, this paper proposes a Generalizable Trajectory Generation model (GTG). The model consists of three parts: 1) Extracting city-invariant road representation based on Space Syntax method; 2) Cross-city travel cost prediction through disentangled adversarial training; 3) Travel preference learning by shortest path search and preference update. By learning invariant movement patterns, the model is capable of generating trajectories in new cities. Experiments on three datasets demonstrates that our model significantly outperforms existing models in terms of generalization ability.

Jiabao Wang, Zepeng Wu, Qian Dong, Lingzhong Meng, Yunzhi Xue, Yukuan Yang

Recent advancements have underscored the exceptional analytical and situational understanding capabilities of Large Language Models (LLMs) in autonomous driving decisions. However, the inherent hallucination issues of LLMs pose significant safety concerns when utilized as standalone decision-making systems. To address these challenges, we propose the Hybrid-Driving framework, which leverages LLMs' situational comprehension and reasoning abilities alongside the specialized driving expertise embedded in knowledge graphs and driving rules, thereby enhancing the safety, robustness, and reliability of autonomous driving decisions. To articulate driving experiences clearly, we introduce the Scenario Evolution Knowledge Graph (SEKG), which integrates scenario prediction and action risk analysis in autonomous driving. By delineating observation areas and defining Time-to-Collision (TTC) levels, we effectively control the number of driving scenario nodes and ensure scenario diversity. Based on the scenario evolution relationships within the SEKG, we predict scenarios and assess associated action risks. Additionally, we implement a rule-filtering mechanism to eliminate unreasonable actions and employ prompt engineering to integrate scenario information, optional actions, and SEKG-based action risk analysis into the LLMs for decision-making. Extensive experiments demonstrate that our approach substantially improves decision success rates compared to using LLMs alone (≥37.5%), as well as surpasses the DiLu framework with LLMs and few-shot driving memory (≥7.5%), and other reinforcement learning methods (≥11%). These results validate the effectiveness of the Hybrid-Driving framework in enhancing LLM reliability for autonomous driving and advocate for its broader application of domain-specific knowledge across other fields.

Jia Wang, Liyan Zhu, Zhe Wang, Chenqiu Zhang, Yaoxing Wu, Jun Cui, Jianqiang Li

Missense mutations could affect the Liquid-Liquid Phase Separation (LLPS) propensity of proteins and lead to aberrant phase-separating behaviours, which are recently found to be associated with many diseases including Alzheimer's and cancer. However, the regulatory role of mutations in LLPS remains unclear due to challenges in accurately characterizing the LLPS ability of mutants, including the high similarity in features, lack of labeled data, and vast amounts of data involved. To bridge this gap and facilitate the discovery of therapeutic strategies, we propose the first machine learning-based guider for protein phase-separating behaviour alteration, PScalpel. PScalpel leverages both structural information and an auxiliary tasks-based graph contrastive learning framework to distinguish the mutants’ LLPS ability, and incorporates a genetic algorithms-based recommendation method to identify mutants with desired LLPS properties. Comprehensive computational and biological experiments validate the effectiveness of PScalpel as a versatile tool for guiding alterations in protein phase separation behavior.

Chengyue Wang, Haicheng Liao, Bonan Wang, Yanchen Guan, Bin Rao, Ziyuan Pu, Zhiyong Cui, Cheng-Zhong Xu, Zhenning Li

Accurate trajectory prediction is essential for the safety and efficiency of autonomous driving. Traditional models often struggle with real-time processing, capturing non-linearity and uncertainty in traffic environments, efficiency in dense traffic, and modeling temporal dynamics of interactions. We introduce NEST (Neuromodulated Small-world Hypergraph Trajectory Prediction), a novel framework that integrates Small-world Networks and hypergraphs for superior interaction modeling and prediction accuracy. This integration enables the capture of both local and extended vehicle interactions, while the Neuromodulator component adapts dynamically to changing traffic conditions. We validate the NEST model on several real-world datasets, including nuScenes, MoCAD, and HighD. The results consistently demonstrate that NEST outperforms existing methods in various traffic scenarios, showcasing its exceptional generalization capability, efficiency, and temporal foresight. Our comprehensive evaluation illustrates that NEST significantly improves the reliability and operational efficiency of autonomous driving systems, making it a robust solution for trajectory prediction in complex traffic environments.

Renshuai Tao, Manyi Le, Chuangchuang Tan, Huan Liu, Haotong Qin, Yao Zhao

Despite significant advances in deepfake detection, handling varying image quality, especially due to different compressions on online social networks (OSNs), remains challenging. Current methods succeed by leveraging correlations between paired images, whether raw or compressed. However, in open-world scenarios, paired data is scarce, with compressed images readily available but corresponding raw versions difficult to obtain. This imbalance, where unpaired data vastly outnumbers paired data, often leads to reduced detection performance, as existing methods struggle without corresponding raw images. To overcome this issue, we propose a novel approach named the open-world deepfake detection network (ODDN), which comprises two core modules: open-world data aggregation (ODA) and compression-discard gradient correction (CGC). ODA effectively aggregates correlations between compressed and raw samples through both fine-grained and coarse-grained analyses for paired and unpaired data, respectively. CGC incorporates a compression-discard gradient correction to further enhance performance across diverse compression methods in OSN. This technique optimizes the training gradient to ensure the model remains insensitive to compression variations. Extensive experiments conducted on 17 popular deepfake datasets demonstrate the superiority of the ODDN over SOTA baselines.

Lei Tan, Yuliang Xue, Guobiao Li, Zhenxing Qian, Sheng Li, Chunlei Bao

Ceramic artworks with elegant patterns present enormous collectible value and profits. To claim the copyright, the builder usually pastes their conspicuous stamp on the bottom or side of the ceramic artworks, which inevitably affects the external image of the artwork. In addition, the stamp is weak in resisting forgery attacks due to its visible nature. To address the above issues, we propose in this paper a novel framework for embedding invisible watermarking into patterns of the ceramic artworks. In the framework, a template-based watermarking embedding scheme is designed to map the watermark to an invisible template, which is added to the ceramic pattern to create its watermarked version. A distortion layer is further proposed to model the distortion of ceramic patterns in the ceramic manufacturing process, where a color-halftoning and an adaptive brightness adjustment strategy are developed to counter the print and firing operations that introduce the most significant distortions. Finally, a deep decoder is learned to extract the watermarking from the distorted pattern. Various experiments have been conducted to demonstrate the advantage of our proposed method for protecting the copyright of the ceramic artworks, which provides reliable watermark extraction accuracy without the need for a conspicuous stamp.

Cheng Tan, Yijie Zhang, Zhangyang Gao, Yufei Huang, Haitao Lin, Lirong Wu, Fandi Wu, Mathieu Blanchette, Stan Z. Li

The development of therapeutic antibodies heavily relies on accurate predictions of how antigens will interact with antibodies. Existing computational methods in antibody design often overlook crucial conformational changes that antigens undergo during the binding process, significantly impacting the reliability of the resulting antibodies. To bridge this gap, we introduce dyAb, a flexible framework that incorporates AlphaFold2-driven predictions to model pre-binding antigen structures and specifically addresses the dynamic nature of antigen conformation changes. Our dyAb model leverages a unique combination of coarse-grained interface alignment and fine-grained flow matching techniques to simulate the interaction dynamics and structural evolution of the antigen-antibody complex, providing a realistic representation of the binding process. Extensive experiments show that dyAb significantly outperforms existing models in antibody design involving changing antigen conformations. These results highlight dyAb's potential to streamline the design process for therapeutic antibodies, promising more efficient development cycles and improved outcomes in clinical applications.

Nan Sun, Han Fang, Yuxing Lu, Chengxin Zhao, Hefei Ling

DNN-based watermarking methods have rapidly advanced, with the ``Encoder-Noise Layer-Decoder'' (END) framework being the most widely used. To ensure end-to-end training, the noise layer in the framework must be differentiable. However, real-world distortions are often non-differentiable, leading to challenges in end-to-end training. Existing solutions only treat the distortion perturbation as additive noise, which does not fully integrate the effect of distortion in training. To better incorporate non-differentiable distortions into training, we propose a novel dual-decoder architecture (END^2). Unlike conventional END architecture, our method employs two structurally identical decoders: the Teacher Decoder, processing pure watermarked images, and the Student Decoder, handling distortion-perturbed images. The gradient is backpropagated only through the Teacher Decoder branch to optimize the encoder thus bypassing the problem of non-differentiability. To ensure resistance to arbitrary distortions, we enforce alignment of the two decoders' feature representations by maximizing the cosine similarity between their intermediate vectors on a hypersphere. Extensive experiments demonstrate that our scheme outperforms state-of-the-art algorithms under various non-differentiable distortions. Moreover, even without the differentiability constraint, our method surpasses baselines with a differentiable noise layer. Our approach is effective and easily implementable across all END architectures, enhancing practicality and generalizability.