AAAI 2025 论文列表

Molecular representation learning plays a crucial role in various downstream tasks, such as molecular property prediction and drug design. To accurately represent molecules, Graph Neural Networks (GNNs) and Graph Transformers (GTs) have shown potential in the realm of self-supervised pretraining. However, existing approaches often overlook the relationship between molecular structure and electronic information, as well as the internal semantic reasoning within molecules. This omission of fundamental chemical knowledge in graph semantics leads to incomplete molecular representations, missing the integration of structural and electronic data. To address these issues, we introduce MOL-Mamba, a framework that enhances molecular representation by combining structural and electronic insights. MOL-Mamba consists of an Atom & Fragment Mamba-Graph (MG) for hierarchical structural reasoning and a Mamba-Transformer (MT) fuser for integrating molecular structure and electronic correlation learning. Additionally, we propose a Structural Distribution Collaborative Training and E-semantic Fusion Training framework to further enhance molecular representation learning. Extensive experiments demonstrate that MOL-Mamba outperforms state-of-the-art baselines across eleven chemical-biological molecular datasets.

In the field of Moving Infrared Small Target Detection (MIRSTD), current methods typically use sequential modeling with two individual modules for spatial and temporal processing. However, such a modeling strategy lacks clear guidance on the motion and displacement difference between moving targets and background noise, thereby limiting the feature discriminability and resulting in error-prone target localization. This paper addresses this issue from clip and frame levels and proposes a novel architecture MOCID for MIRSTD. For clip-level feature fusion, we design a spatio-temporal backbone consisting of several proposed Fourier-inspired Spatio-temporal Attention (FISTA) layers. Each FISTA layer sequentially processes the features from spatial and temporal views to capture clip-level temporal motion context, where Fourier Transformation and Inverse Fourier Transformation are employed for each view. This context is then embedded into dynamic convolutional kernels for subsequent spatial feature extraction, thereby enabling clear motion difference guidance and generating comprehensive features. For frame-level feature fusion, we design a Displacement-aware Mamba Module (DAM) to capture detailed frame-to-frame displacement information. DAM utilizes an innovative Temporal Interpolation and Displacement-aware Scan technique to perform spatio-temporal difference-aware displacement modeling, introducing elaborate temporal indicators into feature extraction. Combining the above improvements, our model captures comprehensive motion and displacement contexts, significantly improving the detection of the small target. Extensive experiments demonstrate that MOCID achieves state-of-the-art detection accuracy on popular IRDST and DAUB datasets. Furthermore, MOCID offers a superior balance between throughput and performance compared to other methods. The code for this work will be made publicly available.

As the global population ages and the incidence of chronic diseases increases, the demand for early detection of abnormal medical conditions is increasing. Traditional health monitoring methods often require significant resources and specialized personnel, limiting their widespread use. Leveraging advancements in AI technologies, this study proposes a non-invasive method for detecting abnormal medical conditions from image data. A multimodal perception framework is introduced, integrating features from various modalities, including facial expressions and body postures, to enhance detection accuracy. The framework employs a Cascaded Squeeze-Excitation (CSE) module, consisting of Adaptive and Multi-modal Squeeze-Excitation components, to capture complex feature dependencies and improve cross-modal performance. Extensive experiments demonstrate the effectiveness of this approach, showing improved performance over existing methods. In addition, a new dataset that encompasses a wide range of medical conditions has been released, providing a valuable resource for future research in this domain.

As deep learning advances, Large Language Models (LLMs) and their multimodal counterparts, Vision-Language Models (VLMs), have shown exceptional performance in many real-world tasks. However, VLMs face significant security challenges, such as jailbreak attacks, where attackers attempt to bypass the model’s safety alignment to elicit harmful responses. The threat of jailbreak attacks on VLMs arises from both the inherent vulnerabilities of LLMs and the multiple information channels that VLMs process. While various attacks and defenses have been proposed, there is a notable gap in unified and comprehensive evaluations, as each method is evaluated on different dataset and metrics, making it impossible to compare the effectiveness of each method. To address this gap, we introduce MMJ-Bench, a unified pipeline for evaluating jailbreak attacks and defense techniques for VLMs. Through extensive experiments, we assess the effectiveness of various attack methods against SoTA VLMs and evaluate the impact of defense mechanisms on both defense effectiveness and model utility for normal tasks. Our comprehensive evaluation contribute to the field by offering a unified and systematic evaluation framework and the first public-available benchmark for VLM jailbreak research. We also demonstrate several insightful findings that highlights directions for future studies.

Controllable 3D scene generation has extensive applications in virtual reality and interior design, where the generated scenes should exhibit high levels of realism and controllability in terms of geometry. Scene graphs provide a suitable data representation that facilitates these applications. However, current graph-based methods for scene generation are constrained to text-based inputs and exhibit insufficient adaptability to flexible user inputs, hindering the ability to precisely control object geometry. To address this issue, we propose MMGDreamer, a dual-branch diffusion model for scene generation that incorporates a novel Mixed-Modality Graph, visual enhancement module, and relation predictor. The mixed-modality graph allows object nodes to integrate textual and visual modalities, with optional relationships between nodes. It enhances adaptability to flexible user inputs and enables meticulous control over the geometry of objects in the generated scenes. The visual enhancement module enriches the visual fidelity of text-only nodes by constructing visual representations using text embeddings. Furthermore, our relation predictor leverages node representations to infer absent relationships between nodes, resulting in more coherent scene layouts. Extensive experimental results demonstrate that MMGDreamer exhibits superior control of object geometry, achieving state-of-the-art scene generation performance.

Multiple object tracking (MOT) from unmanned aerial vehicle (UAV) platforms requires efficient motion modeling. This is because UAV-MOT faces both local object motion and global camera motion. Motion blur also increases the difficulty of detecting large moving objects. Previous UAV motion modeling approaches either focus only on local motion or ignore motion blurring effects, thus limiting their tracking performance and speed. To address these issues, we propose the Motion Mamba Module, which explores both local and global motion features through cross-correlation and bi-directional Mamba Modules for better motion modeling. To address the detection difficulties caused by motion blur, we also design motion margin loss to effectively improve the detection accuracy of motion blurred objects. Based on the Motion Mamba module and motion margin loss, our proposed MM-Tracker surpasses the state-of-the-art in two widely open-source UAV-MOT datasets.

We introduce MM-Mixing, a multi-modal mixing alignment framework for 3D understanding. MM-Mixing applies mixing-based methods to multi-modal data, preserving and optimizing cross-modal connections while enhancing diversity and improving alignment across modalities. Our proposed two-stage training pipeline combines feature-level and input-level mixing to optimize the 3D encoder. The first stage employs feature-level mixing with contrastive learning to align 3D features with their corresponding modalities. The second stage incorporates both feature-level and input-level mixing, introducing mixed point cloud inputs to further refine 3D feature representations. MM-Mixing enhances intermodality relationships, promotes generalization, and ensures feature consistency while providing diverse and realistic training samples. We demonstrate that MM-Mixing significantly improves baseline performance across various learning scenarios, including zero-shot 3D classification, linear probing 3D classification, and cross-modal 3D shape retrieval. Notably, we improved the zero-shot classification accuracy on ScanObjectNN from 51.3% to 61.9%, and on Objaverse-LVIS from 46.8% to 51.4%. Our findings highlight the potential of multi-modal mixing-based alignment to significantly advance 3D object recognition and understanding while remaining straightforward to implement and integrate into existing frameworks.

Large visual-language models (LVLMs) have achieved great success in multiple applications. However, they still encounter challenges in complex scenes, especially those involving camouflaged objects. This is primarily due to the lack of samples related to camouflaged scenes in the training dataset. To mitigate this issue, we construct the MM-CamObj dataset for the first time, comprising two subsets: CamObj-Align and CamObj-Instruct. Specifically, CamObj-Align contains 11,363 image-text pairs, and it is designed for VL alignment and injecting rich knowledge of camouflaged scenes into LVLMs. CamObj-Instruct is collected for fine-tuning the LVLMs with improved instruction-following capabilities, and it includes 11,363 images and 68,849 conversations with diverse instructions. Based on the MM-CamObj dataset, we propose the CamObj-Llava, an LVLM specifically designed for addressing tasks in camouflaged scenes. To facilitate our model's effective acquisition of knowledge about camouflaged objects and scenes, we introduce a curriculum learning strategy with six distinct modes. Additionally, we construct the CamObj-Bench to evaluate the existing LVLMs' capabilities of understanding, recognition, localization and count in camouflage scenes. This benchmark includes 600 images and 7 tasks, with a total of 9,449 questions. Extensive experiments are conducted on the CamObj-Bench with CamObj-Llava, 8 existing open-source and 3 closed-source LVLMs. Surprisingly, the results indicate that our model achieves a 25.84% improvement in 4 out of 7 tasks compared to GPT-4o.

Long-tailed (LT) data distribution is common in multi-label image classification (MLC) and can significantly impact the performance of classification models. One reason is the challenge of learning unbiased instance representations (i.e. features) for imbalanced datasets. Additionally, the co-occurrence of head/tail classes within the same instance, along with complex label dependencies, introduces further challenges. In this work, we delve into this problem through the lens of neural collapse (NC). NC refers to a phenomenon where the last-layer features and classifier of a deep neural network model exhibit a simplex Equiangular Tight Frame (ETF) structure during its terminal training phase. This structure creates an optimal linearly separable state. However, this phenomenon typically occurs in balanced datasets but rarely applies to the typical imbalanced problem. To induce NC properties under Long-tailed multi-label classification (LT-MLC) conditions, we propose an approach named MLC-NC, which aims to learn high-quality data representations and improve the model’s generalization ability. Specifically, MLC-NC accounts for the fact that different labels correspond to different feature parts located in images. MLC-NC extracts class-wise features from each instance through a cross-attention mechanism. To guide the features toward the ETF structure, we introduce visual-semantic feature alignment with a fixed ETF structured label embedding, which helps to learn evenly distributed class centers. To reduce within-class feature variation, we introduce collapse calibration within a lower-dimensional feature space. To mitigate classification bias, we concatenate features and feed them into a binarized fixed ETF classifier. As an orthogonal approach to existing methods, MLC-NC can be seamlessly integrated into various frameworks. Extensive experiments on widely-used benchmarks demonstrate the effectiveness of our method.

Deep neural networks (DNNs) typically employ an end-to-end (E2E) training paradigm which presents several challenges, including high GPU memory consumption, inefficiency, and difficulties in model parallelization during training. Recent research has sought to address these issues, with one promising approach being local learning. This method involves partitioning the backbone network into gradient-isolated modules and manually designing auxiliary networks to train these local modules. Existing methods often neglect the interaction of information between local modules, leading to myopic issues and a performance gap compared to E2E training. To address these limitations, we propose the Multilaminar Leap Augmented Auxiliary Network (MLAAN). Specifically, MLAAN comprises Multilaminar Local Modules (MLM) and Leap Augmented Modules (LAM). MLM captures both local and global features through independent and cascaded auxiliary networks, alleviating performance issues caused by insufficient global features. However, overly simplistic auxiliary networks can impede MLM's ability to capture global information. To address this, we further design LAM, an enhanced auxiliary network that uses the Exponential Moving Average (EMA) method to facilitate information exchange between local modules, thereby mitigating the shortsightedness resulting from inadequate interaction. The synergy between MLM and LAM has demonstrated excellent performance. Our experiments on the CIFAR-10, STL-10, SVHN, and ImageNet datasets show that MLAAN can be seamlessly integrated into existing local learning frameworks, significantly enhancing their performance and even surpassing end-to-end (E2E) training methods, while also reducing GPU memory consumption.

The out-of-distribution (OOD) detection on graph-structured data is crucial for deploying graph neural networks securely in open-world scenarios. However, existing methods have overlooked the prevalent scenario of multi-label classification in real-world applications. In this work, we investigate the unexplored issue of OOD detection within multi-label node classification tasks. We propose ML-GOOD, a simple yet sufficient approach that utilizes an energy function to gauge the OOD score for each label. We further develop a strategy for amalgamating multiple label energies, allowing for the comprehensive utilization of label information to tackle the primary challenges encountered in multi-label scenarios. Extensive experimentation conducted on seven diverse sets of real-world multi-label graph datasets, encompassing cross-domain scenarios. The results show that the AUROC of ML-GOOD is improved by 5.26% in intra-domain and 6.54% in cross-domain compared to the previous methods. These empirical validations not only affirm the robustness of our methodology but also illuminate new avenues for further exploration within this burgeoning field of research.

Current Siamese and Transformer trackers commonly use various subtask branches like regression and classification to predict object states. Despite the demonstrated success, these subtask branches might introduce location and scale offsets due to discrepancies and misalignment in the respective predictions. To address this, we propose a novel generative tracker, MIMTrack, which defines tracking as a Masked Image Modeling (MIM) process combined with in-context learning (ICL). MIMTrack begins with building the visual prompt image, which consists of a template, a search area, and two target images associated with them. The target image transforms the bounding box into a unified RGB image space as other tracking image. All states prediction are naturally aligned by pixels generation of search target image. In light of this, we perform a MIM process within the visual prompt to reconstruct a masked search target image using the context from other parts. MIM with ICL makes use of implicit cross-relations between template and search area. A singlestream generative framework reduces the offset in the estimation. Furthermore, a latent memory module is introduced as a plugin to enhance pixel generation by leveraging various target appearances over time. The advanced performance observed on leading benchmark datasets highlights the simplicity and effectiveness of our MIMTrack framework.

Network traffic includes data transmitted across a network, such as web browsing and file transfers, and is organized into packets (small units of data) and flows (sequences of packets exchanged between two endpoints). Classifying encrypted traffic is essential for detecting security threats and optimizing network management. Recent advancements have highlighted the superiority of foundation models in this task, particularly for their ability to leverage large amounts of unlabeled data and demonstrate strong generalization to unseen data. However, existing methods that focus on token-level relationships fail to capture broader flow patterns, as tokens, defined as sequences of hexadecimal digits, typically carry limited semantic information in encrypted traffic. These flow patterns, which are crucial for traffic classification, arise from the interactions between packets within a flow, not just their internal structure. To address this limitation, we propose a Multi-Instance Encrypted Traffic Transformer (MIETT), which adopts a multi-instance approach where each packet is treated as a distinct instance within a larger bag representing the entire flow. This enables the model to capture both token-level and packet-level relationships more effectively through Two-Level Attention (TLA) layers, improving the model's ability to learn complex packet dynamics and flow patterns. We further enhance the model's understanding of temporal and flow-specific dynamics by introducing two novel pre-training tasks: Packet Relative Position Prediction (PRPP) and Flow Contrastive Learning (FCL). After fine-tuning, MIETT achieves state-of-the-art (SOTA) performance across five datasets, demonstrating its effectiveness in classifying encrypted traffic and understanding complex network behaviors.

We present and release MIDI-GPT, a generative system based on the Transformer architecture that is designed for computer-assisted music composition workflows. MIDI-GPT supports the infilling of musical material at the track and bar level, and can condition generation on attributes including: instrument type, musical style, note density, polyphony level, and note duration. In order to integrate these features, we employ an alternative representation for musical material, creating a time-ordered sequence of musical events for each track and concatenating several tracks into a single sequence, rather than using a single time-ordered sequence where the musical events corresponding to different tracks are interleaved. We also propose a variation of our representation allowing for expressiveness. We present experimental results that demonstrate that MIDI-GPT is able to consistently avoid duplicating the musical material it was trained on, generate music that is stylistically similar to the training dataset, and that attribute controls allow enforcing various constraints on the generated material. We also outline several real-world applications of MIDI-GPT, including collaborations with industry partners that explore the integration and evaluation of MIDI-GPT into commercial products, as well as several artistic works produced using it.

The increasing parameters and expansive dataset of large lan- guage models (LLMs) highlight the urgent demand for a technical solution to audit the underlying privacy risks and copyright issues associated with LLMs. Existing studies have partially addressed this need through an exploration of the pre-training data detection problem, which is an instance of a membership inference attack (MIA). This problem involves determining whether a given piece of text has been used during the pre-training phase of the target LLM. Although existing methods have designed various sophisticated MIA score functions to achieve considerable detection performance in pre-trained LLMs, how to achieve high-confidence detection and how to perform MIA on aligned LLMs remain challenging. In this paper, we propose MIA-Tuner, a novel instruction-based MIA method, which instructs LLMs themselves to serve as a more precise pre-training data detector internally, rather than design an external MIA score function. Furthermore, we design two instruction-based safeguards to respectively mitigate the privacy risks brought by the existing methods and MIA-Tuner. To comprehensively evaluate the most recent state-of-the-art LLMs, we collect a more up-to-date MIA benchmark dataset, named WIKIMIA-24, to replace the widely adopted benchmark WIKIMIA. We conduct extensive experiments across various aligned and unaligned LLMs over the two benchmark datasets. The results demonstrate that MIA-Tuner increases the AUC of MIAs from 0.7 to a significantly high level of 0.9.

Completely automated public Turing test to tell humans apart (CAPTCHA) is an effective mechanism to protect websites and online applications from malicious bots programs. Image-based CAPTCHA is one of the most widely used schemes. However, deep learning techniques have significantly weakened the security of some image-based CAPTCHA schemes. Mooney images (MIs) are important research materials in the field of cognitive science. Compared to natural images, MI exhibits fewer visual cues, fragmented content, and greater ambiguity, leading to the perception of MI relying more on the iterative process between feedforward and feedback mechanisms. In this paper, we raise an intriguing question: can MIs be used to enhance the security of CAPTCHA? Before this study, we first propose a novel framework HiMI that generates the high-quality MIs from natural images and also allows flexible adjustment of the perceived difficulty. Based on MI, we design two MI-CAPTCHA schemes related to object detection and instance segmentation tasks, respectively. We experimentally demonstrate that HiMI performs better than other baseline methods in terms of both image quality and application potential in two MI-CAPTCHA schemes. Additionally, we conduct experiments to explore the solving performance of humans and CAPTCHA solvers under different parameter settings of schemes, providing valuable reference for the practical application.

Similar to Language or Image LLMs, VideoLLMs are also plagued by hallucination issues. Hallucinations in videos not only manifest in the spatial dimension regarding the perception of the existence of visual objects (static) but also the temporal dimension influencing the perception of actions and events (dynamic). This paper introduces the concept of Motion Hallucination for the first time, exploring the hallucination phenomena caused by insufficient motion perception capabilities in VideoLMMs, as well as how to detect, evaluate, and mitigate the hallucination. To this end, we propose the first benchmark for assessing motion hallucination MHBench, which consists of 1,200 videos of 20 different action categories. By constructing a collection of adversarial triplet types of videos (original/antonym/incomplete), we achieve a comprehensive evaluation of motion hallucination. Furthermore, we present a Motion Contrastive Decoding (MotionCD) method, which employs bidirectional motion elimination between the original video and its reverse playback to construct an amateur model that removes the influence of motion while preserving visual information, thereby effectively suppressing motion hallucination. Extensive experiments on MHBench reveal that current state-of-the-art VideoLLMs significantly suffer from motion hallucination, while the introduction of MotionCD can effectively mitigate this issue, achieving up to a 15.1% performance improvement. We hope this work will guide future efforts in avoiding and mitigating hallucinations in VideoLLMs.

Multi-modal Federated Learning (MFL) is a distributed machine learning paradigm that enables multiple participants with multi-modal data to collaboratively train a global model for multi-modal tasks without sharing their local data. MFL typically deploys the trained global model as an Embedding-as-a-Service (EaaS), allowing participants to obtain embeddings for downstream tasks. However, it increases the risk of unauthorized copying and leakage of the model. Protecting the ownership of the MFL model while maintaining model performance is challenging. In this paper, we propose the first general model ownership protection framework for MFL, named MFL-Owner. MFL-Owner decouples the watermarking process from the model training process and addresses both ownership verification and traceability, effectively safeguarding the interests of the MFL collective. MFL-Owner leverages the concept of orthogonal transformations by incorporating a linear transformation matrix with orthogonal constraints into the model, achieving high-quality ownership verification and traceability with minimal impact on model performance. To enhance the practicality of the watermark and prevent conflicts among multiple clients during tracing, we propose a trigger dataset selection method based on out-of-distribution data combined with Gaussian noise perturbation. Our experiments on multiple datasets demonstrate that MFL-Owner is effective for model ownership verification and traceability for MFL.