Weiqiang Wang

Jintang Li, Ruofan Wu, Wangbin Sun et al.

The last years have witnessed the emergence of a promising self-supervised learning strategy, referred to as masked autoencoding. However, there is a lack of theoretical understanding of how masking matters on graph autoencoders (GAEs). In this work, we present masked graph autoencoder (MaskGAE), a self-supervised learning framework for graph-structured data. Different from standard GAEs, MaskGAE adopts masked graph modeling (MGM) as a principled pretext task - masking a portion of edges and attempting to reconstruct the missing part with partially visible, unmasked graph structure. To understand whether MGM can help GAEs learn better representations, we provide both theoretical and empirical evidence to comprehensively justify the benefits of this pretext task. Theoretically, we establish close connections between GAEs and contrastive learning, showing that MGM significantly improves the self-supervised learning scheme of GAEs. Empirically, we conduct extensive experiments on a variety of graph benchmarks, demonstrating the superiority of MaskGAE over several state-of-the-arts on both link prediction and node classification tasks.

Haoxing Chen, Yaohui Li, Yan Hong et al.

Audio-visual zero-shot learning aims to recognize unseen classes based on paired audio-visual sequences. Recent methods mainly focus on learning multi-modal features aligned with class names to enhance the generalization ability to unseen categories. However, these approaches ignore the obscure event concepts in class names and may inevitably introduce complex network structures with difficult training objectives. In this paper, we introduce a straightforward yet efficient framework called KnowleDge-Augmented audio-visual learning (KDA), which aids the model in more effectively learning novel event content by leveraging an external knowledge base. Specifically, we first propose to utilize the knowledge contained in large language models (LLMs) to generate numerous descriptive sentences that include important distinguishing audio-visual features of event classes, which helps to better understand unseen categories. Furthermore, we propose a knowledge-aware adaptive margin loss to help distinguish similar events, further improving the generalization ability towards unseen classes. Extensive experimental results demonstrate that our proposed KDA can outperform state-of-the-art methods on three popular audio-visual zero-shot learning datasets.Our code will be avaliable at \url{https://github.com/chenhaoxing/KDA}.

Weiqiang Wang, Xuefei Zhe, Qiuhong Ke et al.

We present a neural network-based system for long-term, multi-action human motion synthesis. The system, dubbed as NEURAL MARIONETTE, can produce high-quality and meaningful motions with smooth transitions from simple user input, including a sequence of action tags with expected action duration, and optionally a hand-drawn moving trajectory if the user specifies. The core of our system is a novel Transformer-based motion generation model, namely MARIONET, which can generate diverse motions given action tags. Different from existing motion generation models, MARIONET utilizes contextual information from the past motion clip and future action tag, dedicated to generating actions that can smoothly blend historical and future actions. Specifically, MARIONET first encodes target action tag and contextual information into an action-level latent code. The code is unfolded into frame-level control signals via a time unrolling module, which could be then combined with other frame-level control signals like the target trajectory. Motion frames are then generated in an auto-regressive way. By sequentially applying MARIONET, the system NEURAL MARIONETTE can robustly generate long-term, multi-action motions with the help of two simple schemes, namely "Shadow Start" and "Action Revision". Along with the novel system, we also present a new dataset dedicated to the multi-action motion synthesis task, which contains both action tags and their contextual information. Extensive experiments are conducted to study the action accuracy, naturalism, and transition smoothness of the motions generated by our system.

Jiafu Wu, Mufeng Yao, Dong Wu et al.

Graph representation plays an important role in the field of financial risk control, where the relationship among users can be constructed in a graph manner. In practical scenarios, the relationships between nodes in risk control tasks are bidirectional, e.g., merchants having both revenue and expense behaviors. Graph neural networks designed for undirected graphs usually aggregate discriminative node or edge representations with an attention strategy, but cannot fully exploit the out-degree information when used for the tasks built on directed graph, which leads to the problem of a directional bias. To tackle this problem, we propose a Directed Graph ATtention network called DGAT, which explicitly takes out-degree into attention calculation. In addition to having directional requirements, the same node might have different representations of its input and output, and thus we further propose a dual embedding of DGAT, referred to as DEDGAT. Specifically, DEDGAT assigns in-degree and out-degree representations to each node and uses these two embeddings to calculate the attention weights of in-degree and out-degree nodes, respectively. Experiments performed on the benchmark datasets show that DGAT and DEDGAT obtain better classification performance compared to undirected GAT. Also,the visualization results demonstrate that our methods can fully use both in-degree and out-degree information.

Haoxing Chen, Yaohui Li, Zizheng Huang et al.

Pre-trained large-scale vision-language models (VLMs) have acquired profound understanding of general visual concepts. Recent advancements in efficient transfer learning (ETL) have shown remarkable success in fine-tuning VLMs within the scenario of limited data, introducing only a few parameters to harness task-specific insights from VLMs. Despite significant progress, current leading ETL methods tend to overfit the narrow distributions of base classes seen during training and encounter two primary challenges: (i) only utilizing uni-modal information to modeling task-specific knowledge; and (ii) using costly and time-consuming methods to supplement knowledge. To address these issues, we propose a Conditional Prototype Rectification Prompt Learning (CPR) method to correct the bias of base examples and augment limited data in an effective way. Specifically, we alleviate overfitting on base classes from two aspects. First, each input image acquires knowledge from both textual and visual prototypes, and then generates sample-conditional text tokens. Second, we extract utilizable knowledge from unlabeled data to further refine the prototypes. These two strategies mitigate biases stemming from base classes, yielding a more effective classifier. Extensive experiments on 11 benchmark datasets show that our CPR achieves state-of-the-art performance on both few-shot classification and base-to-new generalization tasks. Our code is avaliable at \url{https://github.com/chenhaoxing/CPR}.

Zizheng Huang, Haoxing Chen, Jiaqi Li et al.

Recent Vision Mamba (Vim) models exhibit nearly linear complexity in sequence length, making them highly attractive for processing visual data. However, the training methodologies and their potential are still not sufficiently explored. In this paper, we investigate strategies for Vim and propose Stochastic Layer-Wise Shuffle (SLWS), a novel regularization method that can effectively improve the Vim training. Without architectural modifications, this approach enables the non-hierarchical Vim to get leading performance on ImageNet-1K compared with the similar type counterparts. Our method operates through four simple steps per layer: probability allocation to assign layer-dependent shuffle rates, operation sampling via Bernoulli trials, sequence shuffling of input tokens, and order restoration of outputs. SLWS distinguishes itself through three principles: \textit{(1) Plug-and-play:} No architectural modifications are needed, and it is deactivated during inference. \textit{(2) Simple but effective:} The four-step process introduces only random permutations and negligible overhead. \textit{(3) Intuitive design:} Shuffling probabilities grow linearly with layer depth, aligning with the hierarchical semantic abstraction in vision models. Our work underscores the importance of tailored training strategies for Vim models and provides a helpful way to explore their scalability.

Xiaofei Yin, Yijie Hong, Ya Guo et al.

In the evolving landscape of multimodal language models, understanding the nuanced meanings conveyed through visual cues - such as satire, insult, or critique - remains a significant challenge. Existing evaluation benchmarks primarily focus on direct tasks like image captioning or are limited to a narrow set of categories, such as humor or satire, for deep semantic understanding. To address this gap, we introduce, for the first time, a comprehensive, multi-level Chinese-based benchmark designed specifically for evaluating the understanding of implicit meanings in images. This benchmark is systematically categorized into four subtasks: surface-level content understanding, symbolic meaning interpretation, background knowledge comprehension, and implicit meaning comprehension. We propose an innovative semi-automatic method for constructing datasets, adhering to established construction protocols. Using this benchmark, we evaluate 15 open-source large vision language models (LVLMs) and GPT-4o, revealing that even the best-performing model lags behind human performance by nearly 14% in understanding implicit meaning. Our findings underscore the intrinsic challenges current LVLMs face in grasping nuanced visual semantics, highlighting significant opportunities for future research and development in this domain. We will publicly release our InsightVision dataset, code upon acceptance of the paper.

Haoxing Chen, Zhuoer Xu, Zhangxuan Gu et al.

Diffusion model based language-guided image editing has achieved great success recently. However, existing state-of-the-art diffusion models struggle with rendering correct text and text style during generation. To tackle this problem, we propose a universal self-supervised text editing diffusion model (DiffUTE), which aims to replace or modify words in the source image with another one while maintaining its realistic appearance. Specifically, we build our model on a diffusion model and carefully modify the network structure to enable the model for drawing multilingual characters with the help of glyph and position information. Moreover, we design a self-supervised learning framework to leverage large amounts of web data to improve the representation ability of the model. Experimental results show that our method achieves an impressive performance and enables controllable editing on in-the-wild images with high fidelity. Our code will be avaliable in \url{https://github.com/chenhaoxing/DiffUTE}.

Weixi Zhao, Yunjie Tian, Qixiang Ye et al.

Exploiting relations among 2D joints plays a crucial role yet remains semi-developed in 2D-to-3D pose estimation. To alleviate this issue, we propose GraFormer, a novel transformer architecture combined with graph convolution for 3D pose estimation. The proposed GraFormer comprises two repeatedly stacked core modules, GraAttention and ChebGConv block. GraAttention enables all 2D joints to interact in global receptive field without weakening the graph structure information of joints, which introduces vital features for later modules. Unlike vanilla graph convolutions that only model the apparent relationship of joints, ChebGConv block enables 2D joints to interact in the high-order sphere, which formulates their hidden implicit relations. We empirically show the superiority of GraFormer through conducting extensive experiments across popular benchmarks. Specifically, GraFormer outperforms state of the art on Human3.6M dataset while using 18$\%$ parameters. The code is available at https://github.com/Graformer/GraFormer .

Hongzhu Li, Weiqiang Wang

The connectionist temporal classification (CTC) enables end-to-end sequence learning by maximizing the probability of correctly recognizing sequences during training. The outputs of a CTC-trained model tend to form a series of spikes separated by strongly predicted blanks, know as the spiky problem. To figure out the reason for it, we reinterpret the CTC training process as an iterative fitting task that is based on frame-wise cross-entropy loss. It offers us an intuitive way to compare target probabilities with model outputs for each iteration, and explain how the model outputs gradually turns spiky. Inspired by it, we put forward two ways to modify the CTC training. The experiments demonstrate that our method can well solve the spiky problem and moreover, lead to faster convergence over various training settings. Beside this, the reinterpretation of CTC, as a brand new perspective, may be potentially useful in other situations. The code is publicly available at https://github.com/hzli-ucas/caffe/tree/ctc.

Weixi Zhao, Weiqiang Wang

We propose a novel attention-based 2D-to-3D pose estimation network for graph-structured data, named KOG-Transformer, and a 3D pose-to-shape estimation network for hand data, named GASE-Net. Previous 3D pose estimation methods have focused on various modifications to the graph convolution kernel, such as abandoning weight sharing or increasing the receptive field. Some of these methods employ attention-based non-local modules as auxiliary modules. In order to better model the relationship between nodes in graph-structured data and fuse the information of different neighbor nodes in a differentiated way, we make targeted modifications to the attention module and propose two modules designed for graph-structured data, graph relative positional encoding multi-head self-attention (GR-MSA) and K-order graph-oriented multi-head self-attention (KOG-MSA). By stacking GR-MSA and KOG-MSA, we propose a novel network KOG-Transformer for 2D-to-3D pose estimation. Furthermore, we propose a network for shape estimation on hand data, called GraAttention shape estimation network (GASE-Net), which takes a 3D pose as input and gradually models the shape of the hand from sparse to dense. We have empirically shown the superiority of KOG-Transformer through extensive experiments. Experimental results show that KOG-Transformer significantly outperforms the previous state-of-the-art methods on the benchmark dataset Human3.6M. We evaluate the effect of GASE-Net on two public available hand datasets, ObMan and InterHand2.6M. GASE-Net can predict the corresponding shape for input pose with strong generalization ability.

Zeyang Sha, Hanling Tian, Zhuoer Xu et al.

The emergence of autonomous Large Language Model (LLM) agents capable of tool usage has introduced new safety risks that go beyond traditional conversational misuse. These agents, empowered to execute external functions, are vulnerable to both user-initiated threats (e.g., adversarial prompts) and tool-initiated threats (e.g., malicious outputs from compromised tools). In this paper, we propose the first unified safety-alignment framework for tool-using agents, enabling models to handle both channels of threat via structured reasoning and sandboxed reinforcement learning. We introduce a tri-modal taxonomy, including benign, malicious, and sensitive for both user prompts and tool responses, and define a policy-driven decision model. Our framework employs a custom-designed sandbox environment that simulates real-world tool execution and allows fine-grained reward shaping. Through extensive evaluations on public and self-built benchmarks, including Agent SafetyBench, InjecAgent, and BFCL, we demonstrate that our safety-aligned agents significantly improve resistance to security threats while preserving strong utility on benign tasks. Our results show that safety and effectiveness can be jointly optimized, laying the groundwork for trustworthy deployment of autonomous LLM agents.

ZhenZhe Ying, Zhuoer Xu, Zhifeng Li et al.

Despite the success of deep learning in computer vision and natural language processing, Gradient Boosted Decision Tree (GBDT) is yet one of the most powerful tools for applications with tabular data such as e-commerce and FinTech. However, applying GBDT to multi-task learning is still a challenge. Unlike deep models that can jointly learn a shared latent representation across multiple tasks, GBDT can hardly learn a shared tree structure. In this paper, we propose Multi-task Gradient Boosting Machine (MT-GBM), a GBDT-based method for multi-task learning. The MT-GBM can find the shared tree structures and split branches according to multi-task losses. First, it assigns multiple outputs to each leaf node. Next, it computes the gradient corresponding to each output (task). Then, we also propose an algorithm to combine the gradients of all tasks and update the tree. Finally, we apply MT-GBM to LightGBM. Experiments show that our MT-GBM improves the performance of the main task significantly, which means the proposed MT-GBM is efficient and effective.

Ji Gan, Weiqiang Wang, Ke Lu

Chinese is one of the most widely used languages in the world, yet online handwritten Chinese character recognition (OLHCCR) remains challenging. To recognize Chinese characters, one popular choice is to adopt the 2D convolutional neural network (2D-CNN) on the extracted feature images, and another one is to employ the recurrent neural network (RNN) or 1D-CNN on the time-series features. Instead of viewing characters as either static images or temporal trajectories, here we propose to represent characters as geometric graphs, retaining both spatial structures and temporal orders. Accordingly, we propose a novel spatial graph convolution network (SGCN) to effectively classify those character graphs for the first time. Specifically, our SGCN incorporates the local neighbourhood information via spatial graph convolutions and further learns the global shape properties with a hierarchical residual structure. Experiments on IAHCC-UCAS2016, ICDAR-2013, and UNIPEN datasets demonstrate that the SGCN can achieve comparable recognition performance with the state-of-the-art methods for character recognition.

Chao Qu, Hui Li, Chang Liu et al.

We propose a \emph{collaborative} multi-agent reinforcement learning algorithm named variational policy propagation (VPP) to learn a \emph{joint} policy through the interactions over agents. We prove that the joint policy is a Markov Random Field under some mild conditions, which in turn reduces the policy space effectively. We integrate the variational inference as special differentiable layers in policy such that the actions can be efficiently sampled from the Markov Random Field and the overall policy is differentiable. We evaluate our algorithm on several large scale challenging tasks and demonstrate that it outperforms previous state-of-the-arts.

Yuanqiang Cai, Longyin Wen, Libo Zhang et al.

The convention standard for object detection uses a bounding box to represent each individual object instance. However, it is not practical in the industry-relevant applications in the context of warehouses due to severe occlusions among groups of instances of the same categories. In this paper, we propose a new task, ie, simultaneously object localization and counting, abbreviated as Locount, which requires algorithms to localize groups of objects of interest with the number of instances. However, there does not exist a dataset or benchmark designed for such a task. To this end, we collect a large-scale object localization and counting dataset with rich annotations in retail stores, which consists of 50,394 images with more than 1.9 million object instances in 140 categories. Together with this dataset, we provide a new evaluation protocol and divide the training and testing subsets to fairly evaluate the performance of algorithms for Locount, developing a new benchmark for the Locount task. Moreover, we present a cascaded localization and counting network as a strong baseline, which gradually classifies and regresses the bounding boxes of objects with the predicted numbers of instances enclosed in the bounding boxes, trained in an end-to-end manner. Extensive experiments are conducted on the proposed dataset to demonstrate its significance and the analysis discussions on failure cases are provided to indicate future directions. Dataset is available at https://isrc.iscas.ac.cn/gitlab/research/locount-dataset.

Yuanqiang Cai, Dawei Du, Libo Zhang et al.

Object detection and counting are related but challenging problems, especially for drone based scenes with small objects and cluttered background. In this paper, we propose a new Guided Attention Network (GANet) to deal with both object detection and counting tasks based on the feature pyramid. Different from the previous methods relying on unsupervised attention modules, we fuse different scales of feature maps by using the proposed weakly-supervised Background Attention (BA) between the background and objects for more semantic feature representation. Then, the Foreground Attention (FA) module is developed to consider both global and local appearance of the object to facilitate accurate localization. Moreover, the new data argumentation strategy is designed to train a robust model in various complex scenes. Extensive experiments on three challenging benchmarks (i.e., UAVDT, CARPK and PUCPR+) show the state-of-the-art detection and counting performance of the proposed method compared with existing methods.

XiaoBin Li, WeiQiang Wang

Loss functions play a key role in training superior deep neural networks. In convolutional neural networks (CNNs), the popular cross entropy loss together with softmax does not explicitly guarantee minimization of intra-class variance or maximization of inter-class variance. In the early studies, there is no theoretical analysis and experiments explicitly indicating how to choose the number of units in fully connected layer. To help CNNs learn features more fast and discriminative, there are two contributions in this paper. First, we determine the minimum number of units in FC layer by rigorous theoretical analysis and extensive experiment, which reduces CNNs' parameter memory and training time. Second, we propose a negative-focused weights-biased softmax (W-Softmax) loss to help CNNs learn more discriminative features. The proposed W-Softmax loss not only theoretically formulates the intraclass compactness and inter-class separability, but also can avoid overfitting by enlarging decision margins. Moreover, the size of decision margins can be flexibly controlled by adjusting a hyperparameter $α$. Extensive experimental results on several benchmark datasets show the superiority of W-Softmax in image classification tasks.