Bing Su

Zezhong Lv, Bing Su, Ji-Rong Wen

Video moment localization aims to retrieve the target segment of an untrimmed video according to the natural language query. Weakly supervised methods gains attention recently, as the precise temporal location of the target segment is not always available. However, one of the greatest challenges encountered by the weakly supervised method is implied in the mismatch between the video and language induced by the coarse temporal annotations. To refine the vision-language alignment, recent works contrast the cross-modality similarities driven by reconstructing masked queries between positive and negative video proposals. However, the reconstruction may be influenced by the latent spurious correlation between the unmasked and the masked parts, which distorts the restoring process and further degrades the efficacy of contrastive learning since the masked words are not completely reconstructed from the cross-modality knowledge. In this paper, we discover and mitigate this spurious correlation through a novel proposed counterfactual cross-modality reasoning method. Specifically, we first formulate query reconstruction as an aggregated causal effect of cross-modality and query knowledge. Then by introducing counterfactual cross-modality knowledge into this aggregation, the spurious impact of the unmasked part contributing to the reconstruction is explicitly modeled. Finally, by suppressing the unimodal effect of masked query, we can rectify the reconstructions of video proposals to perform reasonable contrastive learning. Extensive experimental evaluations demonstrate the effectiveness of our proposed method. The code is available at \href{https://github.com/sLdZ0306/CCR}{https://github.com/sLdZ0306/CCR}.

Zhao Yang, Bing Su, Ji-Rong Wen

Text-to-motion generation has gained increasing attention, but most existing methods are limited to generating short-term motions that correspond to a single sentence describing a single action. However, when a text stream describes a sequence of continuous motions, the generated motions corresponding to each sentence may not be coherently linked. Existing long-term motion generation methods face two main issues. Firstly, they cannot directly generate coherent motions and require additional operations such as interpolation to process the generated actions. Secondly, they generate subsequent actions in an autoregressive manner without considering the influence of future actions on previous ones. To address these issues, we propose a novel approach that utilizes a past-conditioned diffusion model with two optional coherent sampling methods: Past Inpainting Sampling and Compositional Transition Sampling. Past Inpainting Sampling completes subsequent motions by treating previous motions as conditions, while Compositional Transition Sampling models the distribution of the transition as the composition of two adjacent motions guided by different text prompts. Our experimental results demonstrate that our proposed method is capable of generating compositional and coherent long-term 3D human motions controlled by a user-instructed long text stream. The code is available at \href{https://github.com/yangzhao1230/PCMDM}{https://github.com/yangzhao1230/PCMDM}.

Yujie Zhou, Wenwen Qiang, Anyi Rao et al.

Zero-shot skeleton-based action recognition aims to recognize actions of unseen categories after training on data of seen categories. The key is to build the connection between visual and semantic space from seen to unseen classes. Previous studies have primarily focused on encoding sequences into a singular feature vector, with subsequent mapping the features to an identical anchor point within the embedded space. Their performance is hindered by 1) the ignorance of the global visual/semantic distribution alignment, which results in a limitation to capture the true interdependence between the two spaces. 2) the negligence of temporal information since the frame-wise features with rich action clues are directly pooled into a single feature vector. We propose a new zero-shot skeleton-based action recognition method via mutual information (MI) estimation and maximization. Specifically, 1) we maximize the MI between visual and semantic space for distribution alignment; 2) we leverage the temporal information for estimating the MI by encouraging MI to increase as more frames are observed. Extensive experiments on three large-scale skeleton action datasets confirm the effectiveness of our method. Code: https://github.com/YujieOuO/SMIE.

Jiexin Wang, Yujie Zhou, Wenwen Qiang et al.

Human motion prediction (HMP) has emerged as a popular research topic due to its diverse applications, but it remains a challenging task due to the stochastic and aperiodic nature of future poses. Traditional methods rely on hand-crafted features and machine learning techniques, which often struggle to model the complex dynamics of human motion. Recent deep learning-based methods have achieved success by learning spatio-temporal representations of motion, but these models often overlook the reliability of motion data. Additionally, the temporal and spatial dependencies of skeleton nodes are distinct. The temporal relationship captures motion information over time, while the spatial relationship describes body structure and the relationships between different nodes. In this paper, we propose a novel spatio-temporal branching network using incremental information for HMP, which decouples the learning of temporal-domain and spatial-domain features, extracts more motion information, and achieves complementary cross-domain knowledge learning through knowledge distillation. Our approach effectively reduces noise interference and provides more expressive information for characterizing motion by separately extracting temporal and spatial features. We evaluate our approach on standard HMP benchmarks and outperform state-of-the-art methods in terms of prediction accuracy.

Jiangmeng Li, Wenwen Qiang, Changwen Zheng et al.

While self-supervised learning techniques are often used to mining implicit knowledge from unlabeled data via modeling multiple views, it is unclear how to perform effective representation learning in a complex and inconsistent context. To this end, we propose a methodology, specifically consistency and complementarity network (CoCoNet), which avails of strict global inter-view consistency and local cross-view complementarity preserving regularization to comprehensively learn representations from multiple views. On the global stage, we reckon that the crucial knowledge is implicitly shared among views, and enhancing the encoder to capture such knowledge from data can improve the discriminability of the learned representations. Hence, preserving the global consistency of multiple views ensures the acquisition of common knowledge. CoCoNet aligns the probabilistic distribution of views by utilizing an efficient discrepancy metric measurement based on the generalized sliced Wasserstein distance. Lastly on the local stage, we propose a heuristic complementarity-factor, which joints cross-view discriminative knowledge, and it guides the encoders to learn not only view-wise discriminability but also cross-view complementary information. Theoretically, we provide the information-theoretical-based analyses of our proposed CoCoNet. Empirically, to investigate the improvement gains of our approach, we conduct adequate experimental validations, which demonstrate that CoCoNet outperforms the state-of-the-art self-supervised methods by a significant margin proves that such implicit consistency and complementarity preserving regularization can enhance the discriminability of latent representations.

Wenwen Qiang, Jiangmeng Li, Changwen Zheng et al.

Unsupervised domain adaptation (UDA) requires source domain samples with clean ground truth labels during training. Accurately labeling a large number of source domain samples is time-consuming and laborious. An alternative is to utilize samples with noisy labels for training. However, training with noisy labels can greatly reduce the performance of UDA. In this paper, we address the problem that learning UDA models only with access to noisy labels and propose a novel method called robust local preserving and global aligning network (RLPGA). RLPGA improves the robustness of the label noise from two aspects. One is learning a classifier by a robust informative-theoretic-based loss function. The other is constructing two adjacency weight matrices and two negative weight matrices by the proposed local preserving module to preserve the local topology structures of input data. We conduct theoretical analysis on the robustness of the proposed RLPGA and prove that the robust informative-theoretic-based loss and the local preserving module are beneficial to reduce the empirical risk of the target domain. A series of empirical studies show the effectiveness of our proposed RLPGA.

Jiahao Chen, Bin Qin, Jiangmeng Li et al.

Long-tailed learning has garnered increasing attention due to its practical significance. Among the various approaches, the fine-tuning paradigm has gained considerable interest with the advent of foundation models. However, most existing methods primarily focus on leveraging knowledge from these models, overlooking the inherent biases introduced by the imbalanced training data they rely on. In this paper, we examine how such imbalances from pre-training affect long-tailed downstream tasks. Specifically, we find the imbalance biases inherited in foundation models on downstream task as parameter imbalance and data imbalance. During fine-tuning, we observe that parameter imbalance plays a more critical role, while data imbalance can be mitigated using existing re-balancing strategies. Moreover, we find that parameter imbalance cannot be effectively addressed by current re-balancing techniques, such as adjusting the logits, during training, unlike data imbalance. To tackle both imbalances simultaneously, we build our method on causal learning and view the incomplete semantic factor as the confounder, which brings spurious correlations between input samples and labels. To resolve the negative effects of this, we propose a novel backdoor adjustment method that learns the true causal effect between input samples and labels, rather than merely fitting the correlations in the data. Notably, we achieve an average performance increase of about $1.67\%$ on each dataset. Code is available: https://github.com/JiahaoChen1/Pre-train-Imbalance

Bing Su, Ji-Rong Wen

Convolutional neural networks use regular quadrilateral convolution kernels to extract features. Since the number of parameters increases quadratically with the size of the convolution kernel, many popular models use small convolution kernels, resulting in small local receptive fields in lower layers. This paper proposes a novel log-polar space convolution (LPSC) method, where the convolution kernel is elliptical and adaptively divides its local receptive field into different regions according to the relative directions and logarithmic distances. The local receptive field grows exponentially with the number of distance levels. Therefore, the proposed LPSC not only naturally encodes local spatial structures, but also greatly increases the single-layer receptive field while maintaining the number of parameters. We show that LPSC can be implemented with conventional convolution via log-polar space pooling and can be applied in any network architecture to replace conventional convolutions. Experiments on different tasks and datasets demonstrate the effectiveness of the proposed LPSC. Code is available at https://github.com/BingSu12/Log-Polar-Space-Convolution.

Jiexin Wang, Yiju Guo, Bing Su

Human motion prediction (HMP) involves forecasting future human motion based on historical data. Graph Convolutional Networks (GCNs) have garnered widespread attention in this field for their proficiency in capturing relationships among joints in human motion. However, existing GCN-based methods tend to focus on either temporal-domain or spatial-domain features, or they combine spatio-temporal features without fully leveraging the complementarity and cross-dependency of these two features. In this paper, we propose the Spatial-Temporal Multi-Subgraph Graph Convolutional Network (STMS-GCN) to capture complex spatio-temporal dependencies in human motion. Specifically, we decouple the modeling of temporal and spatial dependencies, enabling cross-domain knowledge transfer at multiple scales through a spatio-temporal information consistency constraint mechanism. Besides, we utilize multiple subgraphs to extract richer motion information and enhance the learning associations of diverse subgraphs through a homogeneous information constraint mechanism. Extensive experiments on the standard HMP benchmarks demonstrate the superiority of our method.

Jiexin Wang, Wenwen Qiang, Zhao Yang et al.

Expressive representation of pose sequences is crucial for accurate motion modeling in human motion prediction (HMP). While recent deep learning-based methods have shown promise in learning motion representations, these methods tend to overlook the varying relevance and dependencies between historical information and future moments, with a stronger correlation for short-term predictions and weaker for distant future predictions. This limits the learning of motion representation and then hampers prediction performance. In this paper, we propose a novel approach called multi-range decoupling decoding with gating-adjusting aggregation ($MD2GA$), which leverages the temporal correlations to refine motion representation learning. This approach employs a two-stage strategy for HMP. In the first stage, a multi-range decoupling decoding adeptly adjusts feature learning by decoding the shared features into distinct future lengths, where different decoders offer diverse insights into motion patterns. In the second stage, a gating-adjusting aggregation dynamically combines the diverse insights guided by input motion data. Extensive experiments demonstrate that the proposed method can be easily integrated into other motion prediction methods and enhance their prediction performance.

Shuang Cui, Yi Li, Jiangmeng Li et al.

Single image defocus deblurring (SIDD) aims to restore an all-in-focus image from a defocused one. Distribution shifts in defocused images generally lead to performance degradation of existing methods during out-of-distribution inferences. In this work, we gauge the intrinsic reason behind the performance degradation, which is identified as the heterogeneity of lens-specific point spread functions. Empirical evidence supports this finding, motivating us to employ a continual test-time adaptation (CTTA) paradigm for SIDD. However, traditional CTTA methods, which primarily rely on entropy minimization, cannot sufficiently explore task-dependent information for pixel-level regression tasks like SIDD. To address this issue, we propose a novel Siamese networks-based continual test-time adaptation framework, which adapts source models to continuously changing target domains only requiring unlabeled target data in an online manner. To further mitigate semantically erroneous textures introduced by source SIDD models under severe degradation, we revisit the learning paradigm through a structural causal model and propose Causal Siamese networks (CauSiam). Our method leverages large-scale pre-trained vision-language models to derive discriminative universal semantic priors and integrates these priors into Siamese networks, ensuring causal identifiability between blurry inputs and restored images. Extensive experiments demonstrate that CauSiam effectively improves the generalization performance of existing SIDD methods in continuously changing domains.

Jiangmeng Li, Wenwen Qiang, Hang Gao et al.

Multi-view representation learning captures comprehensive information from multiple views of a shared context. Recent works intuitively apply contrastive learning (CL) to learn representations, regarded as a pairwise manner, which is still scalable: view-specific noise is not filtered in learning view-shared representations; the fake negative pairs, where the negative terms are actually within the same class as the positive, and the real negative pairs are coequally treated; and evenly measuring the similarities between terms might interfere with optimization. Importantly, few works research the theoretical framework of generalized self-supervised multi-view learning, especially for more than two views. To this end, we rethink the existing multi-view learning paradigm from the information theoretical perspective and then propose a novel information theoretical framework for generalized multi-view learning. Guided by it, we build a multi-view coding method with a three-tier progressive architecture, namely Information theory-guided heuristic Progressive Multi-view Coding (IPMC). In the distribution-tier, IPMC aligns the distribution between views to reduce view-specific noise. In the set-tier, IPMC builds self-adjusted pools for contrasting, which utilizes a view filter to adaptively modify the pools. Lastly, in the instance-tier, we adopt a designed unified loss to learn discriminative representations and reduce the gradient interference. Theoretically and empirically, we demonstrate the superiority of IPMC over state-of-the-art methods.

Jiahuan Zhou, Yansong Tang, Bing Su et al.

Existing popular unsupervised embedding learning methods focus on enhancing the instance-level local discrimination of the given unlabeled images by exploring various negative data. However, the existed sample outliers which exhibit large intra-class divergences or small inter-class variations severely limit their learning performance. We justify that the performance limitation is caused by the gradient vanishing on these sample outliers. Moreover, the shortage of positive data and disregard for global discrimination consideration also pose critical issues for unsupervised learning but are always ignored by existing methods. To handle these issues, we propose a novel solution to explicitly model and directly explore the uncertainty of the given unlabeled learning samples. Instead of learning a deterministic feature point for each sample in the embedding space, we propose to represent a sample by a stochastic Gaussian with the mean vector depicting its space localization and covariance vector representing the sample uncertainty. We leverage such uncertainty modeling as momentum to the learning which is helpful to tackle the outliers. Furthermore, abundant positive candidates can be readily drawn from the learned instance-specific distributions which are further adopted to mitigate the aforementioned issues. Thorough rationale analyses and extensive experiments are presented to verify our superiority.