Hang Wang

Hang Wang, Xuanhong Chen, Bingbing Ni et al.

While lightweight ViT framework has made tremendous progress in image super-resolution, its uni-dimensional self-attention modeling, as well as homogeneous aggregation scheme, limit its effective receptive field (ERF) to include more comprehensive interactions from both spatial and channel dimensions. To tackle these drawbacks, this work proposes two enhanced components under a new Omni-SR architecture. First, an Omni Self-Attention (OSA) block is proposed based on dense interaction principle, which can simultaneously model pixel-interaction from both spatial and channel dimensions, mining the potential correlations across omni-axis (i.e., spatial and channel). Coupling with mainstream window partitioning strategies, OSA can achieve superior performance with compelling computational budgets. Second, a multi-scale interaction scheme is proposed to mitigate sub-optimal ERF (i.e., premature saturation) in shallow models, which facilitates local propagation and meso-/global-scale interactions, rendering an omni-scale aggregation building block. Extensive experiments demonstrate that Omni-SR achieves record-high performance on lightweight super-resolution benchmarks (e.g., 26.95 dB@Urban100 $\times 4$ with only 792K parameters). Our code is available at \url{https://github.com/Francis0625/Omni-SR}.

Xiaohang Wang, Xuanhong Chen, Bingbing Ni et al.

Depth map super-resolution (DSR) has been a fundamental task for 3D computer vision. While arbitrary scale DSR is a more realistic setting in this scenario, previous approaches predominantly suffer from the issue of inefficient real-numbered scale upsampling. To explicitly address this issue, we propose a novel continuous depth representation for DSR. The heart of this representation is our proposed Geometric Spatial Aggregator (GSA), which exploits a distance field modulated by arbitrarily upsampled target gridding, through which the geometric information is explicitly introduced into feature aggregation and target generation. Furthermore, bricking with GSA, we present a transformer-style backbone named GeoDSR, which possesses a principled way to construct the functional mapping between local coordinates and the high-resolution output results, empowering our model with the advantage of arbitrary shape transformation ready to help diverse zooming demand. Extensive experimental results on standard depth map benchmarks, e.g., NYU v2, have demonstrated that the proposed framework achieves significant restoration gain in arbitrary scale depth map super-resolution compared with the prior art. Our codes are available at https://github.com/nana01219/GeoDSR.

Cheng Cheng, Lin Song, Ruoyi Xue et al. · tencent-ai

The contrastive vision-language pre-training, known as CLIP, demonstrates remarkable potential in perceiving open-world visual concepts, enabling effective zero-shot image recognition. Nevertheless, few-shot learning methods based on CLIP typically require offline fine-tuning of the parameters on few-shot samples, resulting in longer inference time and the risk of over-fitting in certain domains. To tackle these challenges, we propose the Meta-Adapter, a lightweight residual-style adapter, to refine the CLIP features guided by the few-shot samples in an online manner. With a few training samples, our method can enable effective few-shot learning capabilities and generalize to unseen data or tasks without additional fine-tuning, achieving competitive performance and high efficiency. Without bells and whistles, our approach outperforms the state-of-the-art online few-shot learning method by an average of 3.6\% on eight image classification datasets with higher inference speed. Furthermore, our model is simple and flexible, serving as a plug-and-play module directly applicable to downstream tasks. Without further fine-tuning, Meta-Adapter obtains notable performance improvements in open-vocabulary object detection and segmentation tasks.

PSBC LLM Team, Huawei LLM Team, Ruihan Long et al.

Large language models (LLMs) drive significant financial innovations, yet their high-concurrency deployment is severely bottlenecked by KV cache memory overhead, which inflates infrastructure costs and throttles scalability. To address this, we propose YouZhi-LLM, a highly efficient financial LLM empowered by a comprehensive structural transition and training pipeline natively built on the Huawei Ascend ecosystem. At its algorithmic core, YouZhi-LLM features a layer-adaptive GQA-to-MLA transition framework that dynamically assigns per-layer FreqFold sizes, maximizing KV-cache compression while minimizing perplexity degradation. To recover representation capacity and inject domain expertise, the Ascend-based training pipeline seamlessly integrates generalized knowledge distillation with financial-specific supervised fine-tuning. Evaluations demonstrate the superiority of this systematic approach, with the adaptive transition reducing perplexity degradation by up to 35% over uniform baselines. Crucially, when evaluated on Ascend NPUs via vLLM-Ascend, the massive KV-cache reduction translates directly into deployment efficiency. Compared to their respective base models, YouZhi-7B yields a 12.3% improvement in average financial benchmark score alongside a 2.69$\times$ increase in maximum concurrency; similarly, YouZhi-14B achieves a 7.0% accuracy gain and a 2.43$\times$ concurrency boost, establishing a new paradigm for cost-effective, high-throughput financial inference.

Hang Wang, Zhen Xiang, David J. Miller et al.

Backdoor attacks are an important type of adversarial threat against deep neural network classifiers, wherein test samples from one or more source classes will be (mis)classified to the attacker's target class when a backdoor pattern is embedded. In this paper, we focus on the post-training backdoor defense scenario commonly considered in the literature, where the defender aims to detect whether a trained classifier was backdoor-attacked without any access to the training set. Many post-training detectors are designed to detect attacks that use either one or a few specific backdoor embedding functions (e.g., patch-replacement or additive attacks). These detectors may fail when the backdoor embedding function used by the attacker (unknown to the defender) is different from the backdoor embedding function assumed by the defender. In contrast, we propose a post-training defense that detects backdoor attacks with arbitrary types of backdoor embeddings, without making any assumptions about the backdoor embedding type. Our detector leverages the influence of the backdoor attack, independent of the backdoor embedding mechanism, on the landscape of the classifier's outputs prior to the softmax layer. For each class, a maximum margin statistic is estimated. Detection inference is then performed by applying an unsupervised anomaly detector to these statistics. Thus, our detector does not need any legitimate clean samples, and can efficiently detect backdoor attacks with arbitrary numbers of source classes. These advantages over several state-of-the-art methods are demonstrated on four datasets, for three different types of backdoor patterns, and for a variety of attack configurations. Finally, we propose a novel, general approach for backdoor mitigation once a detection is made. The mitigation approach was the runner-up at the first IEEE Trojan Removal Competition. The code is online available.

Hang Wang, Sen Lin, Junshan Zhang

The ensemble method is a promising way to mitigate the overestimation issue in Q-learning, where multiple function approximators are used to estimate the action values. It is known that the estimation bias hinges heavily on the ensemble size (i.e., the number of Q-function approximators used in the target), and that determining the `right' ensemble size is highly nontrivial, because of the time-varying nature of the function approximation errors during the learning process. To tackle this challenge, we first derive an upper bound and a lower bound on the estimation bias, based on which the ensemble size is adapted to drive the bias to be nearly zero, thereby coping with the impact of the time-varying approximation errors accordingly. Motivated by the theoretic findings, we advocate that the ensemble method can be combined with Model Identification Adaptive Control (MIAC) for effective ensemble size adaptation. Specifically, we devise Adaptive Ensemble Q-learning (AdaEQ), a generalized ensemble method with two key steps: (a) approximation error characterization which serves as the feedback for flexibly controlling the ensemble size, and (b) ensemble size adaptation tailored towards minimizing the estimation bias. Extensive experiments are carried out to show that AdaEQ can improve the learning performance than the existing methods for the MuJoCo benchmark.

Hang Wang, Zhen Xiang, David J. Miller et al.

Deep neural networks are vulnerable to backdoor attacks (Trojans), where an attacker poisons the training set with backdoor triggers so that the neural network learns to classify test-time triggers to the attacker's designated target class. Recent work shows that backdoor poisoning induces over-fitting (abnormally large activations) in the attacked model, which motivates a general, post-training clipping method for backdoor mitigation, i.e., with bounds on internal-layer activations learned using a small set of clean samples. We devise a new such approach, choosing the activation bounds to explicitly limit classification margins. This method gives superior performance against peer methods for CIFAR-10 image classification. We also show that this method has strong robustness against adaptive attacks, X2X attacks, and on different datasets. Finally, we demonstrate a method extension for test-time detection and correction based on the output differences between the original and activation-bounded networks. The code of our method is online available.

Hang Wang, David J. Miller, George Kesidis

Well-known (non-malicious) sources of overfitting in deep neural net (DNN) classifiers include: i) large class imbalances; ii) insufficient training-set diversity; and iii) over-training. In recent work, it was shown that backdoor data-poisoning also induces overfitting, with unusually large classification margins to the attacker's target class, mediated particularly by (unbounded) ReLU activations that allow large signals to propagate in the DNN. Thus, an effective post-training (with no knowledge of the training set or training process) mitigation approach against backdoors was proposed, leveraging a small clean dataset, based on bounding neural activations. Improving upon that work, we threshold activations specifically to limit maximum margins (MMs), which yields performance gains in backdoor mitigation. We also provide some analytical support for this mitigation approach. Most importantly, we show that post-training MM-based regularization substantially mitigates non-malicious overfitting due to class imbalances and overtraining. Thus, unlike adversarial training, which provides some resilience against attacks but which harms clean (attack-free) generalization, we demonstrate an approach originating from adversarial learning that helps clean generalization accuracy. Experiments on CIFAR-10 and CIFAR-100, in comparison with peer methods, demonstrate strong performance of our methods.

Chenhao Lin, Fangbin Yi, Hang Wang et al.

Face forgery techniques have emerged as a forefront concern, and numerous detection approaches have been proposed to address this challenge. However, existing methods predominantly concentrate on single-face manipulation detection, leaving the more intricate and realistic realm of multi-face forgeries relatively unexplored. This paper proposes a novel framework explicitly tailored for multi-face forgery detection,filling a critical gap in the current research. The framework mainly involves two modules:(i) a facial relationships learning module, which generates distinguishable local features for each face within images,(ii) a global feature aggregation module that leverages the mutual constraints between global and local information to enhance forgery detection accuracy.Our experimental results on two publicly available multi-face forgery datasets demonstrate that the proposed approach achieves state-of-the-art performance in multi-face forgery detection scenarios.

Hang Wang, Sen Lin, Junshan Zhang

Warm-Start reinforcement learning (RL), aided by a prior policy obtained from offline training, is emerging as a promising RL approach for practical applications. Recent empirical studies have demonstrated that the performance of Warm-Start RL can be improved \textit{quickly} in some cases but become \textit{stagnant} in other cases, especially when the function approximation is used. To this end, the primary objective of this work is to build a fundamental understanding on ``\textit{whether and when online learning can be significantly accelerated by a warm-start policy from offline RL?}''. Specifically, we consider the widely used Actor-Critic (A-C) method with a prior policy. We first quantify the approximation errors in the Actor update and the Critic update, respectively. Next, we cast the Warm-Start A-C algorithm as Newton's method with perturbation, and study the impact of the approximation errors on the finite-time learning performance with inaccurate Actor/Critic updates. Under some general technical conditions, we derive the upper bounds, which shed light on achieving the desired finite-learning performance in the Warm-Start A-C algorithm. In particular, our findings reveal that it is essential to reduce the algorithm bias in online learning. We also obtain lower bounds on the sub-optimality gap of the Warm-Start A-C algorithm to quantify the impact of the bias and error propagation.

Hang Wang, Chao Shen, Chenhao Lin et al.

The proliferation of advanced AI video synthesis techniques poses an unprecedented challenge to digital video authenticity. Existing AI-generated video (AIGV) detection methods primarily focus on uni-modal or spatiotemporal artifacts, but they overlook the rich cues within the visual-textual cross-modal space, especially the temporal stability of semantic alignment. In this work, we identify a distinctive fingerprint in AIGVs, termed cross-modal temporal artifact (CMTA). Unlike real videos that exhibit natural temporal fluctuations in cross-modal alignment due to semantic variations, AIGVs display unnaturally stable semantic trajectories governed by given input prompts. To bridge this gap, we propose the CMTA framework, a cross-modal detection approach that captures these unique temporal artifacts through joint cross-modal embedding and multi-grained temporal modeling. Specifically, CMTA leverages BLIP to generate frame-level image captions and utilizes CLIP to extract corresponding visual-textual representations. A coarse-grained temporal modeling branch is then designed to characterize temporal fluctuations in cross-modal alignment with a GRU. In parallel, a fine-grained branch is constructed to capture intricate inter-frame variations from integrated visual-textual features with a Transformer encoder. Extensive experiments on 40 subsets across four large-scale datasets, including GenVideo, EvalCrafter, VideoPhy, and VidProM, validate that our approach sets a new state-of-the-art while exhibiting superior cross-generator generalization. Code and models of CMTA will be released at https://github.com/hwang-cs-ime/CMTA

Pan Wang, Yihao Hu, Xiujin Liu et al.

Vision-language model (VLM) agents increasingly rely on memory-augmented reinforcement learning to reuse experience across long-horizon tasks, yet most existing frameworks store memory as text and depend on proprietary teacher models to summarize or refine it. This design is poorly matched to spatial decision making: geometric priors are compressed into lossy language, and sparse interaction is often supervised through delayed textual feedback rather than dense visually grounded signals. We argue that reusable experience for VLM agents should remain visually grounded. Based on this insight, we propose \textbf{AtlasVA}, a teacher-free visual skill memory framework that organizes memory into three complementary layers: spatial heatmaps, visual exemplars, and symbolic text skills. AtlasVA further evolves danger and affinity atlases directly from trajectory statistics and lightweight grid heuristics, and reuses these self-evolving atlases as potential-based shaping rewards for reinforcement learning. This unifies perception, memory, and optimization without external LLM supervision. Experiments on \textsc{Sokoban}, \textsc{FrozenLake}, 3D embodied navigation, and 3D robotic manipulation benchmarks show that AtlasVA consistently outperforms text-centric memory baselines and competitive VLM agents, with especially strong gains on spatially intensive tasks. Homepage: https://wangpan-ustc.github.io/AtlasvaWeb

Dechen Gao, Shuangyu Cai, Hanchu Zhou et al.

To safely navigate intricate real-world scenarios, autonomous vehicles must be able to adapt to diverse road conditions and anticipate future events. World model (WM) based reinforcement learning (RL) has emerged as a promising approach by learning and predicting the complex dynamics of various environments. Nevertheless, to the best of our knowledge, there does not exist an accessible platform for training and testing such algorithms in sophisticated driving environments. To fill this void, we introduce CarDreamer, the first open-source learning platform designed specifically for developing WM based autonomous driving algorithms. It comprises three key components: 1) World model backbone: CarDreamer has integrated some state-of-the-art WMs, which simplifies the reproduction of RL algorithms. The backbone is decoupled from the rest and communicates using the standard Gym interface, so that users can easily integrate and test their own algorithms. 2) Built-in tasks: CarDreamer offers a comprehensive set of highly configurable driving tasks which are compatible with Gym interfaces and are equipped with empirically optimized reward functions. 3) Task development suite: This suite streamlines the creation of driving tasks, enabling easy definition of traffic flows and vehicle routes, along with automatic collection of multi-modal observation data. A visualization server allows users to trace real-time agent driving videos and performance metrics through a browser. Furthermore, we conduct extensive experiments using built-in tasks to evaluate the performance and potential of WMs in autonomous driving. Thanks to the richness and flexibility of CarDreamer, we also systematically study the impact of observation modality, observability, and sharing of vehicle intentions on AV safety and efficiency. All code and documents are accessible on https://github.com/ucd-dare/CarDreamer.

Dechen Gao, Hang Wang, Hanchu Zhou et al.

Imitation learning (IL) and reinforcement learning (RL) each offer distinct advantages for robotics policy learning: IL provides stable learning from demonstrations, and RL promotes generalization through exploration. While existing robot learning approaches using IL-based pre-training followed by RL-based fine-tuning are promising, this two-step learning paradigm often suffers from instability and poor sample efficiency during the RL fine-tuning phase. In this work, we introduce IN-RIL, INterleaved Reinforcement learning and Imitation Learning, for policy fine-tuning, which periodically injects IL updates after multiple RL updates and hence can benefit from the stability of IL and the guidance of expert data for more efficient exploration throughout the entire fine-tuning process. Since IL and RL involve different optimization objectives, we develop gradient separation mechanisms to prevent destructive interference during \ABBR fine-tuning, by separating possibly conflicting gradient updates in orthogonal subspaces. Furthermore, we conduct rigorous analysis, and our findings shed light on why interleaving IL with RL stabilizes learning and improves sample-efficiency. Extensive experiments on 14 robot manipulation and locomotion tasks across 3 benchmarks, including FurnitureBench, OpenAI Gym, and Robomimic, demonstrate that \ABBR can significantly improve sample efficiency and mitigate performance collapse during online finetuning in both long- and short-horizon tasks with either sparse or dense rewards. IN-RIL, as a general plug-in compatible with various state-of-the-art RL algorithms, can significantly improve RL fine-tuning, e.g., from 12\% to 88\% with 6.3x improvement in the success rate on Robomimic Transport. Project page: https://github.com/ucd-dare/IN-RIL.

Hang Wang, Chao Shen, Lei Zhang et al.

AI-generated videos (AIGVs) have achieved unprecedented photorealism, posing severe threats to digital forensics. Existing AIGV detectors focus mainly on localized artifacts or short-term temporal inconsistencies, thus often fail to capture the underlying generative logic governing global temporal evolution, limiting AIGV detection performance. In this paper, we identify a distinctive fingerprint in AIGVs, termed anomalous temporal self-similarity (ATSS). Unlike real videos that exhibit stochastic natural dynamics, AIGVs follow deterministic anchor-driven trajectories (e.g., text or image prompts), inducing unnaturally repetitive correlations across visual and semantic domains. To exploit this, we propose the ATSS method, a multimodal detection framework that exploits this insight via a triple-similarity representation and a cross-attentive fusion mechanism. Specifically, ATSS reconstructs semantic trajectories by leveraging frame-wise descriptions to construct visual, textual, and cross-modal similarity matrices, which jointly quantify the inherent temporal anomalies. These matrices are encoded by dedicated Transformer encoders and integrated via a bidirectional cross-attentive fusion module to effectively model intra- and inter-modal dynamics. Extensive experiments on four large-scale benchmarks, including GenVideo, EvalCrafter, VideoPhy, and VidProM, demonstrate that ATSS significantly outperforms state-of-the-art methods in terms of AP, AUC, and ACC metrics, exhibiting superior generalization across diverse video generation models. Code and models of ATSS will be released at https://github.com/hwang-cs-ime/ATSS.

Hang Wang, Zhi-Qi Cheng, Chenhao Lin et al.

Text-to-image synthesis has progressed to the point where models can generate visually compelling images from natural language prompts. Yet, existing methods often fail to reconcile high-level semantic fidelity with explicit spatial control, particularly in scenes involving multiple objects, nuanced relations, or complex layouts. To bridge this gap, we propose a Hierarchical Cross-Modal Alignment (HCMA) framework for grounded text-to-image generation. HCMA integrates two alignment modules into each diffusion sampling step: a global module that continuously aligns latent representations with textual descriptions to ensure scene-level coherence, and a local module that employs bounding-box layouts to anchor objects at specified locations, enabling fine-grained spatial control. Extensive experiments on the MS-COCO 2014 validation set show that HCMA surpasses state-of-the-art baselines, achieving a 0.69 improvement in Frechet Inception Distance (FID) and a 0.0295 gain in CLIP Score. These results demonstrate HCMA's effectiveness in faithfully capturing intricate textual semantics while adhering to user-defined spatial constraints, offering a robust solution for semantically grounded image generation. Our code is available at https://github.com/hwang-cs-ime/HCMA.

Minghao Xu, Hang Wang, Bingbing Ni et al.

Autonomous highlight detection is crucial for enhancing the efficiency of video browsing on social media platforms. To attain this goal in a data-driven way, one may often face the situation where highlight annotations are not available on the target video category used in practice, while the supervision on another video category (named as source video category) is achievable. In such a situation, one can derive an effective highlight detector on target video category by transferring the highlight knowledge acquired from source video category to the target one. We call this problem cross-category video highlight detection, which has been rarely studied in previous works. For tackling such practical problem, we propose a Dual-Learner-based Video Highlight Detection (DL-VHD) framework. Under this framework, we first design a Set-based Learning module (SL-module) to improve the conventional pair-based learning by assessing the highlight extent of a video segment under a broader context. Based on such learning manner, we introduce two different learners to acquire the basic distinction of target category videos and the characteristics of highlight moments on source video category, respectively. These two types of highlight knowledge are further consolidated via knowledge distillation. Extensive experiments on three benchmark datasets demonstrate the superiority of the proposed SL-module, and the DL-VHD method outperforms five typical Unsupervised Domain Adaptation (UDA) algorithms on various cross-category highlight detection tasks. Our code is available at https://github.com/ChrisAllenMing/Cross_Category_Video_Highlight .

Linguo Li, Minsi Wang, Bingbing Ni et al.

In this work, we propose a Cross-view Contrastive Learning framework for unsupervised 3D skeleton-based action Representation (CrosSCLR), by leveraging multi-view complementary supervision signal. CrosSCLR consists of both single-view contrastive learning (SkeletonCLR) and cross-view consistent knowledge mining (CVC-KM) modules, integrated in a collaborative learning manner. It is noted that CVC-KM works in such a way that high-confidence positive/negative samples and their distributions are exchanged among views according to their embedding similarity, ensuring cross-view consistency in terms of contrastive context, i.e., similar distributions. Extensive experiments show that CrosSCLR achieves remarkable action recognition results on NTU-60 and NTU-120 datasets under unsupervised settings, with observed higher-quality action representations. Our code is available at https://github.com/LinguoLi/CrosSCLR.

Hang Wang, Minghao Xu, Bingbing Ni et al.

Transferring knowledges learned from multiple source domains to target domain is a more practical and challenging task than conventional single-source domain adaptation. Furthermore, the increase of modalities brings more difficulty in aligning feature distributions among multiple domains. To mitigate these problems, we propose a Learning to Combine for Multi-Source Domain Adaptation (LtC-MSDA) framework via exploring interactions among domains. In the nutshell, a knowledge graph is constructed on the prototypes of various domains to realize the information propagation among semantically adjacent representations. On such basis, a graph model is learned to predict query samples under the guidance of correlated prototypes. In addition, we design a Relation Alignment Loss (RAL) to facilitate the consistency of categories' relational interdependency and the compactness of features, which boosts features' intra-class invariance and inter-class separability. Comprehensive results on public benchmark datasets demonstrate that our approach outperforms existing methods with a remarkable margin. Our code is available at \url{https://github.com/ChrisAllenMing/LtC-MSDA}

Minghao Xu, Hang Wang, Bingbing Ni et al.

Applying the knowledge of an object detector trained on a specific domain directly onto a new domain is risky, as the gap between two domains can severely degrade model's performance. Furthermore, since different instances commonly embody distinct modal information in object detection scenario, the feature alignment of source and target domain is hard to be realized. To mitigate these problems, we propose a Graph-induced Prototype Alignment (GPA) framework to seek for category-level domain alignment via elaborate prototype representations. In the nutshell, more precise instance-level features are obtained through graph-based information propagation among region proposals, and, on such basis, the prototype representation of each class is derived for category-level domain alignment. In addition, in order to alleviate the negative effect of class-imbalance on domain adaptation, we design a Class-reweighted Contrastive Loss to harmonize the adaptation training process. Combining with Faster R-CNN, the proposed framework conducts feature alignment in a two-stage manner. Comprehensive results on various cross-domain detection tasks demonstrate that our approach outperforms existing methods with a remarkable margin. Our code is available at https://github.com/ChrisAllenMing/GPA-detection.

Pan Wang, Yihao Hu, Xiujin Liu et al.

Traditional shadow removal networks often treat image restoration as an unconstrained mapping, lacking the physical interpretability required to balance localized texture recovery with global illumination consistency. To address this, we propose CFSR, a multi-modal prior-driven framework that reframes shadow removal as a physics-constrained restoration process. By seamlessly integrating 3D geometric cues with large-scale foundation model semantics, CFSR effectively bridges the 2D-3D domain gap. Specifically, we first map observations into a custom HVI color space to suppress shadow-induced noise and robustly fuse RGB data with estimated depth priors. At its core, our Geometric & Semantic Dual Explicit Guided Attention mechanism utilizes DINO features and 3D surface normals to directly modulate the attention affinity matrix, structurally enforcing physical lighting constraints. To recover severely degraded regions, we inject holistic priors via a frozen CLIP encoder. Finally, our Frequency Collaborative Reconstruction Module (FCRM) achieves an optimal synthesis by decoupling the decoding process. Conditioned on geometric priors, FCRM seamlessly harmonizes the reconstruction of sharp high-frequency occlusion boundaries with the restoration of low-frequency global illumination. Extensive experiments demonstrate that CFSR achieves state-of-the-art performance across multiple challenging benchmarks.

Cheng Cheng, Hang Wang, Hongbin Sun

The prevalence of convolution neural networks (CNNs) and vision transformers (ViTs) has markedly revolutionized the area of single-image super-resolution (SISR). To further boost the SR performances, several techniques, such as residual learning and attention mechanism, are introduced, which can be largely attributed to a wider range of activated area, that is, the input pixels that strongly influence the SR results. However, the possibility of further improving SR performance through another versatile vision backbone remains an unresolved challenge. To address this issue, in this paper, we unleash the representation potential of the modern state space model, i.e., Vision Mamba (Vim), in the context of SISR. Specifically, we present three recipes for better utilization of Vim-based models: 1) Integration into a MetaFormer-style block; 2) Pre-training on a larger and broader dataset; 3) Employing complementary attention mechanism, upon which we introduce the MMA. The resulting network MMA is capable of finding the most relevant and representative input pixels to reconstruct the corresponding high-resolution images. Comprehensive experimental analysis reveals that MMA not only achieves competitive or even superior performance compared to state-of-the-art SISR methods but also maintains relatively low memory and computational overheads (e.g., +0.5 dB PSNR elevation on Manga109 dataset with 19.8 M parameters at the scale of 2). Furthermore, MMA proves its versatility in lightweight SR applications. Through this work, we aim to illuminate the potential applications of state space models in the broader realm of image processing rather than SISR, encouraging further exploration in this innovative direction.

Hang Wang, Xin Ye, Feng Tao et al.

World model based reinforcement learning (RL) has emerged as a promising approach for autonomous driving, which learns a latent dynamics model and uses it to train a planning policy. To speed up the learning process, the pretrain-finetune paradigm is often used, where online RL is initialized by a pretrained model and a policy learned offline. However, naively performing such initialization in RL may result in dramatic performance degradation during the online interactions in the new task. To tackle this challenge, we first analyze the performance degradation and identify two primary root causes therein: the mismatch of the planning policy and the mismatch of the dynamics model, due to distribution shift. We further analyze the effects of these factors on performance degradation during finetuning, and our findings reveal that the choice of finetuning strategies plays a pivotal role in mitigating these effects. We then introduce AdaWM, an Adaptive World Model based planning method, featuring two key steps: (a) mismatch identification, which quantifies the mismatches and informs the finetuning strategy, and (b) alignment-driven finetuning, which selectively updates either the policy or the model as needed using efficient low-rank updates. Extensive experiments on the challenging CARLA driving tasks demonstrate that AdaWM significantly improves the finetuning process, resulting in more robust and efficient performance in autonomous driving systems.

Zhoulin Ji, Chenhao Lin, Hang Wang et al.

Detecting synthetic from real speech is increasingly crucial due to the risks of misinformation and identity impersonation. While various datasets for synthetic speech analysis have been developed, they often focus on specific areas, limiting their utility for comprehensive research. To fill this gap, we propose the Speech-Forensics dataset by extensively covering authentic, synthetic, and partially forged speech samples that include multiple segments synthesized by different high-quality algorithms. Moreover, we propose a TEmporal Speech LocalizaTion network, called TEST, aiming at simultaneously performing authenticity detection, multiple fake segments localization, and synthesis algorithms recognition, without any complex post-processing. TEST effectively integrates LSTM and Transformer to extract more powerful temporal speech representations and utilizes dense prediction on multi-scale pyramid features to estimate the synthetic spans. Our model achieves an average mAP of 83.55% and an EER of 5.25% at the utterance level. At the segment level, it attains an EER of 1.07% and a 92.19% F1 score. These results highlight the model's robust capability for a comprehensive analysis of synthetic speech, offering a promising avenue for future research and practical applications in this field.

Shibin Mei, Hang Wang, Bingbing Ni

The cameras equipped on mobile terminals employ different sensors in different photograph modes, and the transferability of raw domain denoising models between these sensors is significant but remains sufficient exploration. Industrial solutions either develop distinct training strategies and models for different sensors or ignore the differences between sensors and simply extend existing models to new sensors, which leads to tedious training or unsatisfactory performance. In this paper, we introduce a new benchmark, the Multi-Sensor SIDD (MSSIDD) dataset, which is the first raw-domain dataset designed to evaluate the sensor transferability of denoising models. The MSSIDD dataset consists of 60,000 raw images of six distinct sensors, derived through the degeneration of sRGB images via different camera sensor parameters. Furthermore, we propose a sensor consistency training framework that enables denoising models to learn the sensor-invariant features, thereby facilitating the generalization of the consistent model to unseen sensors. We evaluate previous arts on the newly proposed MSSIDD dataset, and the experimental results validate the effectiveness of our proposed method. Our dataset is available at https://www.kaggle.com/datasets/sjtuwh/mssidd.

Hang Wang, Huijie Tang, Ningai Leng et al.

Based on economic theories and integrated with machine learning technology, this study explores a collaborative Supply Chain Management and Financial Supply Chain Management (SCM - FSCM) model to solve issues like efficiency loss, financing constraints, and risk transmission. We combine Transaction Cost and Information Asymmetry theories and use algorithms such as random forests to process multi-dimensional data and build a data-driven, three-dimensional (cost-efficiency-risk) analysis framework. We then apply an FSCM model of "core enterprise credit empowerment plus dynamic pledge financing." We use Long Short-Term Memory (LSTM) networks for demand forecasting and clustering/regression algorithms for benefit allocation. The study also combines Game Theory and reinforcement learning to optimize the inventory-procurement mechanism and uses eXtreme Gradient Boosting (XGBoost) for credit assessment to enable rapid monetization of inventory. Verified with 20 core and 100 supporting enterprises, the results show a 30\% increase in inventory turnover, an 18\%-22\% decrease in SME financing costs, a stable order fulfillment rate above 95\%, and excellent model performance (demand forecasting error <= 8\%, credit assessment accuracy >= 90\%). This SCM-FSCM model effectively reduces operating costs, alleviates financing constraints, and supports high-quality supply chain development.

Fengyi Wu, Yifei Dong, Zhi-Qi Cheng et al.

We introduce Goal-Conditioned Visual Navigation Instruction Generation (GoViG), a new task that aims to autonomously generate precise and contextually coherent navigation instructions solely from egocentric visual observations of initial and goal states. Unlike conventional approaches that rely on structured inputs such as semantic annotations or environmental maps, GoViG exclusively leverages raw egocentric visual data, substantially improving its adaptability to unseen and unstructured environments. Our method addresses this task by decomposing it into two interconnected subtasks: (1) visual forecasting, which predicts intermediate visual states bridging the initial and goal views; and (2) instruction generation, which synthesizes linguistically coherent instructions grounded in both observed and anticipated visuals. These subtasks are integrated within an autoregressive multimodal large language model trained with tailored objectives to ensure spatial accuracy and linguistic clarity. Furthermore, we introduce two complementary multimodal reasoning strategies, one-pass and interleaved reasoning, to mimic incremental human cognitive processes during navigation. To evaluate our method, we propose the R2R-Goal dataset, combining diverse synthetic and real-world trajectories. Empirical results demonstrate significant improvements over state-of-the-art methods, achieving superior BLEU-4 and CIDEr scores along with robust cross-domain generalization.

Hang Wang, Dechen Gao, Junshan Zhang

We study multi-agent reinforcement learning (MARL) for tasks in complex high-dimensional environments, such as autonomous driving. MARL is known to suffer from the \textit{partial observability} and \textit{non-stationarity} issues. To tackle these challenges, information sharing is often employed, which however faces major hurdles in practice, including overwhelming communication overhead and scalability concerns. By making use of generative AI embodied in world model together with its latent representation, we develop {\it CALL}, \underline{C}ommunic\underline{a}tive Wor\underline{l}d Mode\underline{l}, for MARL, where 1) each agent first learns its world model that encodes its state and intention into low-dimensional latent representation with smaller memory footprint, which can be shared with other agents of interest via lightweight communication; and 2) each agent carries out ego-centric learning while exploiting lightweight information sharing to enrich her world model, and then exploits its generalization capacity to improve prediction for better planning. We characterize the gain on the prediction accuracy from the information sharing and its impact on performance gap. Extensive experiments are carried out on the challenging local trajectory planning tasks in the CARLA platform to demonstrate the performance gains of using \textit{CALL}.

Hang Wang, Zhi-Qi Cheng, Youtian Du et al. · cmu, uw

Video Action Counting (VAC) is crucial in analyzing sports, fitness, and everyday activities by quantifying repetitive actions in videos. However, traditional VAC methods have overlooked the complexity of action repetitions, such as interruptions and the variability in cycle duration. Our research addresses the shortfall by introducing a novel approach to VAC, called Irregular Video Action Counting (IVAC). IVAC prioritizes modeling irregular repetition patterns in videos, which we define through two primary aspects: Inter-cycle Consistency and Cycle-interval Inconsistency. Inter-cycle Consistency ensures homogeneity in the spatial-temporal representations of cycle segments, signifying action uniformity within cycles. Cycle-interval inconsistency highlights the importance of distinguishing between cycle segments and intervals based on their inherent content differences. To encapsulate these principles, we propose a new methodology that includes consistency and inconsistency modules, supported by a unique pull-push loss (P2L) mechanism. The IVAC-P2L model applies a pull loss to promote coherence among cycle segment features and a push loss to clearly distinguish features of cycle segments from interval segments. Empirical evaluations conducted on the RepCount dataset demonstrate that the IVAC-P2L model sets a new benchmark in VAC task performance. Furthermore, the model demonstrates exceptional adaptability and generalization across various video contents, outperforming existing models on two additional datasets, UCFRep and Countix, without the need for dataset-specific optimization. These results confirm the efficacy of our approach in addressing irregular repetitions in videos and pave the way for further advancements in video analysis and understanding.

Dechen Gao, Boqi Zhao, Andrew Lee et al.

Conventional flow matching and diffusion-based policies sample through iterative denoising from standard noise distributions (e.g., Gaussian), and require conditioning mechanisms to incorporate visual information during the generative process, incurring substantial time and memory overhead. To reduce the complexity, we develop VITA(VIsion-To-Action policy), a noise-free and conditioning-free policy learning framework that directly maps visual representations to latent actions using flow matching. VITA treats latent visual representations as the source of the flow, thus eliminating the need of conditioning. As expected, bridging vision and action is challenging, because actions are lower-dimensional, less structured, and sparser than visual representations; moreover, flow matching requires the source and target to have the same dimensionality. To overcome this, we introduce an action autoencoder that maps raw actions into a structured latent space aligned with visual latents, trained jointly with flow matching. To further prevent latent space collapse, we propose flow latent decoding, which anchors the latent generation process by backpropagating the action reconstruction loss through the flow matching ODE (ordinary differential equations) solving steps. We evaluate VITA on 8 simulation and 2 real-world tasks from ALOHA and Robomimic. VITA outperforms or matches state-of-the-art generative policies, while achieving 1.5-2.3x faster inference compared to conventional methods with conditioning. Project page: https://ucd-dare.github.io/VITA/

Guangmingmei Yang, Xi Li, Hang Wang et al.

A variety of defenses have been proposed against Trojans planted in (backdoor attacks on) deep neural network (DNN) classifiers. Backdoor-agnostic methods seek to reliably detect and/or to mitigate backdoors irrespective of the incorporation mechanism used by the attacker, while inversion methods explicitly assume one. In this paper, we describe a new detector that: relies on embedded feature representations to estimate (invert) the backdoor and to identify its target class; can operate without access to the training dataset; and is highly effective for various incorporation mechanisms (i.e., is backdoor agnostic). Our detection approach is evaluated -- and found to be favorable - in comparison with an array of published defenses for a variety of different attacks on the CIFAR-10 and CIFAR-100 image-classification domains.

Yihao Hu, Pan Wang, Xiaodong Bai et al.

Pomelo detection is an essential process for their localization, automated robotic harvesting, and maturity analysis. However, detecting Shatian pomelo in complex orchard environments poses significant challenges, including multi-scale issues, obstructions from trunks and leaves, small object detection, etc. To address these issues, this study constructs a custom dataset STP-AgriData and proposes the SDE-DET model for Shatian pomelo detection. SDE-DET first utilizes the Star Block to effectively acquire high-dimensional information without increasing the computational overhead. Furthermore, the presented model adopts Deformable Attention in its backbone, to enhance its ability to detect pomelos under occluded conditions. Finally, multiple Efficient Multi-Scale Attention mechanisms are integrated into our model to reduce the computational overhead and extract deep visual representations, thereby improving the capacity for small object detection. In the experiment, we compared SDE-DET with the Yolo series and other mainstream detection models in Shatian pomelo detection. The presented SDE-DET model achieved scores of 0.883, 0.771, 0.838, 0.497, and 0.823 in Precision, Recall, mAP@0.5, mAP@0.5:0.95 and F1-score, respectively. SDE-DET has achieved state-of-the-art performance on the STP-AgriData dataset. Experiments indicate that the SDE-DET provides a reliable method for Shatian pomelo detection, laying the foundation for the further development of automatic harvest robots.

Hang Wang, Junshan Zhang

Multi-agent reinforcement learning faces fundamental challenges that conventional approaches have failed to overcome: exponentially growing joint action spaces, non-stationary environments where simultaneous learning creates moving targets, and partial observability that constrains coordination. Current methods remain reactive, employing stimulus-response mechanisms that fail when facing novel scenarios. We argue for a transformative paradigm shift from reactive to proactive multi-agent intelligence through generative AI-based reinforcement learning. This position advocates reconceptualizing agents not as isolated policy optimizers, but as sophisticated generative models capable of synthesizing complex multi-agent dynamics and making anticipatory decisions based on predictive understanding of future interactions. Rather than responding to immediate observations, generative-RL agents can model environment evolution, predict other agents' behaviors, generate coordinated action sequences, and engage in strategic reasoning accounting for long-term dynamics. This approach leverages pattern recognition and generation capabilities of generative AI to enable proactive decision-making, seamless coordination through enhanced communication, and dynamic adaptation to evolving scenarios. We envision this paradigm shift will unlock unprecedented possibilities for distributed intelligence, moving beyond individual optimization toward emergent collective behaviors representing genuine collaborative intelligence. The implications extend across autonomous systems, robotics, and human-AI collaboration, promising solutions to coordination challenges intractable under traditional reactive frameworks.

Hang Wang, Qiaoyi Fang, Junshan Zhang

The past few years have witnessed a rapid growth of the deployment of automated vehicles (AVs). Clearly, AVs and human-driven vehicles (HVs) will co-exist for many years, and AVs will have to operate around HVs, pedestrians, cyclists, and more, calling for fundamental breakthroughs in AI designed for mixed traffic to achieve mixed autonomy. Thus motivated, we study heterogeneous decision making by AVs and HVs in a mixed traffic environment, aiming to capture the interactions between human and machine decision-making and develop an AI foundation that enables vehicles to operate safely and efficiently. There are a number of challenges to achieve mixed autonomy, including 1) humans drivers make driving decisions with bounded rationality, and it remains open to develop accurate models for HVs' decision making; and 2) uncertainty-aware planning plays a critical role for AVs to take safety maneuvers in response to the human behavior. In this paper, we introduce a formulation of AV-HV interaction, where the HV makes decisions with bounded rationality and the AV employs uncertainty-aware planning based on the prediction on HV's future actions. We conduct a comprehensive analysis on AV and HV's learning regret to answer the questions: 1) {How does the learning performance depend on HV's bounded rationality and AV's planning}; 2) {How do different decision making strategies impact the overall learning performance}? Our findings reveal some intriguing phenomena, such as Goodhart's Law in AV's learning performance and compounding effects in HV's decision making process. By examining the dynamics of the regrets, we gain insights into the interplay between human and machine decision making.

Qiaoyi Fang, Weiyu Du, Hang Wang et al.

World models have recently emerged as a promising approach to reinforcement learning (RL), achieving state-of-the-art performance across a wide range of visual control tasks. This work aims to obtain a deep understanding of the robustness and generalization capabilities of world models. Thus motivated, we develop a stochastic differential equation formulation by treating the world model learning as a stochastic dynamical system, and characterize the impact of latent representation errors on robustness and generalization, for both cases with zero-drift representation errors and with non-zero-drift representation errors. Our somewhat surprising findings, based on both theoretic and experimental studies, reveal that for the case with zero drift, modest latent representation errors can in fact function as implicit regularization and hence result in improved robustness. We further propose a Jacobian regularization scheme to mitigate the compounding error propagation effects of non-zero drift, thereby enhancing training stability and robustness. Our experimental studies corroborate that this regularization approach not only stabilizes training but also accelerates convergence and improves accuracy of long-horizon prediction.

Xiao-Jun Wu, Cai-Jun Zhao, Chun Meng et al.

The p16/Ki-67 dual staining method is a new approach for cervical cancer screening with high sensitivity and specificity. However, there are issues of mis-detection and inaccurate recognition when the YOLOv5s algorithm is directly applied to dual-stained cell images. This paper Proposes a novel cervical cancer dual-stained image recognition (DSIR-YOLO) model based on an YOLOv5. By fusing the Swin-Transformer module, GAM attention mechanism, multi-scale feature fusion, and EIoU loss function, the detection performance is significantly improved, with mAP@0.5 and mAP@0.5:0.95 reaching 92.6% and 70.5%, respectively. Compared with YOLOv5s in five-fold cross-validation, the accuracy, recall, mAP@0.5, and mAP@0.5:0.95 of the improved algorithm are increased by 2.3%, 4.1%, 4.3%, and 8.0%, respectively, with smaller variances and higher stability. Compared with other detection algorithms, DSIR-YOLO in this paper sacrifices some performance requirements to improve the network recognition effect. In addition, the influence of dataset quality on the detection results is studied. By controlling the sealing property of pixels, scale difference, unlabelled cells, and diagonal annotation, the model detection accuracy, recall, mAP@0.5, and mAP@0.5:0.95 are improved by 13.3%, 15.3%, 18.3%, and 30.5%, respectively.

Yuanxing Liu, Weinan Zhang, Baohua Dong et al.

Conversational recommender systems (CRSs) aim to understand the information needs and preferences expressed in a dialogue to recommend suitable items to the user. Most of the existing conversational recommendation datasets are synthesized or simulated with crowdsourcing, which has a large gap with real-world scenarios. To bridge the gap, previous work contributes a dataset E-ConvRec, based on pre-sales dialogues between users and customer service staff in E-commerce scenarios. However, E-ConvRec only supplies coarse-grained annotations and general tasks for making recommendations in pre-sales dialogues. Different from that, we use real user needs as a clue to explore the E-commerce conversational recommendation in complex pre-sales dialogues, namely user needs-centric E-commerce conversational recommendation (UNECR). In this paper, we construct a user needs-centric E-commerce conversational recommendation dataset (U-NEED) from real-world E-commerce scenarios. U-NEED consists of 3 types of resources: (i) 7,698 fine-grained annotated pre-sales dialogues in 5 top categories (ii) 333,879 user behaviors and (iii) 332,148 product knowledge tuples. To facilitate the research of UNECR, we propose 5 critical tasks: (i) pre-sales dialogue understanding (ii) user needs elicitation (iii) user needs-based recommendation (iv) pre-sales dialogue generation and (v) pre-sales dialogue evaluation. We establish baseline methods and evaluation metrics for each task. We report experimental results of 5 tasks on U-NEED. We also report results in 3 typical categories. Experimental results indicate that the challenges of UNECR in various categories are different.

Huansheng Ning, Hang Wang, Wenxi Wang et al.

E-learning is a widely used learning method, but with the development of society, traditional E-learning method has exposed some shortcomings, such as the boring way of teaching, so that it is difficult to increase the enthusiasm of students and raise their attention in class. The application of serious games in E-learning can make up for these shortcomings and effectively improve the quality of teaching. When applying serious games to E-learning, there are two main considerations: educational goals and game design. A successful serious game should organically combine the two aspects and balance the educational and entertaining nature of serious games. This paper mainly discusses the role of serious games in E-learning, various elements of game design, the classification of the educational goals of serious games and the relationship between educational goals and game design. In addition, we try to classify serious games and match educational goals with game types to provide guidance and assistance in the design of serious games. This paper also summarizes some shortcomings that serious games may have in the application of E-learning.

Minghao Xu, Hang Wang, Bingbing Ni et al.

This paper studies unsupervised/self-supervised whole-graph representation learning, which is critical in many tasks such as molecule properties prediction in drug and material discovery. Existing methods mainly focus on preserving the local similarity structure between different graph instances but fail to discover the global semantic structure of the entire data set. In this paper, we propose a unified framework called Local-instance and Global-semantic Learning (GraphLoG) for self-supervised whole-graph representation learning. Specifically, besides preserving the local similarities, GraphLoG introduces the hierarchical prototypes to capture the global semantic clusters. An efficient online expectation-maximization (EM) algorithm is further developed for learning the model. We evaluate GraphLoG by pre-training it on massive unlabeled graphs followed by fine-tuning on downstream tasks. Extensive experiments on both chemical and biological benchmark data sets demonstrate the effectiveness of the proposed approach.

Xuanhong Chen, Hang Wang, Bingbing Ni

Convolution and self-attention are acting as two fundamental building blocks in deep neural networks, where the former extracts local image features in a linear way while the latter non-locally encodes high-order contextual relationships. Though essentially complementary to each other, i.e., first-/high-order, stat-of-the-art architectures, i.e., CNNs or transformers lack a principled way to simultaneously apply both operations in a single computational module, due to their heterogeneous computing pattern and excessive burden of global dot-product for visual tasks. In this work, we theoretically derive a global self-attention approximation scheme, which approximates a self-attention via the convolution operation on transformed features. Based on the approximated scheme, we establish a multi-branch elementary module composed of both convolution and self-attention operation, capable of unifying both local and non-local feature interaction. Importantly, once trained, this multi-branch module could be conditionally converted into a single standard convolution operation via structural re-parameterization, rendering a pure convolution styled operator named X-volution, ready to be plugged into any modern networks as an atomic operation. Extensive experiments demonstrate that the proposed X-volution, achieves highly competitive visual understanding improvements (+1.2% top-1 accuracy on ImageNet classification, +1.7 box AP and +1.5 mask AP on COCO detection and segmentation).

Hang Wang, David J. Miller, George Kesidis

Deep Neural Networks (DNNs) have been shown vulnerable to Test-Time Evasion attacks (TTEs, or adversarial examples), which, by making small changes to the input, alter the DNN's decision. We propose an unsupervised attack detector on DNN classifiers based on class-conditional Generative Adversarial Networks (GANs). We model the distribution of clean data conditioned on the predicted class label by an Auxiliary Classifier GAN (AC-GAN). Given a test sample and its predicted class, three detection statistics are calculated based on the AC-GAN Generator and Discriminator. Experiments on image classification datasets under various TTE attacks show that our method outperforms previous detection methods. We also investigate the effectiveness of anomaly detection using different DNN layers (input features or internal-layer features) and demonstrate, as one might expect, that anomalies are harder to detect using features closer to the DNN's output layer.

Minghao Xu, Hang Wang, Bingbing Ni

Multi-Source Domain Adaptation (MSDA) focuses on transferring the knowledge from multiple source domains to the target domain, which is a more practical and challenging problem compared to the conventional single-source domain adaptation. In this problem, it is essential to model multiple source domains and target domain jointly, and an effective domain combination scheme is also highly required. The graphical structure among different domains is useful to tackle these challenges, in which the interdependency among various instances/categories can be effectively modeled. In this work, we propose two types of graphical models, i.e. Conditional Random Field for MSDA (CRF-MSDA) and Markov Random Field for MSDA (MRF-MSDA), for cross-domain joint modeling and learnable domain combination. In a nutshell, given an observation set composed of a query sample and the semantic prototypes (i.e. representative category embeddings) on various domains, the CRF-MSDA model seeks to learn the joint distribution of labels conditioned on the observations. We attain this goal by constructing a relational graph over all observations and conducting local message passing on it. By comparison, MRF-MSDA aims to model the joint distribution of observations over different Markov networks via an energy-based formulation, and it can naturally perform label prediction by summing the joint likelihoods over several specific networks. Compared to the CRF-MSDA counterpart, the MRF-MSDA model is more expressive and possesses lower computational cost. We evaluate these two models on four standard benchmark data sets of MSDA with distinct domain shift and data complexity, and both models achieve superior performance over existing methods on all benchmarks. In addition, the analytical studies illustrate the effect of different model components and provide insights about how the cross-domain joint modeling performs.

Hang Wang, Sen Lin, Hamid Jafarkhani et al.

This paper studies distributed Q-learning for Linear Quadratic Regulator (LQR) in a multi-agent network. The existing results often assume that agents can observe the global system state, which may be infeasible in large-scale systems due to privacy concerns or communication constraints. In this work, we consider a setting with unknown system models and no centralized coordinator. We devise a state tracking (ST) based Q-learning algorithm to design optimal controllers for agents. Specifically, we assume that agents maintain local estimates of the global state based on their local information and communications with neighbors. At each step, every agent updates its local global state estimation, based on which it solves an approximate Q-factor locally through policy iteration. Assuming decaying injected excitation noise during the policy evaluation, we prove that the local estimation converges to the true global state, and establish the convergence of the proposed distributed ST-based Q-learning algorithm. The experimental studies corroborate our theoretical results by showing that our proposed method achieves comparable performance with the centralized case.

Sen Lin, Hang Wang, Junshan Zhang

System identification is a fundamental problem in reinforcement learning, control theory and signal processing, and the non-asymptotic analysis of the corresponding sample complexity is challenging and elusive, even for linear time-varying (LTV) systems. To tackle this challenge, we develop an episodic block model for the LTV system where the model parameters remain constant within each block but change from block to block. Based on the observation that the model parameters across different blocks are related, we treat each episodic block as a learning task and then run meta-learning over many blocks for system identification, using two steps, namely offline meta-learning and online adaptation. We carry out a comprehensive non-asymptotic analysis of the performance of meta-learning based system identification. To deal with the technical challenges rooted in the sample correlation and small sample sizes in each block, we devise a new two-scale martingale small-ball approach for offline meta-learning, for arbitrary model correlation structure across blocks. We then quantify the finite time error of online adaptation by leveraging recent advances in linear stochastic approximation with correlated samples.

Zhen Xiang, David J. Miller, Hang Wang et al.

Recently, a backdoor data poisoning attack was proposed, which adds mislabeled examples to the training set, with an embedded backdoor pattern, aiming to have the classifier learn to classify to a target class whenever the backdoor pattern is present in a test sample. Here, we address post-training detection of innocuous perceptible backdoors in DNN image classifiers, wherein the defender does not have access to the poisoned training set, but only to the trained classifier, as well as unpoisoned examples. This problem is challenging because without the poisoned training set, we have no hint about the actual backdoor pattern used during training. This post-training scenario is also of great import because in many practical contexts the DNN user did not train the DNN and does not have access to the training data. We identify two important properties of perceptible backdoor patterns - spatial invariance and robustness - based upon which we propose a novel detector using the maximum achievable misclassification fraction (MAMF) statistic. We detect whether the trained DNN has been backdoor-attacked and infer the source and target classes. Our detector outperforms other existing detectors and, coupled with an imperceptible backdoor detector, helps achieve post-training detection of all evasive backdoors.

Guijin Wang, Cairong Zhang, Xinghao Chen et al.

In human-computer interaction, it is important to accurately estimate the hand pose especially fingertips. However, traditional approaches for fingertip localization mainly rely on depth images and thus suffer considerably from the noise and missing values. Instead of depth images, stereo images can also provide 3D information of hands and promote 3D hand pose estimation. There are nevertheless limitations on the dataset size, global viewpoints, hand articulations and hand shapes in the publicly available stereo-based hand pose datasets. To mitigate these limitations and promote further research on hand pose estimation from stereo images, we propose a new large-scale binocular hand pose dataset called THU-Bi-Hand, offering a new perspective for fingertip localization. In the THU-Bi-Hand dataset, there are 447k pairs of stereo images of different hand shapes from 10 subjects with accurate 3D location annotations of the wrist and five fingertips. Captured with minimal restriction on the range of hand motion, the dataset covers large global viewpoint space and hand articulation space. To better present the performance of fingertip localization on THU-Bi-Hand, we propose a novel scheme termed Bi-stream Pose Guided Region Ensemble Network (Bi-Pose-REN). It extracts more representative feature regions around joint points in the feature maps under the guidance of the previously estimated pose. The feature regions are integrated hierarchically according to the topology of hand joints to regress the refined hand pose. Bi-Pose-REN and several existing methods are evaluated on THU-Bi-Hand so that benchmarks are provided for further research. Experimental results show that our new method has achieved the best performance on THU-Bi-Hand.