Zhi Gao

Junjie Wen, Jinqiang Cui, Zhenjun Zhao et al.

Underwater image enhancement (UIE) is vital for high-level vision-related underwater tasks. Although learning-based UIE methods have made remarkable achievements in recent years, it's still challenging for them to consistently deal with various underwater conditions, which could be caused by: 1) the use of the simplified atmospheric image formation model in UIE may result in severe errors; 2) the network trained solely with synthetic images might have difficulty in generalizing well to real underwater images. In this work, we, for the first time, propose a framework \textit{SyreaNet} for UIE that integrates both synthetic and real data under the guidance of the revised underwater image formation model and novel domain adaptation (DA) strategies. First, an underwater image synthesis module based on the revised model is proposed. Then, a physically guided disentangled network is designed to predict the clear images by combining both synthetic and real underwater images. The intra- and inter-domain gaps are abridged by fully exchanging the domain knowledge. Extensive experiments demonstrate the superiority of our framework over other state-of-the-art (SOTA) learning-based UIE methods qualitatively and quantitatively. The code and dataset are publicly available at https://github.com/RockWenJJ/SyreaNet.git.

Jin Chen, Zhi Gao, Xinxiao Wu et al.

Single domain generalization aims to learn a model from a single training domain (source domain) and apply it to multiple unseen test domains (target domains). Existing methods focus on expanding the distribution of the training domain to cover the target domains, but without estimating the domain shift between the source and target domains. In this paper, we propose a new learning paradigm, namely simulate-analyze-reduce, which first simulates the domain shift by building an auxiliary domain as the target domain, then learns to analyze the causes of domain shift, and finally learns to reduce the domain shift for model adaptation. Under this paradigm, we propose a meta-causal learning method to learn meta-knowledge, that is, how to infer the causes of domain shift between the auxiliary and source domains during training. We use the meta-knowledge to analyze the shift between the target and source domains during testing. Specifically, we perform multiple transformations on source data to generate the auxiliary domain, perform counterfactual inference to learn to discover the causal factors of the shift between the auxiliary and source domains, and incorporate the inferred causality into factor-aware domain alignments. Extensive experiments on several benchmarks of image classification show the effectiveness of our method.

Cong Yang, Zuchao Li, Hongzan Jiao et al.

Recently, while significant progress has been made in remote sensing image change captioning, existing methods fail to filter out areas unrelated to actual changes, making models susceptible to irrelevant features. In this article, we propose a novel multimodal framework for remote sensing image change captioning, guided by Key Change Features and Instruction-tuned (KCFI). This framework aims to fully leverage the intrinsic knowledge of large language models through visual instructions and enhance the effectiveness and accuracy of change features using pixel-level change detection tasks. Specifically, KCFI includes a ViTs encoder for extracting bi-temporal remote sensing image features, a key feature perceiver for identifying critical change areas, a pixel-level change detection decoder to constrain key change features, and an instruction-tuned decoder based on a large language model. Moreover, to ensure that change description and change detection tasks are jointly optimized, we employ a dynamic weight-averaging strategy to balance the losses between the two tasks. We also explore various feature combinations for visual fine-tuning instructions and demonstrate that using only key change features to guide the large language model is the optimal choice. To validate the effectiveness of our approach, we compare it against several state-of-the-art change captioning methods on the LEVIR-CC dataset, achieving the best performance. Our code will be available at https://github.com/yangcong356/KCFI.git.

Zhi Gao, Chen Xu, Feng Li et al.

Continual learning aims to efficiently learn from a non-stationary stream of data while avoiding forgetting the knowledge of old data. In many practical applications, data complies with non-Euclidean geometry. As such, the commonly used Euclidean space cannot gracefully capture non-Euclidean geometric structures of data, leading to inferior results. In this paper, we study continual learning from a novel perspective by exploring data geometry for the non-stationary stream of data. Our method dynamically expands the geometry of the underlying space to match growing geometric structures induced by new data, and prevents forgetting by keeping geometric structures of old data into account. In doing so, making use of the mixed curvature space, we propose an incremental search scheme, through which the growing geometric structures are encoded. Then, we introduce an angular-regularization loss and a neighbor-robustness loss to train the model, capable of penalizing the change of global geometric structures and local geometric structures. Experiments show that our method achieves better performance than baseline methods designed in Euclidean space.

Pengxiang Li, Zhi Gao, Bofei Zhang et al.

Vision language models (VLMs) have achieved impressive progress in diverse applications, becoming a prevalent research direction. In this paper, we build FIRE, a feedback-refinement dataset, consisting of 1.1M multi-turn conversations that are derived from 27 source datasets, empowering VLMs to spontaneously refine their responses based on user feedback across diverse tasks. To scale up the data collection, FIRE is collected in two components: FIRE-100K and FIRE-1M, where FIRE-100K is generated by GPT-4V, and FIRE-1M is freely generated via models trained on FIRE-100K. Then, we build FIRE-Bench, a benchmark to comprehensively evaluate the feedback-refining capability of VLMs, which contains 11K feedback-refinement conversations as the test data, two evaluation settings, and a model to provide feedback for VLMs. We develop the FIRE-LLaVA model by fine-tuning LLaVA on FIRE-100K and FIRE-1M, which shows remarkable feedback-refining capability on FIRE-Bench and outperforms untrained VLMs by 50%, making more efficient user-agent interactions and underscoring the significance of the FIRE dataset.

Duanchu Wang, Cheng Li, Junjie Yang et al.

Point cloud quality plays a critical role in 3D acquisition, reconstruction, rendering, and perception, yet existing point cloud quality assessment (PCQA) research remains largely centered on scalar score prediction. In practical inspection scenarios, quality assessment often involves identifying defects, characterizing dominant issue types, assessing downstream usability, and providing evidence-supported descriptions, which are not explicitly evaluated by current benchmarks. We introduce PointQ-Bench, a benchmark designed to extend PCQA from scalar scoring toward comprehensive quality understanding. PointQ-Bench consists of 3,083 point clouds spanning authentic scans, simulated distortions, and AI-generated content, covering eight major issue types. Each sample is annotated with mean opinion scores (MOS), quality levels, issue tags, expert-grounded descriptions, and 12,332 question-answer pairs. The benchmark supports three perception-oriented tasks: anomaly sensing, defect diagnosis, and usability grading, as well as a cognition-oriented task of open-ended quality reporting. To evaluate free-form quality descriptions, we further propose SSFRQ-5D, a five-dimensional evaluation protocol validated through human-AI agreement analysis. Extensive experiments on 14 vision-language models and traditional PCQA baselines reveal a consistent perception-diagnosis gap: while current models exhibit emerging abilities in coarse defect perception, they struggle with grounded diagnosis and quality calibration. Strong 2D MLLMs generally outperform existing 3D VLMs, and the benefit of additional views or point-level inputs is non-uniform, varying across tasks, data sources, and models, particularly under boundary-ambiguous conditions. Overall, PointQ-Bench provides a diagnostic testbed for advancing reliable and interpretable point cloud quality understanding.

Yue Fan, Xiaojian Ma, Rujie Wu et al.

We explore how reconciling several foundation models (large language models and vision-language models) with a novel unified memory mechanism could tackle the challenging video understanding problem, especially capturing the long-term temporal relations in lengthy videos. In particular, the proposed multimodal agent VideoAgent: 1) constructs a structured memory to store both the generic temporal event descriptions and object-centric tracking states of the video; 2) given an input task query, it employs tools including video segment localization and object memory querying along with other visual foundation models to interactively solve the task, utilizing the zero-shot tool-use ability of LLMs. VideoAgent demonstrates impressive performances on several long-horizon video understanding benchmarks, an average increase of 6.6% on NExT-QA and 26.0% on EgoSchema over baselines, closing the gap between open-sourced models and private counterparts including Gemini 1.5 Pro.

Yixue Zhang, Kun Wu, Zhi Gao et al.

The pursuit of general-purpose robotic manipulation is hindered by the scarcity of diverse, real-world interaction data. Unlike data collection from web in vision or language, robotic data collection is an active process incurring prohibitive physical costs. Consequently, automated task curation to maximize data value remains a critical yet under-explored challenge. Existing manual methods are unscalable and biased toward common tasks, while off-the-shelf foundation models often hallucinate physically infeasible instructions. To address this, we introduce RoboGene, an agentic framework designed to automate the generation of diverse, physically plausible manipulation tasks across single-arm, dual-arm, and mobile robots. RoboGene integrates three core components: diversity-driven sampling for broad task coverage, self-reflection mechanisms to enforce physical constraints, and human-in-the-loop refinement for continuous improvement. We conduct extensive quantitative analysis and large-scale real-world experiments, collecting datasets of 18k trajectories and introducing novel metrics to assess task quality, feasibility, and diversity. Results demonstrate that RoboGene significantly outperforms state-of-the-art foundation models (e.g., GPT-4o, Gemini 2.5 Pro). Furthermore, real-world experiments show that VLA models pre-trained with RoboGene achieve higher success rates and superior generalization, underscoring the importance of high-quality task generation. Our project is available at https://robogene-boost-vla.github.io.

Wu Wei, Xiaomeng Fan, Yuwei Wu et al.

Modality alignment is critical for vision-language models (VLMs) to effectively integrate information across modalities. However, existing methods extract hierarchical features from text while representing each image with a single feature, leading to asymmetric and suboptimal alignment. To address this, we propose Alignment across Trees, a method that constructs and aligns tree-like hierarchical features for both image and text modalities. Specifically, we introduce a semantic-aware visual feature extraction framework that applies a cross-attention mechanism to visual class tokens from intermediate Transformer layers, guided by textual cues to extract visual features with coarse-to-fine semantics. We then embed the feature trees of the two modalities into hyperbolic manifolds with distinct curvatures to effectively model their hierarchical structures. To align across the heterogeneous hyperbolic manifolds with different curvatures, we formulate a KL distance measure between distributions on heterogeneous manifolds, and learn an intermediate manifold for manifold alignment by minimizing the distance. We prove the existence and uniqueness of the optimal intermediate manifold. Experiments on taxonomic open-set classification tasks across multiple image datasets demonstrate that our method consistently outperforms strong baselines under few-shot and cross-domain settings.

Bofei Zhang, Zirui Shang, Zhi Gao et al.

Building Graphical User Interface (GUI) agents is a promising research direction, which simulates human interaction with computers or mobile phones to perform diverse GUI tasks. However, a major challenge in developing generalized GUI agents is the lack of sufficient trajectory data across various operating systems and applications, mainly due to the high cost of manual annotations. In this paper, we propose the TongUI framework that builds generalized GUI agents by learning from rich multimodal web tutorials. Concretely, we crawl and process online GUI tutorials (such as videos and articles) into GUI agent trajectory data, through which we produce the GUI-Net dataset containing 143K trajectory data across five operating systems and more than 200 applications. We develop the TongUI agent by fine-tuning Qwen2.5-VL-3B/7B models on GUI-Net, which show remarkable performance improvements on commonly used grounding and navigation benchmarks, outperforming baseline agents about 10\% on multiple benchmarks, showing the effectiveness of the GUI-Net dataset and underscoring the significance of our TongUI framework. We will fully open-source the code, the GUI-Net dataset, and the trained models soon.

Hanqing Guo, Ye Zheng, Yin Zhang et al.

Visual detection of micro aerial vehicles (MAVs) has received increasing research attention in recent years due to its importance in many applications. However, the existing approaches based on either appearance or motion features of MAVs still face challenges when the background is complex, the MAV target is small, or the computation resource is limited. In this paper, we propose a global-local MAV detector that can fuse both motion and appearance features for MAV detection under challenging conditions. This detector first searches MAV target using a global detector and then switches to a local detector which works in an adaptive search region to enhance accuracy and efficiency. Additionally, a detector switcher is applied to coordinate the global and local detectors. A new dataset is created to train and verify the effectiveness of the proposed detector. This dataset contains more challenging scenarios that can occur in practice. Extensive experiments on three challenging datasets show that the proposed detector outperforms the state-of-the-art ones in terms of detection accuracy and computational efficiency. In particular, this detector can run with near real-time frame rate on NVIDIA Jetson NX Xavier, which demonstrates the usefulness of our approach for real-world applications. The dataset is available at https://github.com/WestlakeIntelligentRobotics/GLAD. In addition, A video summarizing this work is available at https://youtu.be/Tv473mAzHbU.

Chenrui Shi, Zedong Yu, Zhi Gao et al.

Large vision language models (VLMs) have advanced graphical user interface (GUI) task automation but still lag behind humans. We hypothesize this gap stems from missing core GUI knowledge, which existing training schemes (such as supervised fine tuning and reinforcement learning) alone cannot fully address. By analyzing common failure patterns in GUI task execution, we distill GUI knowledge into three dimensions: (1) interface perception, knowledge about recognizing widgets and system states; (2) interaction prediction, knowledge about reasoning action state transitions; and (3) instruction understanding, knowledge about planning, verifying, and assessing task completion progress. We further introduce GUI Knowledge Bench, a benchmark with multiple choice and yes/no questions across six platforms (Web, Android, MacOS, Windows, Linux, IOS) and 292 applications. Our evaluation shows that current VLMs identify widget functions but struggle with perceiving system states, predicting actions, and verifying task completion. Experiments on real world GUI tasks further validate the close link between GUI knowledge and task success. By providing a structured framework for assessing GUI knowledge, our work supports the selection of VLMs with greater potential prior to downstream training and provides insights for building more capable GUI agents.

Enqi Liu, Liyuan Pan, Zhi Gao et al.

Foggy conditions are commonly encountered in real-world applications; however, existing action recognition approaches typically assume favorable weather and high-quality video inputs. On foggy days, unpredictable visibility degradation and reduced contrast obstruct the extraction of semantic cues, posing significant challenges for current action recognition methods. In this paper, we mitigate the issues faced in action recognition under foggy conditions by employing two strategies. First, we present FogAct, the first benchmark dataset for foggy action recognition, consisting of paired clean and foggy videos captured with a stereo camera system. The dataset spans 10 scenes and 55 action categories, comprising nearly 10,000 video clips. Second, we propose FogNet, a two-stream CLIP model that discovers fog-invariant semantic information hidden behind the degraded videos. FogNet learns robust representations of foggy videos with guidance from clean videos, effectively capturing shared structural and motion cues between clean and foggy videos. Extensive experiments on FogAct and three other popular datasets demonstrate that our method achieves competitive performance compared with state-of-the-art (SOTA) approaches. Our FogAct and FogNet are given in our project page.

Xiaowen Zhang, Zhi Gao, Licheng Jiao et al.

In vision-language models (VLMs), misalignment between textual descriptions and visual coordinates often induces hallucinations. This issue becomes particularly severe in dense prediction tasks such as spatial-temporal video grounding (STVG). Prior approaches typically focus on enhancing visual-textual alignment or attaching auxiliary decoders. However, these strategies inevitably introduce additional trainable modules, leading to significant annotation costs and computational overhead. In this work, we propose a novel visual prompting paradigm that avoids the difficult problem of aligning coordinates across modalities. Specifically, we reformulate per-frame coordinate prediction as a compact instance-level identification problem by assigning each object a unique, temporally consistent ID. These IDs are embedded into the video as visual prompts, providing explicit and interpretable inputs to the VLMs. Furthermore, we introduce STVG-R1, the first reinforcement learning framework for STVG, which employs a task-driven reward to jointly optimize temporal accuracy, spatial consistency, and structural format regularization. Extensive experiments on six benchmarks demonstrate the effectiveness of our approach. STVG-R1 surpasses the baseline Qwen2.5-VL-7B by a remarkable margin of 20.9% on m_IoU on the HCSTVG-v2 benchmark, establishing a new state of the art (SOTA). Surprisingly, STVG-R1 also exhibits strong zero-shot generalization to multi-object referring video object segmentation tasks, achieving a SOTA 47.3% J&F on MeViS.

Rui Xie, Zhi Gao, Chenrui Shi et al.

Large vision-language models have endowed GUI agents with strong general capabilities for interface understanding and interaction. However, due to insufficient exposure to domain-specific software operation data during training, these agents exhibit significant domain bias - they lack familiarity with the specific operation workflows (planning) and UI element layouts (grounding) of particular applications, limiting their real-world task performance. In this paper, we present GUIDE (GUI Unbiasing via Instructional-Video Driven Expertise), a training-free, plug-and-play framework that resolves GUI agent domain bias by autonomously acquiring domain-specific expertise from web tutorial videos through a retrieval-augmented automated annotation pipeline. GUIDE introduces two key innovations. First, a subtitle-driven Video-RAG pipeline unlocks video semantics through subtitle analysis, performing progressive three-stage retrieval - domain classification, topic extraction, and relevance matching - to identify task-relevant tutorial videos. Second, a fully automated annotation pipeline built on an inverse dynamics paradigm feeds consecutive keyframes enhanced with UI element detection into VLMs, inferring the required planning and grounding knowledge that are injected into the agent's corresponding modules to address both manifestations of domain bias. Extensive experiments on OSWorld demonstrate GUIDE's generality as a plug-and-play component for both multi-agent systems and single-model agents. It consistently yields over 5% improvements and reduces execution steps - without modifying any model parameters or architecture - validating GUIDE as an architecture-agnostic enhancement to bridge GUI agent domain bias.

Pengxiang Li, Zechen Hu, Zirui Shang et al.

Vision-language model (VLM) based GUI agents show promise for automating complex desktop and mobile tasks, but face significant challenges in applying reinforcement learning (RL): (1) slow multi-turn interactions with GUI environments for policy rollout, and (2) insufficient high-quality agent-environment interactions for policy learning. To address these challenges, we propose DART, a Decoupled Agentic RL Training framework for GUI agents, which coordinates heterogeneous modules in a highly decoupled manner. DART separates the training system into four asynchronous modules: environment cluster, rollout service, data manager, and trainer. This design enables non-blocking communication, asynchronous training, rollout-wise trajectory sampling, and per-worker model synchronization, significantly improving the system efficiency: 1.6*GPU utilization for rollout, 1.9* training throughput, and 5.5* environment utilization. To facilitate effective learning from abundant samples, we introduce an adaptive data curation scheme: (1) pre-collecting successful trajectories for challenging tasks to supplement sparse success in online sampling; (2) dynamically adjusting rollout numbers and trajectory lengths based on task difficulty; (3) training selectively on high-entropy steps to prioritize critical decisions; (4) stabilizing learning via truncated importance sampling for policy mismatch between policy rollout and updating. On the OSWorld benchmark, DART-GUI-7B achieves a 42.13% task success rate, a 14.61% absolute gain over the base model, and 7.34% higher than open-source SOTA. We will fully open-source our training framework, data, and model checkpoints via computer-use-agents.github.io/dart-gui, which we believe is a timely contribution to the open-source community of agentic RL training.

Ziyi Wang, Zhi Gao, Jin Chen et al.

Domain generalization (DG) aims to learn a model from source domains and apply it to unseen target domains with out-of-distribution data. Owing to CLIP's strong ability to encode semantic concepts, it has attracted increasing interest in domain generalization. However, CLIP often struggles to focus on task-relevant regions across domains, i.e., domain-invariant regions, resulting in suboptimal performance on unseen target domains. To address this challenge, we propose an attention-refocusing scheme, called Simulate, Refocus and Ensemble (SRE), which learns to reduce the domain shift by aligning the attention maps in CLIP via attention refocusing. SRE first simulates domain shifts by performing augmentation on the source data to generate simulated target domains. SRE then learns to reduce the domain shifts by refocusing the attention in CLIP between the source and simulated target domains. Finally, SRE utilizes ensemble learning to enhance the ability to capture domain-invariant attention maps between the source data and the simulated target data. Extensive experimental results on several datasets demonstrate that SRE generally achieves better results than state-of-the-art methods. The code is available at: https://github.com/bitPrincy/SRE-DG.

Weiwei Song, Zhi Gao, Renwei Dian et al.

Remote sensing image retrieval (RSIR), aiming at searching for a set of similar items to a given query image, is a very important task in remote sensing applications. Deep hashing learning as the current mainstream method has achieved satisfactory retrieval performance. On one hand, various deep neural networks are used to extract semantic features of remote sensing images. On the other hand, the hashing techniques are subsequently adopted to map the high-dimensional deep features to the low-dimensional binary codes. This kind of methods attempts to learn one hash function for both the query and database samples in a symmetric way. However, with the number of database samples increasing, it is typically time-consuming to generate the hash codes of large-scale database images. In this paper, we propose a novel deep hashing method, named asymmetric hash code learning (AHCL), for RSIR. The proposed AHCL generates the hash codes of query and database images in an asymmetric way. In more detail, the hash codes of query images are obtained by binarizing the output of the network, while the hash codes of database images are directly learned by solving the designed objective function. In addition, we combine the semantic information of each image and the similarity information of pairs of images as supervised information to train a deep hashing network, which improves the representation ability of deep features and hash codes. The experimental results on three public datasets demonstrate that the proposed method outperforms symmetric methods in terms of retrieval accuracy and efficiency. The source code is available at https://github.com/weiweisong415/Demo AHCL for TGRS2022.

Zhi Gao, Yuntao Du, Xintong Zhang et al.

Utilizing large language models (LLMs) to compose off-the-shelf visual tools represents a promising avenue of research for developing robust visual assistants capable of addressing diverse visual tasks. However, these methods often overlook the potential for continual learning, typically by freezing the utilized tools, thus limiting their adaptation to environments requiring new knowledge. To tackle this challenge, we propose CLOVA, a Closed-Loop Visual Assistant, which operates within a framework encompassing inference, reflection, and learning phases. During the inference phase, LLMs generate programs and execute corresponding tools to complete assigned tasks. In the reflection phase, a multimodal global-local reflection scheme analyzes human feedback to determine which tools require updating. Lastly, the learning phase employs three flexible approaches to automatically gather training data and introduces a novel prompt tuning scheme to update the tools, allowing CLOVA to efficiently acquire new knowledge. Experimental findings demonstrate that CLOVA surpasses existing tool-usage methods by 5% in visual question answering and multiple-image reasoning, by 10% in knowledge tagging, and by 20% in image editing. These results underscore the significance of the continual learning capability in general visual assistants.

Xintong Zhang, Zhi Gao, Bofei Zhang et al.

Vision language models (VLMs) have achieved impressive performance across a variety of computer vision tasks. However, the multimodal reasoning capability has not been fully explored in existing models. In this paper, we propose a Chain-of-Focus (CoF) method that allows VLMs to perform adaptive focusing and zooming in on key image regions based on obtained visual cues and the given questions, achieving efficient multimodal reasoning. To enable this CoF capability, we present a two-stage training pipeline, including supervised fine-tuning (SFT) and reinforcement learning (RL). In the SFT stage, we construct the MM-CoF dataset, comprising 3K samples derived from a visual agent designed to adaptively identify key regions to solve visual tasks with different image resolutions and questions. We use MM-CoF to fine-tune the Qwen2.5-VL model for cold start. In the RL stage, we leverage the outcome accuracies and formats as rewards to update the Qwen2.5-VL model, enabling further refining the search and reasoning strategy of models without human priors. Our model achieves significant improvements on multiple benchmarks. On the V* benchmark that requires strong visual reasoning capability, our model outperforms existing VLMs by 5% among 8 image resolutions ranging from 224 to 4K, demonstrating the effectiveness of the proposed CoF method and facilitating the more efficient deployment of VLMs in practical applications.

Zhi Gao, Bofei Zhang, Pengxiang Li et al.

The advancement of large language models (LLMs) prompts the development of multi-modal agents, which are used as a controller to call external tools, providing a feasible way to solve practical tasks. In this paper, we propose a multi-modal agent tuning method that automatically generates multi-modal tool-usage data and tunes a vision-language model (VLM) as the controller for powerful tool-usage reasoning. To preserve the data quality, we prompt the GPT-4o mini model to generate queries, files, and trajectories, followed by query-file and trajectory verifiers. Based on the data synthesis pipeline, we collect the MM-Traj dataset that contains 20K tasks with trajectories of tool usage. Then, we develop the T3-Agent via \underline{T}rajectory \underline{T}uning on VLMs for \underline{T}ool usage using MM-Traj. Evaluations on the GTA and GAIA benchmarks show that the T3-Agent consistently achieves improvements on two popular VLMs: MiniCPM-V-8.5B and {Qwen2-VL-7B}, which outperforms untrained VLMs by $20\%$, showing the effectiveness of the proposed data synthesis pipeline, leading to high-quality data for tool-usage capabilities.

Enqi Liu, Liyuan Pan, Zhi Gao et al.

Recent advancements in Graphical User Interface (GUI) agents have predominantly focused on training paradigms like supervised fine-tuning (SFT) and reinforcement learning (RL). However, the challenge of high-dynamic GUI environments remains largely underexplored. Existing agents typically rely on a single screenshot after each action for decision-making, leading to a partially observable (or even unobservable) Markov decision process, where the key GUI state including important information for actions is often inadequately captured. To systematically explore this challenge, we introduce DynamicGUIBench, a comprehensive online GUI benchmark spanning ten applications and diverse interaction scenarios characterized by important interface changes between actions. Furthermore, we present DynamicUI, an agent designed for dynamic interfaces, which takes screen-recording videos of the interaction process as input and consists of three components: a dynamic perceiver, a refinement strategy, and a reflection. Specifically, the dynamic perceiver clusters frames of the GUI video, generates captions for the centroids, and iteratively selects the most informative frames as the salient dynamic context. Considering that there may be inconsistencies and noise between the selected frames and the textual context of the agent, the refinement strategy employs an action-conditioned filtering to refine thoughts to mitigate thought-action inconsistency and redundancy. Based on the refined agent trajectories, the reflection module provides effective and accurate guidance for further actions. Experiments on DynamicGUIBench demonstrate that DynamicUI significantly improves the performance in dynamic GUI environments, while maintaining competitive performance on other public benchmarks.

Yuntao Du, Kailin Jiang, Zhi Gao et al.

Knowledge editing techniques have emerged as essential tools for updating the factual knowledge of large language models (LLMs) and multimodal models (LMMs), allowing them to correct outdated or inaccurate information without retraining from scratch. However, existing benchmarks for multimodal knowledge editing primarily focus on entity-level knowledge represented as simple triplets, which fail to capture the complexity of real-world multimodal information. To address this issue, we introduce MMKE-Bench, a comprehensive MultiModal Knowledge Editing Benchmark, designed to evaluate the ability of LMMs to edit diverse visual knowledge in real-world scenarios. MMKE-Bench addresses these limitations by incorporating three types of editing tasks: visual entity editing, visual semantic editing, and user-specific editing. Besides, MMKE-Bench uses free-form natural language to represent and edit knowledge, offering a more flexible and effective format. The benchmark consists of 2,940 pieces of knowledge and 8,363 images across 33 broad categories, with evaluation questions automatically generated and human-verified. We assess five state-of-the-art knowledge editing methods on three prominent LMMs, revealing that no method excels across all criteria, and that visual and user-specific edits are particularly challenging. MMKE-Bench sets a new standard for evaluating the robustness of multimodal knowledge editing techniques, driving progress in this rapidly evolving field.

Junjie Wen, Jinqiang Cui, Benyun Zhao et al.

In recent years, significant progress has been made in the field of underwater image enhancement (UIE). However, its practical utility for high-level vision tasks, such as underwater object detection (UOD) in Autonomous Underwater Vehicles (AUVs), remains relatively unexplored. It may be attributed to several factors: (1) Existing methods typically employ UIE as a pre-processing step, which inevitably introduces considerable computational overhead and latency. (2) The process of enhancing images prior to training object detectors may not necessarily yield performance improvements. (3) The complex underwater environments can induce significant domain shifts across different scenarios, seriously deteriorating the UOD performance. To address these challenges, we introduce EnYOLO, an integrated real-time framework designed for simultaneous UIE and UOD with domain-adaptation capability. Specifically, both the UIE and UOD task heads share the same network backbone and utilize a lightweight design. Furthermore, to ensure balanced training for both tasks, we present a multi-stage training strategy aimed at consistently enhancing their performance. Additionally, we propose a novel domain-adaptation strategy to align feature embeddings originating from diverse underwater environments. Comprehensive experiments demonstrate that our framework not only achieves state-of-the-art (SOTA) performance in both UIE and UOD tasks, but also shows superior adaptability when applied to different underwater scenarios. Our efficiency analysis further highlights the substantial potential of our framework for onboard deployment.

Pengxiang Li, Zhi Gao, Bofei Zhang et al.

Multimodal agents, which integrate a controller e.g., a vision language model) with external tools, have demonstrated remarkable capabilities in tackling complex multimodal tasks. Existing approaches for training these agents, both supervised fine-tuning and reinforcement learning, depend on extensive human-annotated task-answer pairs and tool trajectories. However, for complex multimodal tasks, such annotations are prohibitively expensive or impractical to obtain. In this paper, we propose an iterative tool usage exploration method for multimodal agents without any pre-collected data, namely SPORT, via step-wise preference optimization to refine the trajectories of tool usage. Our method enables multimodal agents to autonomously discover effective tool usage strategies through self-exploration and optimization, eliminating the bottleneck of human annotation. SPORT has four iterative components: task synthesis, step sampling, step verification, and preference tuning. We first synthesize multimodal tasks using language models. Then, we introduce a novel trajectory exploration scheme, where step sampling and step verification are executed alternately to solve synthesized tasks. In step sampling, the agent tries different tools and obtains corresponding results. In step verification, we employ a verifier to provide AI feedback to construct step-wise preference data. The data is subsequently used to update the controller for tool usage through preference tuning, producing a SPORT agent. By interacting with real environments, the SPORT agent gradually evolves into a more refined and capable system. Evaluation in the GTA and GAIA benchmarks shows that the SPORT agent achieves 6.41% and 3.64% improvements, underscoring the generalization and effectiveness introduced by our method. The project page is https://SPORT-Agents.github.io.

Yunteng Tan, Zhi Gao, Xinxiao Wu

Large language model-based web agents have shown strong potential in automating web interactions through advanced reasoning and instruction following. While retrieval-based memory derived from historical trajectories enables these agents to handle complex, long-horizon tasks, current methods struggle to generalize across unseen websites. We identify that this challenge arises from the flat memory structures that entangle high-level task logic with site-specific action details. This entanglement induces a workflow mismatch in new environments, where retrieved contents are conflated with current web, leading to logically inconsistent execution. To address this, we propose Hierarchical Memory Tree (HMT), a structured framework designed to explicitly decouple logical planning from action execution. HMT constructs a three-level hierarchy from raw trajectories via an automated abstraction pipeline: the Intent level maps diverse user instructions to standardized task goals; the Stage level defines reusable semantic subgoals characterized by observable pre-conditions and post-conditions; and the Action level stores action patterns paired with transferable semantic element descriptions. Leveraging this structure, we develop a stage-aware inference mechanism comprising a Planner and an Actor. By explicitly validating pre-conditions, the Planner aligns the current state with the correct logical subgoal to prevent workflow mismatch, while the Actor grounds actions by matching the stored semantic descriptions to the target page. Experimental results on Mind2Web and WebArena show that HMT significantly outperforms flat-memory methods, particularly in cross-website and cross-domain scenarios, highlighting the necessity of structured memory for robust generalization of web agents.

Xiaomeng Fan, Yuwei Wu, Zhi Gao et al.

Hyperbolic neural networks (HNNs) have demonstrated notable efficacy in representing real-world data with hierarchical structures via exploiting the geometric properties of hyperbolic spaces characterized by negative curvatures. Curvature plays a crucial role in optimizing HNNs. Inappropriate curvatures may cause HNNs to converge to suboptimal parameters, degrading overall performance. So far, the theoretical foundation of the effect of curvatures on HNNs has not been developed. In this paper, we derive a PAC-Bayesian generalization bound of HNNs, highlighting the role of curvatures in the generalization of HNNs via their effect on the smoothness of the loss landscape. Driven by the derived bound, we propose a sharpness-aware curvature learning method to smooth the loss landscape, thereby improving the generalization of HNNs. In our method, we design a scope sharpness measure for curvatures, which is minimized through a bi-level optimization process. Then, we introduce an implicit differentiation algorithm that efficiently solves the bi-level optimization by approximating gradients of curvatures. We present the approximation error and convergence analyses of the proposed method, showing that the approximation error is upper-bounded, and the proposed method can converge by bounding gradients of HNNs. Experiments on four settings: classification, learning from long-tailed data, learning from noisy data, and few-shot learning show that our method can improve the performance of HNNs.

Mingliang Zhai, Zhi Gao, Yuwei Wu et al.

Embodied Question Answering (EQA) requires agents to autonomously explore and understand the environment to answer context-dependent questions. Existing frameworks typically center around the planner, which guides the stopping module, memory module, and answering module for reasoning. In this paper, we propose a memory-centric EQA framework named MemoryEQA. Unlike planner-centric EQA models where the memory module cannot fully interact with other modules, MemoryEQA flexible feeds memory information into all modules, thereby enhancing efficiency and accuracy in handling complex tasks, such as those involving multiple targets across different regions. Specifically, we establish a multi-modal hierarchical memory mechanism, which is divided into global memory that stores language-enhanced scene maps, and local memory that retains historical observations and state information. When performing EQA tasks, the multi-modal large language model is leveraged to convert memory information into the required input formats for injection into different modules. To evaluate EQA models' memory capabilities, we constructed the MT-HM3D dataset based on HM3D, comprising 1,587 question-answer pairs involving multiple targets across various regions, which requires agents to maintain memory of exploration-acquired target information. Experimental results on HM-EQA, MT-HM3D, and OpenEQA demonstrate the effectiveness of our framework, where a 19.8% performance gain on MT-HM3D compared to baseline model further underscores memory capability's pivotal role in resolving complex tasks.

Kailin Jiang, Yuntao Du, Yukai Ding et al.

Large language/multimodal models (LLMs/LMMs) store extensive pre-trained knowledge but struggle to maintain consistency with real-world updates, making it difficult to avoid catastrophic forgetting while acquiring evolving knowledge. Previous work focused on constructing textual knowledge datasets and exploring knowledge injection in LLMs, lacking exploration of multimodal evolving knowledge injection in LMMs. To address this, we propose the EVOKE benchmark to evaluate LMMs' ability to inject multimodal evolving knowledge in real-world scenarios. Meanwhile, a comprehensive evaluation of multimodal evolving knowledge injection revealed two challenges: (1) Existing knowledge injection methods perform terribly on evolving knowledge. (2) Supervised fine-tuning causes catastrophic forgetting, particularly instruction following ability is severely compromised. Additionally, we provide pathways and find that: (1) Text knowledge augmentation during the training phase improves performance, while image augmentation cannot achieve it. (2) Continual learning methods, especially Replay and MoELoRA, effectively mitigate forgetting. Our findings indicate that current knowledge injection methods have many limitations on evolving knowledge, which motivates further research on more efficient and stable knowledge injection methods.

Kailin Jiang, Hongbo Jiang, Ning Jiang et al.

Large Multimodal Models encode extensive factual knowledge in their pre-trained weights. However, its knowledge remains static and limited, unable to keep pace with real-world developments, which hinders continuous knowledge acquisition. Effective knowledge injection thus becomes critical, involving two goals: knowledge adaptation (injecting new knowledge) and knowledge retention (preserving old knowledge). Existing methods often struggle to learn new knowledge and suffer from catastrophic forgetting. To address this, we propose KORE, a synergistic method of KnOwledge-oRientEd augmentations and constraints for injecting new knowledge into large multimodal models while preserving old knowledge. Unlike general text or image data augmentation, KORE automatically converts individual knowledge items into structured and comprehensive knowledge to ensure that the model accurately learns new knowledge, enabling accurate adaptation. Meanwhile, KORE stores previous knowledge in the covariance matrix of LMM's linear layer activations and initializes the adapter by projecting the original weights into the matrix's null space, defining a fine-tuning direction that minimizes interference with previous knowledge, enabling powerful retention. Extensive experiments on various LMMs, including LLaVA-v1.5-7B, LLaVA-v1.5-13B, and Qwen2.5-VL-7B, show that KORE achieves superior new knowledge injection performance and effectively mitigates catastrophic forgetting.

Xiaomeng Fan, Yuchuan Mao, Zhi Gao et al.

Open-vocabulary learning requires modeling the data distribution in open environments, which consists of both seen-class and unseen-class data. Existing methods estimate the distribution in open environments using seen-class data, where the absence of unseen classes makes the estimation error inherently unidentifiable. Intuitively, learning beyond the seen classes is crucial for distribution estimation to bound the estimation error. We theoretically demonstrate that the distribution can be effectively estimated by generating unseen-class data, through which the estimation error is upper-bounded. Building on this theoretical insight, we propose a novel open-vocabulary learning method, which generates unseen-class data for estimating the distribution in open environments. The method consists of a class-domain-wise data generation pipeline and a distribution alignment algorithm. The data generation pipeline generates unseen-class data under the guidance of a hierarchical semantic tree and domain information inferred from the seen-class data, facilitating accurate distribution estimation. With the generated data, the distribution alignment algorithm estimates and maximizes the posterior probability to enhance generalization in open-vocabulary learning. Extensive experiments on $11$ datasets demonstrate that our method outperforms baseline approaches by up to $14\%$, highlighting its effectiveness and superiority.

Quan Chen, Chenrui Shi, Qi Chen et al.

Learning from long-horizon demonstrations with complex action sequences presents significant challenges for visual imitation learning, particularly in understanding temporal relationships of actions and spatial relationships between objects. In this paper, we propose a new agent framework that incorporates two dedicated reflection modules to enhance both plan and code generation. The plan generation module produces an initial action sequence, which is then verified by the plan reflection module to ensure temporal coherence and spatial alignment with the demonstration video. The code generation module translates the plan into executable code, while the code reflection module verifies and refines the generated code to ensure correctness and consistency with the generated plan. These two reflection modules jointly enable the agent to detect and correct errors in both the plan generation and code generation, improving performance in tasks with intricate temporal and spatial dependencies. To support systematic evaluation, we introduce LongVILBench, a benchmark comprising 300 human demonstrations with action sequences of up to 18 steps. LongVILBench emphasizes temporal and spatial complexity across multiple task types. Experimental results demonstrate that existing methods perform poorly on this benchmark, whereas our new framework establishes a strong baseline for long-horizon visual imitation learning.

Pengxiang Li, Yuwei Wu, Zhi Gao et al.

Learning in hyperbolic spaces has attracted increasing attention due to its superior ability to model hierarchical structures of data. Most existing hyperbolic learning methods use fixed distance measures for all data, assuming a uniform hierarchy across all data points. However, real-world hierarchical structures exhibit significant diversity, making this assumption overly restrictive. In this paper, we propose a geometry-aware distance measure in hyperbolic spaces, which dynamically adapts to varying hierarchical structures. Our approach derives the distance measure by generating tailored projections and curvatures for each pair of data points, effectively mapping them to an appropriate hyperbolic space. We introduce a revised low-rank decomposition scheme and a hard-pair mining mechanism to mitigate the computational cost of pair-wise distance computation without compromising accuracy. We present an upper bound on the low-rank approximation error using Talagrand's concentration inequality, ensuring theoretical robustness. Extensive experiments on standard image classification (MNIST, CIFAR-10 and CIFAR-100), hierarchical classification (5-level CIFAR-100), and few-shot learning tasks (mini-ImageNet, tiered-ImageNet) demonstrate the effectiveness of our method. Our approach consistently outperforms learning methods that use fixed distance measures, with notable improvements on few-shot learning tasks, where it achieves over 5\% gains on mini-ImageNet. The results reveal that adaptive distance measures better capture diverse hierarchical structures, with visualization showing clearer class boundaries and improved prototype separation in hyperbolic spaces.

Peilin Yu, Yuwei Wu, Zhi Gao et al.

Feature augmentation generates novel samples in the feature space, providing an effective way to enhance the generalization ability of learning algorithms with hyperbolic geometry. Most hyperbolic feature augmentation is confined to closed-environment, assuming the number of classes is fixed (\emph{i.e.}, seen classes) and generating features only for these classes. In this paper, we propose a hyperbolic dual feature augmentation method for open-environment, which augments features for both seen and unseen classes in the hyperbolic space. To obtain a more precise approximation of the real data distribution for efficient training, (1) we adopt a neural ordinary differential equation module, enhanced by meta-learning, estimating the feature distributions of both seen and unseen classes; (2) we then introduce a regularizer to preserve the latent hierarchical structures of data in the hyperbolic space; (3) we also derive an upper bound for the hyperbolic dual augmentation loss, allowing us to train a hyperbolic model using infinite augmentations for seen and unseen classes. Extensive experiments on five open-environment tasks: class-incremental learning, few-shot open-set recognition, few-shot learning, zero-shot learning, and general image classification, demonstrate that our method effectively enhances the performance of hyperbolic algorithms in open-environment.

Yapeng Mi, Zhi Gao, Xiaojian Ma et al.

LLM-driven autonomous agents have emerged as a promising direction in recent years. However, many of these LLM agents are designed empirically or based on intuition, often lacking systematic design principles, which results in diverse agent structures with limited generality and scalability. In this paper, we advocate for building LLM agents by incorporating insights from computer systems. Inspired by the von Neumann architecture, we propose a structured framework for LLM agentic systems, emphasizing modular design and universal principles. Specifically, this paper first provides a comprehensive review of LLM agents from the computer system perspective, then identifies key challenges and future directions inspired by computer system design, and finally explores the learning mechanisms for LLM agents beyond the computer system. The insights gained from this comparative analysis offer a foundation for systematic LLM agent design and advancement.

Peilin Yu, Yuwei Wu, Zhi Gao et al.

Riemannian meta-optimization provides a promising approach to solving non-linear constrained optimization problems, which trains neural networks as optimizers to perform optimization on Riemannian manifolds. However, existing Riemannian meta-optimization methods take up huge memory footprints in large-scale optimization settings, as the learned optimizer can only adapt gradients of a fixed size and thus cannot be shared across different Riemannian parameters. In this paper, we propose an efficient Riemannian meta-optimization method that significantly reduces the memory burden for large-scale optimization via a subspace adaptation scheme. Our method trains neural networks to individually adapt the row and column subspaces of Riemannian gradients, instead of directly adapting the full gradient matrices in existing Riemannian meta-optimization methods. In this case, our learned optimizer can be shared across Riemannian parameters with different sizes. Our method reduces the model memory consumption by six orders of magnitude when optimizing an orthogonal mainstream deep neural network (e.g., ResNet50). Experiments on multiple Riemannian tasks show that our method can not only reduce the memory consumption but also improve the performance of Riemannian meta-optimization.

Haoyu Zhang, Yuwei Wu, Pengxiang Li et al.

In the era of Vision-Language Models (VLMs), enhancing multimodal reasoning capabilities remains a critical challenge, particularly in handling ambiguous or complex visual inputs, where initial inferences often lead to hallucinations or logic errors. Existing VLMs often produce plausible yet ungrounded answers, and even when prompted to "reflect", their corrections may remain detached from the image evidence. To address this, we propose the MIRROR framework for Multimodal Iterative Reasoning via Reflection On visual Regions. By embedding visual reflection as a core mechanism, MIRROR is formulated as a closed-loop process comprising draft, critique, region-based verification, and revision, which are repeated until the output is visually grounded. To facilitate training of this model, we construct **ReflectV**, a visual reflective dataset for multi-turn supervision that explicitly contains reflection triggers, region-based verification actions, and answer revision grounded in visual evidence. Experiments on both general vision-language benchmarks and representative vision-language reasoning benchmarks show that MIRROR improves correctness and reduces visual hallucinations, demonstrating the value of training reflection as an evidence-seeking, region-aware verification process rather than a purely textual revision step.

Xu Chen, Rui Gao, Xinjie Zhang et al.

Achieving natural dyadic interaction requires generating facial expressions that are emotionally appropriate and socially aligned with human preference. Human feedback offers a compelling mechanism to guide such alignment, yet how to effectively incorporate this feedback into facial expression generation remains underexplored. In this paper, we propose a facial expression generation method aligned with human preference by leveraging human feedback to produce contextually and emotionally appropriate expressions for natural dyadic interaction. A key to our method is framing the generation of identity-independent facial expressions as an action learning process, allowing human feedback to assess their validity free from visual or identity bias. We establish a closed feedback loop in which listener expressions dynamically respond to evolving conversational cues of the speaker. Concretely, we train a vision-language-action model via supervised fine-tuning to map the speaker's multimodal signals into controllable low-dimensional expression representations of a 3D morphable model. We further introduce a human-feedback reinforcement learning strategy that integrates the imitation of high-quality expression response with critic-guided optimization. Experiments on two benchmarks demonstrate that our method effectively aligns facial expressions with human preference and achieves superior performance.

Xintong Zhang, Xiaowen Zhang, Jongrong Wu et al.

Adaptive multimodal reasoning has emerged as a promising frontier in Vision-Language Models (VLMs), aiming to dynamically modulate between tool-augmented visual reasoning and text reasoning to enhance both effectiveness and efficiency. However, existing evaluations rely on static difficulty labels and simplistic metrics, which fail to capture the dynamic nature of difficulty relative to varying model capacities. Consequently, they obscure the distinction between adaptive mode selection and general performance while neglecting fine-grained process analyses. In this paper, we propose AdaptMMBench, a comprehensive benchmark for adaptive multimodal reasoning across five domains: real-world, OCR, GUI, knowledge, and math, encompassing both direct perception and complex reasoning tasks. AdaptMMBench utilizes a Matthews Correlation Coefficient (MCC) metric to evaluate the selection rationality of different reasoning modes, isolating this meta-cognition ability by dynamically identifying task difficulties based on models' capability boundaries. Moreover, AdaptMMBench facilitates multi-dimensional process evaluation across key step coverage, tool effectiveness, and computational efficiency. Our evaluation reveals that while adaptive mode selection scales with model capacity, it notably decouples from final accuracy. Conversely, key step coverage aligns with performance, though tool effectiveness remains highly inconsistent across model architectures.

Mingliang Zhai, Hansheng Liang, Xiaomeng Fan et al.

Embodied Question Answering (EQA) requires agents to explore 3D environments to obtain observations and answer questions related to the scene. Existing methods leverage VLMs to directly explore the environment and answer questions without explicit thinking or planning, which limits their reasoning ability and results in excessive or inefficient exploration as well as ineffective responses. In this paper, we introduce ToolEQA, an agent that integrates external tools with multi-step reasoning, where external tools can provide more useful information for completing the task, helping the model derive better exploration directions in the next step of reasoning and thus obtaining additional effective information. This enables ToolEQA to generate more accurate responses with a shorter exploration distance. To enhance the model's ability for tool-usage and multi-step reasoning, we further design a novel EQA data generation pipeline that automatically constructs large-scale EQA tasks with reasoning trajectories and corresponding answers. Based on the pipeline, we collect the EQA-RT dataset that contains about 18K tasks, divided into a training set EQA-RT-Train, and two test sets EQA-RT-Seen (scenes overlapping with the training set) and EQA-RT-Unseen (novel scenes). Experiments on EQA-RT-Seen and EQA-RT-Unseen show that ToolEQA improves the success rate by 9.2~20.2% over state-of-the-art baselines, while outperforming the zero-shot ToolEQA by 10% in success rate. In addition, ToolEQA also achieves state-of-the-art performance on the HM-EQA, OpenEQA, and EXPRESS-Bench datasets, demonstrating its generality. Our homepage see https://tooleqa.github.io.

Yuchuan Mao, Zhi Gao, Xiaomeng Fan et al.

Model ensemble is an effective strategy in continual learning, which alleviates catastrophic forgetting by interpolating model parameters, achieving knowledge fusion learned from different tasks. However, existing model ensemble methods usually encounter the knowledge conflict issue at task and layer levels, causing compromised learning performance in both old and new tasks. To solve this issue, we propose meta-weight-ensembler that adaptively fuses knowledge of different tasks for continual learning. Concretely, we employ a mixing coefficient generator trained via meta-learning to generate appropriate mixing coefficients for model ensemble to address the task-level knowledge conflict. The mixing coefficient is individually generated for each layer to address the layer-level knowledge conflict. In this way, we learn the prior knowledge about adaptively accumulating knowledge of different tasks in a fused model, achieving efficient learning in both old and new tasks. Meta-weight-ensembler can be flexibly combined with existing continual learning methods to boost their ability of alleviating catastrophic forgetting. Experiments on multiple continual learning datasets show that meta-weight-ensembler effectively alleviates catastrophic forgetting and achieves state-of-the-art performance.

Pengxiang Li, Wei Wu, Zhi Gao et al.

We propose a hyperbolic set-to-set distance measure for computing dissimilarity between sets in hyperbolic space. While point-to-point distances in hyperbolic space effectively capture hierarchical relationships between data points, many real-world applications require comparing sets of hyperbolic data points, where the local structure and the global structure of the sets carry crucial semantic information. The proposed the \underline{h}yperbolic \underline{s}et-\underline{to}-\underline{s}et \underline{d}istance measure (HS2SD) integrates both global and local structural information: global structure through geodesic distances between Einstein midpoints of hyperbolic sets, and local structure through topological characteristics of the two sets. To efficiently compute topological differences, we prove that using a finite Thue-Morse sequence of degree and adjacency matrices can serve as a robust approximation to capture the topological structure of a set. In this case, by considering the topological differences, HS2SD provides a more nuanced understanding of the relationships between two hyperbolic sets. Empirical evaluation on entity matching, standard image classification, and few-shot image classification demonstrates that our distance measure outperforms existing methods by effectively modeling the hierarchical and complex relationships inherent in hyperbolic sets.

Ziyi Wang, Zhi Gao, Boxuan Yu et al.

Video understanding has made remarkable progress in recent years, largely driven by advances in deep models and the availability of large-scale annotated datasets. However, existing works typically ignore the inherent domain shifts encountered in real-world video applications, leaving domain generalization (DG) in video understanding underexplored. Hence, we propose Video Understanding Domain Generalization (VUDG), a novel dataset designed specifically for evaluating the DG performance in video understanding. VUDG contains videos from 11 distinct domains that cover three types of domain shifts, and maintains semantic similarity across different domains to ensure fair and meaningful evaluation. We propose a multi-expert progressive annotation framework to annotate each video with both multiple-choice and open-ended question-answer pairs. Extensive experiments on 9 representative large video-language models (LVLMs) and several traditional video question answering methods show that most models (including state-of-the-art LVLMs) suffer performance degradation under domain shifts. These results highlight the challenges posed by VUDG and the difference in the robustness of current models to data distribution shifts. We believe VUDG provides a valuable resource for prompting future research in domain generalization video understanding.

Chenguang Wang, Ensieh Sharifnia, Zhi Gao et al.

For planning of power systems and for the calibration of operational tools, it is essential to analyse system performance in a large range of representative scenarios. When the available historical data is limited, generative models are a promising solution, but modelling high-dimensional dependencies is challenging. In this paper, a multivariate load state generating model on the basis of a conditional variational autoencoder (CVAE) neural network is proposed. Going beyond common CVAE implementations, the model includes stochastic variation of output samples under given latent vectors and co-optimizes the parameters for this output variability. It is shown that this improves statistical properties of the generated data. The quality of generated multivariate loads is evaluated using univariate and multivariate performance metrics. A generation adequacy case study on the European network is used to illustrate model's ability to generate realistic tail distributions. The experiments demonstrate that the proposed generator outperforms other data generating mechanisms.

Weng Fei Low, Ankit Sonthalia, Zhi Gao et al.

Most successful computer vision models transform low-level features, such as Gabor filter responses, into richer representations of intermediate or mid-level complexity for downstream visual tasks. These mid-level representations have not been explored for event cameras, although it is especially relevant to the visually sparse and often disjoint spatial information in the event stream. By making use of locally consistent intermediate representations, termed as superevents, numerous visual tasks ranging from semantic segmentation, visual tracking, depth estimation shall benefit. In essence, superevents are perceptually consistent local units that delineate parts of an object in a scene. Inspired by recent deep learning architectures, we present a novel method that employs lifetime augmentation for obtaining an event stream representation that is fed to a fully convolutional network to extract superevents. Our qualitative and quantitative experimental results on several sequences of a benchmark dataset highlights the significant potential for event-based downstream applications.

Jindou Dai, Yuwei Wu, Zhi Gao et al.

Hyperbolic graph convolutional networks (GCNs) demonstrate powerful representation ability to model graphs with hierarchical structure. Existing hyperbolic GCNs resort to tangent spaces to realize graph convolution on hyperbolic manifolds, which is inferior because tangent space is only a local approximation of a manifold. In this paper, we propose a hyperbolic-to-hyperbolic graph convolutional network (H2H-GCN) that directly works on hyperbolic manifolds. Specifically, we developed a manifold-preserving graph convolution that consists of a hyperbolic feature transformation and a hyperbolic neighborhood aggregation. The hyperbolic feature transformation works as linear transformation on hyperbolic manifolds. It ensures the transformed node representations still lie on the hyperbolic manifold by imposing the orthogonal constraint on the transformation sub-matrix. The hyperbolic neighborhood aggregation updates each node representation via the Einstein midpoint. The H2H-GCN avoids the distortion caused by tangent space approximations and keeps the global hyperbolic structure. Extensive experiments show that the H2H-GCN achieves substantial improvements on the link prediction, node classification, and graph classification tasks.

Kangcheng Liu, Zhi Gao, Feng Lin et al.

This work presents FG-Net, a general deep learning framework for large-scale point clouds understanding without voxelizations, which achieves accurate and real-time performance with a single NVIDIA GTX 1080 GPU. First, a novel noise and outlier filtering method is designed to facilitate subsequent high-level tasks. For effective understanding purpose, we propose a deep convolutional neural network leveraging correlated feature mining and deformable convolution based geometric-aware modelling, in which the local feature relationships and geometric patterns can be fully exploited. For the efficiency issue, we put forward an inverse density sampling operation and a feature pyramid based residual learning strategy to save the computational cost and memory consumption respectively. Extensive experiments on real-world challenging datasets demonstrated that our approaches outperform state-of-the-art approaches in terms of accuracy and efficiency. Moreover, weakly supervised transfer learning is also conducted to demonstrate the generalization capacity of our method.

Xiaodong Liu, Zhi Gao, Ben M. Chen

Color correction for underwater images has received increasing interests, due to its critical role in facilitating available mature vision algorithms for underwater scenarios. Inspired by the stunning success of deep convolutional neural networks (DCNNs) techniques in many vision tasks, especially the strength in extracting features in multiple scales, we propose a deep multi-scale feature fusion net based on the conditional generative adversarial network (GAN) for underwater image color correction. In our network, multi-scale features are extracted first, followed by augmenting local features on each scale with global features. This design was verified to facilitate more effective and faster network learning, resulting in better performance in both color correction and detail preservation. We conducted extensive experiments and compared with the state-of-the-art approaches quantitatively and qualitatively, showing that our method achieves significant improvements.

Zhi Gao, Yuwei Wu, Xingyuan Bu et al.

Recent studies have shown that aggregating convolutional features of a pre-trained Convolutional Neural Network (CNN) can obtain impressive performance for a variety of visual tasks. The symmetric Positive Definite (SPD) matrix becomes a powerful tool due to its remarkable ability to learn an appropriate statistic representation to characterize the underlying structure of visual features. In this paper, we propose to aggregate deep convolutional features into an SPD matrix representation through the SPD generation and the SPD transformation under an end-to-end deep network. To this end, several new layers are introduced in our network, including a nonlinear kernel aggregation layer, an SPD matrix transformation layer, and a vectorization layer. The nonlinear kernel aggregation layer is employed to aggregate the convolutional features into a real SPD matrix directly. The SPD matrix transformation layer is designed to construct a more compact and discriminative SPD representation. The vectorization and normalization operations are performed in the vectorization layer for reducing the redundancy and accelerating the convergence. The SPD matrix in our network can be considered as a mid-level representation bridging convolutional features and high-level semantic features. To demonstrate the effectiveness of our method, we conduct extensive experiments on visual classification. Experiment results show that our method notably outperforms state-of-the-art methods.

Mo Shan, Fei Wang, Feng Lin et al.

We propose a framework for Google Map aided UAV navigation in GPS-denied environment. Geo-referenced navigation provides drift-free localization and does not require loop closures. The UAV position is initialized via correlation, which is simple and efficient. We then use optical flow to predict its position in subsequent frames. During pose tracking, we obtain inter-frame translation either by motion field or homography decomposition, and we use HOG features for registration on Google Map. We employ particle filter to conduct a coarse to fine search to localize the UAV. Offline test using aerial images collected by our quadrotor platform shows promising results as our approach eliminates the drift in dead-reckoning, and the small localization error indicates the superiority of our approach as a supplement to GPS.