Qi Zhao

Shi Chen, Qi Zhao

Effectiveness and interpretability are two essential properties for trustworthy AI systems. Most recent studies in visual reasoning are dedicated to improving the accuracy of predicted answers, and less attention is paid to explaining the rationales behind the decisions. As a result, they commonly take advantage of spurious biases instead of actually reasoning on the visual-textual data, and have yet developed the capability to explain their decision making by considering key information from both modalities. This paper aims to close the gap from three distinct perspectives: first, we define a new type of multi-modal explanations that explain the decisions by progressively traversing the reasoning process and grounding keywords in the images. We develop a functional program to sequentially execute different reasoning steps and construct a new dataset with 1,040,830 multi-modal explanations. Second, we identify the critical need to tightly couple important components across the visual and textual modalities for explaining the decisions, and propose a novel explanation generation method that explicitly models the pairwise correspondence between words and regions of interest. It improves the visual grounding capability by a considerable margin, resulting in enhanced interpretability and reasoning performance. Finally, with our new data and method, we perform extensive analyses to study the effectiveness of our explanation under different settings, including multi-task learning and transfer learning. Our code and data are available at https://github.com/szzexpoi/rex.

Chang Shao, Qi Zhao, Nana Pu et al.

The field of Dynamic Multi-Objective Optimization (DMOO) has witnessed a surge of interest from both academia and industry, as numerous time-evolving real-world applications can be naturally formulated as Dynamic Multi-Objective Optimization Problems (DMOPs). This growing demand thus necessitates advanced benchmarks to rigorously evaluate optimization algorithms under realistic conditions. This paper introduces a comprehensive and principled framework for constructing highly realistic and challenging DMOO benchmarks. The proposed framework incorporates several novel components, including: a generalized formulation that allows the Pareto-optimal Set (PS) to change on hypersurfaces; a mechanism for creating controlled variable contribution imbalances to generate heterogeneous landscapes; and dynamic rotation matrices for inducing time-varying variable interactions and non-separability. Furthermore, we incorporate a temporal perturbation mechanism to simulate irregular environmental changes and propose a generalized time-linkage mechanism that systematically embeds historical solution quality into future problems, thereby capturing critical real-world phenomena such as error accumulation and time-deception. Extensive experimental results validate the effectiveness of the proposed framework, demonstrating its superiority over conventional benchmarks in terms of realism, complexity, and its capability for discriminating state-of-the-art algorithmic performance. Thus, this work establishes a new standard for dynamic multi-objective optimization benchmarking and provides a powerful tool for the development and evaluation of next-generation algorithms capable of addressing the complexities of real-world dynamic systems.

Chunlei Wang, Wenquan Feng, Xiangtai Li et al.

Open-vocabulary learning has emerged as a cutting-edge research area, particularly in light of the widespread adoption of vision-based foundational models. Its primary objective is to comprehend novel concepts that are not encompassed within a predefined vocabulary. One key facet of this endeavor is Visual Grounding, which entails locating a specific region within an image based on a corresponding language description. While current foundational models excel at various visual language tasks, there's a noticeable absence of models specifically tailored for open-vocabulary visual grounding. This research endeavor introduces novel and challenging OV tasks, namely Open-Vocabulary Visual Grounding and Open-Vocabulary Phrase Localization. The overarching aim is to establish connections between language descriptions and the localization of novel objects. To facilitate this, we have curated a comprehensive annotated benchmark, encompassing 7,272 OV-VG images and 1,000 OV-PL images. In our pursuit of addressing these challenges, we delved into various baseline methodologies rooted in existing open-vocabulary object detection, VG, and phrase localization frameworks. Surprisingly, we discovered that state-of-the-art methods often falter in diverse scenarios. Consequently, we developed a novel framework that integrates two critical components: Text-Image Query Selection and Language-Guided Feature Attention. These modules are designed to bolster the recognition of novel categories and enhance the alignment between visual and linguistic information. Extensive experiments demonstrate the efficacy of our proposed framework, which consistently attains SOTA performance across the OV-VG task. Additionally, ablation studies provide further evidence of the effectiveness of our innovative models. Codes and datasets will be made publicly available at https://github.com/cv516Buaa/OV-VG.

Yilun Kong, Hangyu Mao, Qi Zhao et al.

Prompt engineering has demonstrated remarkable success in enhancing the performance of large language models (LLMs) across diverse tasks. However, most existing prompt optimization methods only focus on the task-level performance, overlooking the importance of query-preferred prompts, which leads to suboptimal performances. Additionally, these methods rely heavily on frequent interactions with LLMs to obtain feedback for guiding the optimization process, incurring substantial redundant interaction costs. In this paper, we introduce Query-dependent Prompt Optimization (QPO), which leverages multi-loop offline reinforcement learning to iteratively fine-tune a small pretrained language model to generate optimal prompts tailored to the input queries, thus significantly improving the prompting effect on the large target LLM. We derive insights from offline prompting demonstration data, which already exists in large quantities as a by-product of benchmarking diverse prompts on open-sourced tasks, thereby circumventing the expenses of online interactions. Furthermore, we continuously augment the offline dataset with the generated prompts in each loop, as the prompts from the fine-tuned model are supposed to outperform the source prompts in the original dataset. These iterative loops bootstrap the model towards generating optimal prompts. Experiments on various LLM scales and diverse NLP and math tasks demonstrate the efficacy and cost-efficiency of our method in both zero-shot and few-shot scenarios.

Shi Chen, Ming Jiang, Jinhui Yang et al.

While attention has been an increasingly popular component in deep neural networks to both interpret and boost the performance of models, little work has examined how attention progresses to accomplish a task and whether it is reasonable. In this work, we propose an Attention with Reasoning capability (AiR) framework that uses attention to understand and improve the process leading to task outcomes. We first define an evaluation metric based on a sequence of atomic reasoning operations, enabling a quantitative measurement of attention that considers the reasoning process. We then collect human eye-tracking and answer correctness data, and analyze various machine and human attention mechanisms on their reasoning capability and how they impact task performance. To improve the attention and reasoning ability of visual question answering models, we propose to supervise the learning of attention progressively along the reasoning process and to differentiate the correct and incorrect attention patterns. We demonstrate the effectiveness of the proposed framework in analyzing and modeling attention with better reasoning capability and task performance. The code and data are available at https://github.com/szzexpoi/AiR

Qi Zhao, Shuchang Lyu, Binghao Liu et al.

Deep convolutional neural networks (DCNNs) based remote sensing (RS) image semantic segmentation technology has achieved great success used in many real-world applications such as geographic element analysis. However, strong dependency on annotated data of specific scene makes it hard for DCNNs to fit different RS scenes. To solve this problem, recent works gradually focus on cross-domain RS image semantic segmentation task. In this task, different ground sampling distance, remote sensing sensor variation and different geographical landscapes are three main factors causing dramatic domain shift between source and target images. To decrease the negative influence of domain shift, we propose a self-training guided disentangled adaptation network (ST-DASegNet). We first propose source student backbone and target student backbone to respectively extract the source-style and target-style feature for both source and target images. Towards the intermediate output feature maps of each backbone, we adopt adversarial learning for alignment. Then, we propose a domain disentangled module to extract the universal feature and purify the distinct feature of source-style and target-style features. Finally, these two features are fused and served as input of source student decoder and target student decoder to generate final predictions. Based on our proposed domain disentangled module, we further propose exponential moving average (EMA) based cross-domain separated self-training mechanism to ease the instability and disadvantageous effect during adversarial optimization. Extensive experiments and analysis on benchmark RS datasets show that ST-DASegNet outperforms previous methods on cross-domain RS image semantic segmentation task and achieves state-of-the-art (SOTA) results. Our code is available at https://github.com/cv516Buaa/ST-DASegNet.

Menghao Li, Chunlei Wang, Wenquan Feng et al.

Visual Grounding (VG) aims at localizing target objects from an image based on given expressions and has made significant progress with the development of detection and vision transformer. However, existing VG methods tend to generate false-alarm objects when presented with inaccurate or irrelevant descriptions, which commonly occur in practical applications. Moreover, existing methods fail to capture fine-grained features, accurate localization, and sufficient context comprehension from the whole image and textual descriptions. To address both issues, we propose an Iterative Robust Visual Grounding (IR-VG) framework with Masked Reference based Centerpoint Supervision (MRCS). The framework introduces iterative multi-level vision-language fusion (IMVF) for better alignment. We use MRCS to ahieve more accurate localization with point-wised feature supervision. Then, to improve the robustness of VG, we also present a multi-stage false-alarm sensitive decoder (MFSD) to prevent the generation of false-alarm objects when presented with inaccurate expressions. The proposed framework is evaluated on five regular VG datasets and two newly constructed robust VG datasets. Extensive experiments demonstrate that IR-VG achieves new state-of-the-art (SOTA) results, with improvements of 25\% and 10\% compared to existing SOTA approaches on the two newly proposed robust VG datasets. Moreover, the proposed framework is also verified effective on five regular VG datasets. Codes and models will be publicly at https://github.com/cv516Buaa/IR-VG.

Qi Zhao, Shuchang Lyu, Wenpei Bai et al.

Ovarian cancer is one of the most harmful gynecological diseases. Detecting ovarian tumors in early stage with computer-aided techniques can efficiently decrease the mortality rate. With the improvement of medical treatment standard, ultrasound images are widely applied in clinical treatment. However, recent notable methods mainly focus on single-modality ultrasound ovarian tumor segmentation or recognition, which means there still lacks researches on exploring the representation capability of multi-modality ultrasound ovarian tumor images. To solve this problem, we propose a Multi-Modality Ovarian Tumor Ultrasound (MMOTU) image dataset containing 1469 2d ultrasound images and 170 contrast enhanced ultrasonography (CEUS) images with pixel-wise and global-wise annotations. Based on MMOTU, we mainly focus on unsupervised cross-domain semantic segmentation task. To solve the domain shift problem, we propose a feature alignment based architecture named Dual-Scheme Domain-Selected Network (DS2Net). Specifically, we first design source-encoder and target-encoder to extract two-style features of source and target images. Then, we propose Domain-Distinct Selected Module (DDSM) and Domain-Universal Selected Module (DUSM) to extract the distinct and universal features in two styles (source-style or target-style). Finally, we fuse these two kinds of features and feed them into the source-decoder and target-decoder to generate final predictions. Extensive comparison experiments and analysis on MMOTU image dataset show that DS2Net can boost the segmentation performance for bidirectional cross-domain adaptation of 2d ultrasound images and CEUS images. Our proposed dataset and code are all available at https://github.com/cv516Buaa/MMOTU_DS2Net.

Shi Chen, Ming Jiang, Qi Zhao

In recent years, deep saliency models have made significant progress in predicting human visual attention. However, the mechanisms behind their success remain largely unexplained due to the opaque nature of deep neural networks. In this paper, we present a novel analytic framework that sheds light on the implicit features learned by saliency models and provides principled interpretation and quantification of their contributions to saliency prediction. Our approach decomposes these implicit features into interpretable bases that are explicitly aligned with semantic attributes and reformulates saliency prediction as a weighted combination of probability maps connecting the bases and saliency. By applying our framework, we conduct extensive analyses from various perspectives, including the positive and negative weights of semantics, the impact of training data and architectural designs, the progressive influences of fine-tuning, and common failure patterns of state-of-the-art deep saliency models. Additionally, we demonstrate the effectiveness of our framework by exploring visual attention characteristics in various application scenarios, such as the atypical attention of people with autism spectrum disorder, attention to emotion-eliciting stimuli, and attention evolution over time. Our code is publicly available at \url{https://github.com/szzexpoi/saliency_analysis}.

Qi Zhao, Shijie Wang, Ce Zhang et al.

Can we better anticipate an actor's future actions (e.g. mix eggs) by knowing what commonly happens after his/her current action (e.g. crack eggs)? What if we also know the longer-term goal of the actor (e.g. making egg fried rice)? The long-term action anticipation (LTA) task aims to predict an actor's future behavior from video observations in the form of verb and noun sequences, and it is crucial for human-machine interaction. We propose to formulate the LTA task from two perspectives: a bottom-up approach that predicts the next actions autoregressively by modeling temporal dynamics; and a top-down approach that infers the goal of the actor and plans the needed procedure to accomplish the goal. We hypothesize that large language models (LLMs), which have been pretrained on procedure text data (e.g. recipes, how-tos), have the potential to help LTA from both perspectives. It can help provide the prior knowledge on the possible next actions, and infer the goal given the observed part of a procedure, respectively. To leverage the LLMs, we propose a two-stage framework, AntGPT. It first recognizes the actions already performed in the observed videos and then asks an LLM to predict the future actions via conditioned generation, or to infer the goal and plan the whole procedure by chain-of-thought prompting. Empirical results on the Ego4D LTA v1 and v2 benchmarks, EPIC-Kitchens-55, as well as EGTEA GAZE+ demonstrate the effectiveness of our proposed approach. AntGPT achieves state-of-the-art performance on all above benchmarks, and can successfully infer the goal and thus perform goal-conditioned "counterfactual" prediction via qualitative analysis. Code and model will be released at https://brown-palm.github.io/AntGPT

Yuqing Wang, Zhijie Lin, Ceyuan Yang et al.

Unified multimodal models (UMMs) aim to handle perception and generation in a single model. Yet existing UMMs still rely on a frozen, separately pretrained VAE for image generation, imposing a structural bottleneck. Naively removing it introduces a quality gap, as the model must learn both high-level structure and low-level details from raw pixels. In this paper, we propose Representation Forcing (RF), a technique that closes this gap by making representation prediction a native capability of the model. Concretely, RF forces the decoder to autoregressively predict visual representations as intermediate tokens before pixels; these tokens then stay in context to guide pixel diffusion within the same backbone. By turning representations from perception outputs into generation targets, RF eliminates the need for any external generative latent space. We find that RF benefits both understanding and generation. On image generation, our pixel-space model with RF matches state-of-the-art VAE-based unified models. On image understanding, pixel-space RF generally outperforms its VAE-based variant. Together, these results offer an effective step toward end-to-end, bottleneck-free UMMs.

Shi Chen, Qi Zhao

Humans have the innate capability to answer diverse questions, which is rooted in the natural ability to correlate different concepts based on their semantic relationships and decompose difficult problems into sub-tasks. On the contrary, existing visual reasoning methods assume training samples that capture every possible object and reasoning problem, and rely on black-boxed models that commonly exploit statistical priors. They have yet to develop the capability to address novel objects or spurious biases in real-world scenarios, and also fall short of interpreting the rationales behind their decisions. Inspired by humans' reasoning of the visual world, we tackle the aforementioned challenges from a compositional perspective, and propose an integral framework consisting of a principled object factorization method and a novel neural module network. Our factorization method decomposes objects based on their key characteristics, and automatically derives prototypes that represent a wide range of objects. With these prototypes encoding important semantics, the proposed network then correlates objects by measuring their similarity on a common semantic space and makes decisions with a compositional reasoning process. It is capable of answering questions with diverse objects regardless of their availability during training, and overcoming the issues of biased question-answer distributions. In addition to the enhanced generalizability, our framework also provides an interpretable interface for understanding the decision-making process of models. Our code is available at https://github.com/szzexpoi/POEM.

Qi Zhao, Qiqi Duan, Bai Yan et al.

Metaheuristics have gained great success in academia and practice because their search logic can be applied to any problem with available solution representation, solution quality evaluation, and certain notions of locality. Manually designing metaheuristic algorithms for solving a target problem is criticized for being laborious, error-prone, and requiring intensive specialized knowledge. This gives rise to increasing interest in automated design of metaheuristic algorithms. With computing power to fully explore potential design choices, the automated design could reach and even surpass human-level design and could make high-performance algorithms accessible to a much wider range of researchers and practitioners. This paper presents a broad picture of automated design of metaheuristic algorithms, by conducting a survey on the common grounds and representative techniques in terms of design space, design strategies, performance evaluation strategies, and target problems in this field.

Qi Zhao, Haotian Fu, Chen Sun et al.

Long-horizon decision-making tasks present significant challenges for LLM-based agents due to the need for extensive planning over multiple steps. In this paper, we propose a hierarchical framework that decomposes complex tasks into manageable subgoals, utilizing separate LLMs for subgoal prediction and low-level action generation. To address the challenge of creating training signals for unannotated datasets, we develop a reward model that leverages multimodal environment feedback to automatically generate reward signals. We introduce Environment Preference Optimization (EPO), a novel method that generates preference signals from the environment's feedback and uses them to train LLM-based agents. Extensive experiments on ALFRED demonstrate the state-of-the-art performance of our framework, achieving first place on the ALFRED public leaderboard and showcasing its potential to improve long-horizon decision-making in diverse environments.

Yuwen Lu, Meng Chen, Qi Zhao et al.

Collecting mobile datasets remains challenging for academic researchers due to limited data access and technical barriers. Commercial organizations often possess exclusive access to mobile data, leading to a "data monopoly" that restricts the independence of academic research. Existing open-source mobile data collection frameworks primarily focus on mobile sensing data rather than screen content, which is crucial for various research studies. We present Crepe, a no-code Android app that enables researchers to collect information displayed on screen through simple demonstrations of target data. Crepe utilizes a novel Graph Query technique which augments the structures of mobile UI screens to support flexible identification, location, and collection of specific data pieces. The tool emphasizes participants' privacy and agency by providing full transparency over collected data and allowing easy opt-out. We designed and built Crepe for research purposes only and in scenarios where researchers obtain explicit consent from participants. Code for Crepe will be open-sourced to support future academic research data collection.

Qi Zhao, M. Salman Asif, Zhan Ma

Existing implicit neural representation (INR) methods do not fully exploit spatiotemporal redundancies in videos. Index-based INRs ignore the content-specific spatial features and hybrid INRs ignore the contextual dependency on adjacent frames, leading to poor modeling capability for scenes with large motion or dynamics. We analyze this limitation from the perspective of function fitting and reveal the importance of frame difference. To use explicit motion information, we propose Difference Neural Representation for Videos (DNeRV), which consists of two streams for content and frame difference. We also introduce a collaborative content unit for effective feature fusion. We test DNeRV for video compression, inpainting, and interpolation. DNeRV achieves competitive results against the state-of-the-art neural compression approaches and outperforms existing implicit methods on downstream inpainting and interpolation for $960 \times 1920$ videos.

Qi Zhao, Bai Yan, Taiwei Hu et al.

Metaheuristics are widely recognized gradient-free solvers to hard problems that do not meet the rigorous mathematical assumptions of conventional solvers. The automated design of metaheuristic algorithms provides an attractive path to relieve manual design effort and gain enhanced performance beyond human-made algorithms. However, the specific algorithm prototype and linear algorithm representation in the current automated design pipeline restrict the design within a fixed algorithm structure, which hinders discovering novelties and diversity across the metaheuristic family. To address this challenge, this paper proposes a general framework, AutoOpt, for automatically designing metaheuristic algorithms with diverse structures. AutoOpt contains three innovations: (i) A general algorithm prototype dedicated to covering the metaheuristic family as widely as possible. It promotes high-quality automated design on different problems by fully discovering potentials and novelties across the family. (ii) A directed acyclic graph algorithm representation to fit the proposed prototype. Its flexibility and evolvability enable discovering various algorithm structures in a single run of design, thus boosting the possibility of finding high-performance algorithms. (iii) A graph representation embedding method offering an alternative compact form of the graph to be manipulated, which ensures AutoOpt's generality. Experiments on numeral functions and real applications validate AutoOpt's efficiency and practicability.

Diu Khue Luu, Anh Tuan Nguyen, Ming Jiang et al.

Objective: The next generation prosthetic hand that moves and feels like a real hand requires a robust neural interconnection between the human minds and machines. Methods: Here we present a neuroprosthetic system to demonstrate that principle by employing an artificial intelligence (AI) agent to translate the amputee's movement intent through a peripheral nerve interface. The AI agent is designed based on the recurrent neural network (RNN) and could simultaneously decode six degree-of-freedom (DOF) from multichannel nerve data in real-time. The decoder's performance is characterized in motor decoding experiments with three human amputees. Results: First, we show the AI agent enables amputees to intuitively control a prosthetic hand with individual finger and wrist movements up to 97-98% accuracy. Second, we demonstrate the AI agent's real-time performance by measuring the reaction time and information throughput in a hand gesture matching task. Third, we investigate the AI agent's long-term uses and show the decoder's robust predictive performance over a 16-month implant duration. Conclusion & significance: Our study demonstrates the potential of AI-enabled nerve technology, underling the next generation of dexterous and intuitive prosthetic hands.

Qi Zhao, Bai Yan, Taiwei Hu et al.

Metaheuristics are prominent gradient-free optimizers for solving hard problems that do not meet the rigorous mathematical assumptions of analytical solvers. The canonical manual optimizer design could be laborious, untraceable and error-prone, let alone human experts are not always available. This arises increasing interest and demand in automating the optimizer design process. In response, this paper proposes AutoOptLib, the first platform for accessible automated design of metaheuristic optimizers. AutoOptLib leverages computing resources to conceive, build up, and verify the design choices of the optimizers. It requires much less labor resources and expertise than manual design, democratizing satisfactory metaheuristic optimizers to a much broader range of researchers and practitioners. Furthermore, by fully exploring the design choices with computing resources, AutoOptLib has the potential to surpass human experience, subsequently gaining enhanced performance compared with human problem-solving. To realize the automated design, AutoOptLib provides 1) a rich library of metaheuristic components for continuous, discrete, and permutation problems; 2) a flexible algorithm representation for evolving diverse algorithm structures; 3) different design objectives and techniques for different optimization scenarios; and 4) a graphic user interface for accessibility and practicability. AutoOptLib is fully written in Matlab/Octave; its source code and documentation are available at https://github.com/qz89/AutoOpt and https://AutoOpt.readthedocs.io/, respectively.

Qiuhao Chen, Yuxuan Du, Qi Zhao et al.

Efficient quantum compiling tactics greatly enhance the capability of quantum computers to execute complicated quantum algorithms. Due to its fundamental importance, a plethora of quantum compilers has been designed in past years. However, there are several caveats to current protocols, which are low optimality, high inference time, limited scalability, and lack of universality. To compensate for these defects, here we devise an efficient and practical quantum compiler assisted by advanced deep reinforcement learning (RL) techniques, i.e., data generation, deep Q-learning, and AQ* search. In this way, our protocol is compatible with various quantum machines and can be used to compile multi-qubit operators. We systematically evaluate the performance of our proposal in compiling quantum operators with both inverse-closed and inverse-free universal basis sets. In the task of single-qubit operator compiling, our proposal outperforms other RL-based quantum compilers in the measure of compiling sequence length and inference time. Meanwhile, the output solution is near-optimal, guaranteed by the Solovay-Kitaev theorem. Notably, for the inverse-free universal basis set, the achieved sequence length complexity is comparable with the inverse-based setting and dramatically advances previous methods. These empirical results contribute to improving the inverse-free Solovay-Kitaev theorem. In addition, for the first time, we demonstrate how to leverage RL-based quantum compilers to accomplish two-qubit operator compiling. The achieved results open an avenue for integrating RL with quantum compiling to unify efficiency and practicality and thus facilitate the exploration of quantum advantages.

Qi Zhao, Jun Chen, Ivor Tsang et al.

While modern diffusion models excel at generating diverse single images, extending this to sequential generation reveals a fundamental challenge: balancing narrative dynamism with multi-character coherence. Existing methods often falter at this trade-off, leading to artifacts where characters lose their identity or the story stagnates. To resolve this critical tension, we introduce RealDiffusion, a unified framework designed to reconcile robust coherence with narrative dynamism. Heat diffusion serves as a dissipative prior that averages neighboring features along the sequence and removes high-frequency noise within the subject region. This suppresses attribute drift and stabilizes identity across frames. A region-aware stochastic process then introduces small perturbations that explore nearby modes and prevent collapse so the story maintains pose change and scene evolution. We thus introduce a lightweight, training-free Physics-informed Attention mechanism that injects controllable physical priors into the self-attention layers during inference. By modeling feature evolution as a configurable physical system, our method regularizes spatio-temporal relationships without suppressing intentional, prompt-driven changes. Extensive experiments demonstrate that RealDiffusion achieves substantial gains in character coherence while preserving narrative dynamism, outperforming state-of-the-art approaches. Code is available at https://github.com/ShmilyQi-CN/RealDiffusion.

Shijie Wang, Qi Zhao, Minh Quan Do et al.

What makes good representations for video understanding, such as anticipating future activities, or answering video-conditioned questions? While earlier approaches focus on end-to-end learning directly from video pixels, we propose to revisit text-based representations, such as general-purpose video captions, which are interpretable and can be directly consumed by large language models (LLMs). Intuitively, different video understanding tasks may require representations that are complementary and at different granularity. To this end, we propose versatile action models (Vamos), a learning framework powered by a large language model as the ``reasoner'', and can flexibly leverage visual embedding and free-form text descriptions as its input. To interpret the important text evidence for question answering, we generalize the concept bottleneck model to work with tokens and nonlinear models, which uses hard attention to select a small subset of tokens from the free-form text as inputs to the LLM reasoner. We evaluate Vamos on five complementary benchmarks, Ego4D, NeXT-QA, IntentQA, Spacewalk-18, and EgoSchema, on its capability to model temporal dynamics, encode visual history, and perform reasoning. Surprisingly, we observe that text-based representations consistently achieve competitive performance on all benchmarks, and that visual embeddings provide marginal or no performance improvement, demonstrating the effectiveness of text-based video representation in the LLM era. We also demonstrate that our token bottleneck model is able to select relevant evidence from free-form text, support test-time intervention, and achieves nearly 5 times inference speedup while keeping a competitive question answering performance. Code and models are publicly released at https://brown-palm.github.io/Vamos/

Xianyu Chen, Ming Jiang, Qi Zhao

While exploring visual scenes, humans' scanpaths are driven by their underlying attention processes. Understanding visual scanpaths is essential for various applications. Traditional scanpath models predict the where and when of gaze shifts without providing explanations, creating a gap in understanding the rationale behind fixations. To bridge this gap, we introduce GazeXplain, a novel study of visual scanpath prediction and explanation. This involves annotating natural-language explanations for fixations across eye-tracking datasets and proposing a general model with an attention-language decoder that jointly predicts scanpaths and generates explanations. It integrates a unique semantic alignment mechanism to enhance the consistency between fixations and explanations, alongside a cross-dataset co-training approach for generalization. These novelties present a comprehensive and adaptable solution for explainable human visual scanpath prediction. Extensive experiments on diverse eye-tracking datasets demonstrate the effectiveness of GazeXplain in both scanpath prediction and explanation, offering valuable insights into human visual attention and cognitive processes.

Rong Fu, Yiqing Lyu, Chunlei Meng et al.

Automatic generation of radiology reports seeks to reduce clinician workload while improving documentation consistency. Existing methods that adopt encoder-decoder or retrieval-augmented pipelines achieve progress in fluency but remain vulnerable to visual-linguistic biases, factual inconsistency, and lack of explicit multi-hop clinical reasoning. We present NeuroSymb-MRG, a unified framework that integrates NeuroSymbolic abductive reasoning with active uncertainty minimization to produce structured, clinically grounded reports. The system maps image features to probabilistic clinical concepts, composes differentiable logic-based reasoning chains, decodes those chains into templated clauses, and refines the textual output via retrieval and constrained language-model editing. An active sampling loop driven by rule-level uncertainty and diversity guides clinician-in-the-loop adjudication and promptbook refinement. Experiments on standard benchmarks demonstrate consistent improvements in factual consistency and standard language metrics compared to representative baselines.

Hang Ding, Qiming Feng, Dongqi Liu et al.

Reward modeling has become a cornerstone of aligning large language models (LLMs) with human preferences. Yet, when extended to subjective and open-ended domains such as role play, existing reward models exhibit severe degradation, struggling to capture nuanced and persona-grounded human judgments. To address this gap, we introduce RoleRMBench, the first systematic benchmark for reward modeling in role-playing dialogue, covering seven fine-grained capabilities from narrative management to role consistency and engagement. Evaluation on RoleRMBench reveals large and consistent gaps between general-purpose reward models and human judgment, particularly in narrative and stylistic dimensions. We further propose RoleRM, a reward model trained with Continuous Implicit Preferences (CIP), which reformulates subjective evaluation as continuous consistent pairwise supervision under multiple structuring strategies. Comprehensive experiments show that RoleRM surpasses strong open- and closed-source reward models by over 24% on average, demonstrating substantial gains in narrative coherence and stylistic fidelity. Our findings highlight the importance of continuous preference representation and annotation consistency, establishing a foundation for subjective alignment in human-centered dialogue systems.

Zhefan Xu, Ghassen Jerfel, Marina Haliem et al.

The rapid growth of autonomous driving datasets has enabled the scaling of powerful motion forecasting models. While large-scale pretraining provides strong performance, the standard imitation objective may not fully capture the complex nuances of human driving preferences. Meanwhile, recent advances in vision-language models (VLMs) have demonstrated impressive reasoning and commonsense understanding. Building on these capabilities, this paper presents VL-DPO, a vision-language-guided framework that aligns ego-vehicle motion forecasting models with human preferences. Our approach leverages a VLM as a zero-shot reasoner to automatically generate preference pairs from a pretrained model's rollouts, which are then used to finetune the model via Direct Preference Optimization (DPO). We finetune our models on the Waymo Open End-to-End Driving Dataset (WOD-E2E) and evaluate performance against held-out human preference annotations using rater feedback score (RFS) and average displacement error (ADE). Our experiments confirm that the VLM's trajectory selection is a high-quality proxy for human preference. Our final model, VL-DPO, yields an 11.94% increase in RFS and a 10.01% reduction in ADE over the pretrained model.

Menghao Li, Wenquan Feng, Shuchang Lyu et al.

Cell instance segmentation is a new and challenging task aiming at joint detection and segmentation of every cell in an image. Recently, many instance segmentation methods have applied in this task. Despite their great success, there still exists two main weaknesses caused by uncertainty of localizing cell center points. First, densely packed cells can easily be recognized into one cell. Second, elongated cell can easily be recognized into two cells. To overcome these two weaknesses, we propose a novel cell instance segmentation network based on multi-scheme regression guidance. With multi-scheme regression guidance, the network has the ability to look each cell in different views. Specifically, we first propose a gaussian guidance attention mechanism to use gaussian labels for guiding the network's attention. We then propose a point-regression module for assisting the regression of cell center. Finally, we utilize the output of the above two modules to further guide the instance segmentation. With multi-scheme regression guidance, we can take full advantage of the characteristics of different regions, especially the central region of the cell. We conduct extensive experiments on benchmark datasets, DSB2018, CA2.5 and SCIS. The encouraging results show that our network achieves SOTA (state-of-the-art) performance. On the DSB2018 and CA2.5, our network surpasses previous methods by 1.2% (AP50). Particularly on SCIS dataset, our network performs stronger by large margin (3.0% higher AP50). Visualization and analysis further prove that our proposed method is interpretable.

Safayat Bin Hakim, Aniqa Afzal, Qi Zhao et al.

Privacy-critical domains require phishing detection systems that satisfy contradictory constraints: near-zero false positives to prevent workflow disruption, transparent explanations for non-expert staff, strict regulatory compliance prohibiting sensitive data exposure to external APIs, and robustness against AI-generated attacks. Existing rule-based systems are brittle to novel campaigns, while LLM-based detectors violate privacy regulations through unredacted data transmission. We introduce CyberCane, a neuro-symbolic framework integrating deterministic symbolic analysis with privacy-preserving retrieval-augmented generation (RAG). Our dual-phase pipeline applies lightweight symbolic rules to email metadata, then escalates borderline cases to semantic classification via RAG with automated sensitive data redaction and retrieval from a phishing-only corpus. We further introduce PhishOnt, an OWL ontology enabling verifiable attack classification through formal reasoning chains. Evaluation on DataPhish2025 (12.3k emails; mixed human/LLM) and Nazario/SpamAssassin demonstrates a 78.6-point recall gain over symbolic-only detection on AI-generated threats, with precision exceeding 98% and FPR as low as 0.16%. Healthcare deployment projects a 542x ROI; tunable operating points support diverse risk tolerances, with open-source implementation at https://github.com/sbhakim/Cybercane.

Jue Wang, Yuxiang Lin, Qi Zhao et al.

The widespread use of various chemical gases in industrial processes necessitates effective measures to prevent their leakage during transportation and storage, given their high toxicity. Thermal infrared-based computer vision detection techniques provide a straightforward approach to identify gas leakage areas. However, the development of high-quality algorithms has been challenging due to the low texture in thermal images and the lack of open-source datasets. In this paper, we present the RGB-Thermal Cross Attention Network (RT-CAN), which employs an RGB-assisted two-stream network architecture to integrate texture information from RGB images and gas area information from thermal images. Additionally, to facilitate the research of invisible gas detection, we introduce Gas-DB, an extensive open-source gas detection database including about 1.3K well-annotated RGB-thermal images with eight variant collection scenes. Experimental results demonstrate that our method successfully leverages the advantages of both modalities, achieving state-of-the-art (SOTA) performance among RGB-thermal methods, surpassing single-stream SOTA models in terms of accuracy, Intersection of Union (IoU), and F2 metrics by 4.86%, 5.65%, and 4.88%, respectively. The code and data can be found at https://github.com/logic112358/RT-CAN.

Shaoxuan Cui, Qi Zhao, Guanlin Li et al.

This paper studies stabilization and its corresponding closed-loop region-of-attraction (ROA) for homogeneous polynomial dynamical systems whose nonlinear term admits an orthogonally decomposable (ODECO) tensor representation. While recent tensor-based results provide explicit solutions and sharp global characterizations for open-loop ODECO systems, closed-loop synthesis and computable ROA estimates are still often dominated by local linearization or Lyapunov/SOS (sum of squares) methods, which can be conservative and computationally demanding. We propose a structure-preserving linear feedback design that shares the ODECO eigenbasis of the system's tensor, thereby enabling closed-form trajectory expressions, explicit convergence/escape thresholds, and sharp ROA characterizations. Under mild conditions, we further derive robustness/ISS-type bounds for bounded disturbances. Numerical examples validate the theoretical results.

Qi Zhao, Shaoxuan Cui, Baolin Zhang et al.

Winner-take-all (WTA)--type selection is a fundamental mechanism in networked competition, yet its dependence on higher-order interactions remains insufficiently understood. We study a Lotka--Volterra competitive dynamics on higher-order networks, where classical pairwise inhibition is augmented by multi-way interaction terms induced by hyperedges of uniform hypergraphs. The proposed model shows multiple competitive outcomes, including WTA, winner-share-all (WSA), and variant winner-take-all (VWTA). The existence, uniqueness and stability of equilibria are rigorously proved through mathematical analysis, which relies on classical stability theory and recent advances in tensor algebra. We show that the eventual selection outcome is relatively insensitive to the hyperedge order and the specific higher-order coupling structure, and is instead determined by a small set of interpretable scalar parameters, such as the ratio between self-inhibition and lateral-inhibition and the external inputs. Numerical experiments support the theory by showing that higher-order interactions affect convergence and steady states, yet yield the similar outcome taxonomy (WTA/WSA/VWTA) as in standard graphs. These results provide a network-scientific explanation of the robustness of WTA-type outcomes under complex group interactions and offer principled guidance for designing selection mechanisms on higher-order networks.

Zhuoyuan Ma, Qi Zhao, Jin Zhang et al.

Aerial simultaneous transmitting and reflecting reconfigurable intelligent surfaces (STAR-RIS) enables full-space coverage in dynamic wireless networks. However, most existing works assume fixed user grouping, overlooking the fact that STAR-RIS deployment inherently determines whether users are served via transmission or reflection. To address this, we propose a joint deployment and beamforming framework, where an aerial STAR-RIS dynamically adjusts its location and orientation to adaptively control user grouping and enhance hybrid beamforming. We formulate a Markov decision process (MDP) capturing the coupling among deployment, grouping, and signal design. To solve the resulting non-convex and time-varying problem, we develop a PPO-based reinforcement learning algorithm that adaptively balances user grouping and beamforming resources through online policy learning. Simulation results show 57.1\% and 285\% sum-rate gains over fixed-deployment and RIS-free baselines, respectively, demonstrating the benefit of user-grouping-aware control in STAR-RIS-aided systems.

Yanting Miao, Yutao Sun, Dexin Wang et al.

Visual latent reasoning lets a multimodal large language model (MLLM) create intermediate visual evidence as continuous tokens, avoiding external tools or image generators. However, existing methods usually follow an output-as-input latent paradigm and yield unstable gains. We identify evidence for a feature-space mismatch that can contribute to this instability: dominant visual-latent models build on pre-norm MLLMs and reuse decoder hidden states as predicted latent inputs, even though these states occupy a substantially different norm regime from the input embeddings the model was trained to consume~\citep{xie2025mhc,li2026siamesenorm,team2026attention}. This mismatch can make direct latent feedback unreliable. Motivated by this diagnosis, we propose \textbf{GAP}, a \textbf{G}ranular \textbf{A}lignment \textbf{P}aradigm for visual latent modeling. GAP aligns visual latent reasoning at three levels: feature-level alignment maps decoder outputs into input-compatible visual latents through a lightweight PCA-aligned latent head; context-level alignment grounds latent targets with inspectable auxiliary visual supervision; and capacity-guided alignment assigns latent supervision selectively to examples where the base MLLM struggles. On Qwen2.5-VL 7B, the resulting model achieves the best mean aggregate perception and reasoning performance among our supervised variants. Inference-time intervention probing further suggests that generated latents provide task-relevant visual signal beyond merely adding token slots.

Junli Liang, Pengfei Zhou, Wangqiu Zhou et al.

Traditional Retrieval-Augmented Generation (RAG) effectively supports single-hop question answering with large language models but faces significant limitations in multi-hop question answering tasks, which require combining evidence from multiple documents. Existing chunk-based retrieval often provides irrelevant and logically incoherent context, leading to incomplete evidence chains and incorrect reasoning during answer generation. To address these challenges, we propose SentGraph, a sentence-level graph-based RAG framework that explicitly models fine-grained logical relationships between sentences for multi-hop question answering. Specifically, we construct a hierarchical sentence graph offline by first adapting Rhetorical Structure Theory to distinguish nucleus and satellite sentences, and then organizing them into topic-level subgraphs with cross-document entity bridges. During online retrieval, SentGraph performs graph-guided evidence selection and path expansion to retrieve fine-grained sentence-level evidence. Extensive experiments on four multi-hop question answering benchmarks demonstrate the effectiveness of SentGraph, validating the importance of explicitly modeling sentence-level logical dependencies for multi-hop reasoning.

Qi Zhao, Hongyu Yang, Qi Song et al.

Large language models (LLMs) have demonstrated remarkable capabilities in various complex tasks, yet they still suffer from hallucinations. By incorporating and exploring external knowledge, such as knowledge graphs(KGs), LLM's ability to provide factual answers has been enhanced. This approach carries significant practical implications. However, existing methods suffer from three key limitations: insufficient mining of LLMs' internal knowledge, constrained generation of interpretable reasoning paths, and unclear fusion of internal and external knowledge. Therefore, we propose KnowPath, a knowledge-enhanced large model framework driven by the collaboration of internal and external knowledge. It relies on the internal knowledge of the LLM to guide the exploration of interpretable directed subgraphs in external knowledge graphs, better integrating the two knowledge sources for more accurate reasoning. Extensive experiments on multiple real-world datasets demonstrate the effectiveness of KnowPath. Our code and data are available at https://github.com/tize-72/KnowPath.

Can Lin, Zhengwang Jiang, Ling Zheng et al.

Knowledge graph question answering (KGQA) aims to answer natural language questions using knowledge graphs. Recent research leverages large language models (LLMs) to enhance KGQA reasoning, but faces limitations: retrieval-based methods are constrained by the quality of retrieved information, while agent-based methods rely heavily on proprietary LLMs. To address these limitations, we propose Retrieval-Judgment-Exploration (RJE), a framework that retrieves refined reasoning paths, evaluates their sufficiency, and conditionally explores additional evidence. Moreover, RJE introduces specialized auxiliary modules enabling small-sized LLMs to perform effectively: Reasoning Path Ranking, Question Decomposition, and Retriever-assisted Exploration. Experiments show that our approach with proprietary LLMs (such as GPT-4o-mini) outperforms existing baselines while enabling small open-source LLMs (such as 3B and 8B parameters) to achieve competitive results without fine-tuning LLMs. Additionally, RJE substantially reduces the number of LLM calls and token usage compared to agent-based methods, yielding significant efficiency improvements.

Shuchang Lyu, Qi Zhao, Guangliang Cheng et al.

Unsupervised Domain Adaptation for Remote Sensing Semantic Segmentation (UDA-RSSeg) addresses the challenge of adapting a model trained on source domain data to target domain samples, thereby minimizing the need for annotated data across diverse remote sensing scenes. This task presents two principal challenges: (1) severe inconsistencies in feature representation across different remote sensing domains, and (2) a domain gap that emerges due to the representation bias of source domain patterns when translating features to predictive logits. To tackle these issues, we propose a joint-optimized adversarial network incorporating the "Segment Anything Model (SAM) (SAM-JOANet)" for UDA-RSSeg. Our approach integrates SAM to leverage its robust generalized representation capabilities, thereby alleviating feature inconsistencies. We introduce a finetuning decoder designed to convert SAM-Encoder features into predictive logits. Additionally, a feature-level adversarial-based prompted segmentor is employed to generate class-agnostic maps, which guide the finetuning decoder's feature representations. The network is optimized end-to-end, combining the prompted segmentor and the finetuning decoder. Extensive evaluations on benchmark datasets, including ISPRS (Potsdam/Vaihingen) and CITY-OSM (Paris/Chicago), demonstrate the effectiveness of our method. The results, supported by visualization and analysis, confirm the method's interpretability and robustness. The code of this paper is available at https://github.com/CV-ShuchangLyu/SAM-JOANet.

You Zhou, Lijiang Chen, Shuchang Lyu et al.

Federated learning enables collaborative training of machine learning models among different clients while ensuring data privacy, emerging as the mainstream for breaking data silos in the healthcare domain. However, the imbalance of medical resources, data corruption or improper data preservation may lead to a situation where different clients possess medical images of different modality. This heterogeneity poses a significant challenge for cross-domain medical image segmentation within the federated learning framework. To address this challenge, we propose a new Federated Domain Adaptation (FedDA) segmentation training framework. Specifically, we propose a feature-level adversarial learning among clients by aligning feature maps across clients through embedding an adversarial training mechanism. This design can enhance the model's generalization on multiple domains and alleviate the negative impact from domain-shift. Comprehensive experiments on three medical image datasets demonstrate that our proposed FedDA substantially achieves cross-domain federated aggregation, endowing single modality client with cross-modality processing capabilities, and consistently delivers robust performance compared to state-of-the-art federated aggregation algorithms in objective and subjective assessment. Our code are available at https://github.com/GGbond-study/FedDA.

You Zhou, Lijiang Chen, Guangxia Cui et al.

Ovarian tumor, as a common gynecological disease, can rapidly deteriorate into serious health crises when undetected early, thus posing significant threats to the health of women. Deep neural networks have the potential to identify ovarian tumors, thereby reducing mortality rates, but limited public datasets hinder its progress. To address this gap, we introduce a vital ovarian tumor pathological recognition dataset called \textbf{ViTaL} that contains \textbf{V}isual, \textbf{T}abular and \textbf{L}inguistic modality data of 496 patients across six pathological categories. The ViTaL dataset comprises three subsets corresponding to different patient data modalities: visual data from 2216 two-dimensional ultrasound images, tabular data from medical examinations of 496 patients, and linguistic data from ultrasound reports of 496 patients. It is insufficient to merely distinguish between benign and malignant ovarian tumors in clinical practice. To enable multi-pathology classification of ovarian tumor, we propose a ViTaL-Net based on the Triplet Hierarchical Offset Attention Mechanism (THOAM) to minimize the loss incurred during feature fusion of multi-modal data. This mechanism could effectively enhance the relevance and complementarity between information from different modalities. ViTaL-Net serves as a benchmark for the task of multi-pathology, multi-modality classification of ovarian tumors. In our comprehensive experiments, the proposed method exhibited satisfactory performance, achieving accuracies exceeding 90\% on the two most common pathological types of ovarian tumor and an overall performance of 85\%. Our dataset and code are available at https://github.com/GGbond-study/vitalnet.

Xian-Xian Liu, Yuanyuan Wei, Mingkun Xu et al.

Early detection of gastric cancer, a leading cause of cancer-related mortality worldwide, remains hampered by the limitations of current diagnostic technologies, leading to high rates of misdiagnosis and missed diagnoses. To address these challenges, we propose an integrated system that synergizes advanced hardware and software technologies to balance speed-accuracy. Our study introduces the One Class Twin Cross Learning (OCT-X) algorithm. Leveraging a novel fast double-threshold grid search strategy (FDT-GS) and a patch-based deep fully convolutional network, OCT-X maximizes diagnostic accuracy through real-time data processing and seamless lesion surveillance. The hardware component includes an all-in-one point-of-care testing (POCT) device with high-resolution imaging sensors, real-time data processing, and wireless connectivity, facilitated by the NI CompactDAQ and LabVIEW software. Our integrated system achieved an unprecedented diagnostic accuracy of 99.70%, significantly outperforming existing models by up to 4.47%, and demonstrated a 10% improvement in multirate adaptability. These findings underscore the potential of OCT-X as well as the integrated system in clinical diagnostics, offering a path toward more accurate, efficient, and less invasive early gastric cancer detection. Future research will explore broader applications, further advancing oncological diagnostics. Code is available at https://github.com/liu37972/Multirate-Location-on-OCT-X-Learning.git.

Muyi Bao, Shuchang Lyu, Zhaoyang Xu et al.

Skin lesion segmentation is a critical challenge in computer vision, and it is essential to separate pathological features from healthy skin for diagnostics accurately. Traditional Convolutional Neural Networks (CNNs) are limited by narrow receptive fields, and Transformers face significant computational burdens. This paper presents a novel skin lesion segmentation framework, the Atrous Shifted Parallel Vision Mamba UNet (ASP-VMUNet), which integrates the efficient and scalable Mamba architecture to overcome limitations in traditional CNNs and computationally demanding Transformers. The framework introduces an atrous scan technique that minimizes background interference and expands the receptive field, enhancing Mamba's scanning capabilities. Additionally, the inclusion of a Parallel Vision Mamba (PVM) layer and a shift round operation optimizes feature segmentation and fosters rich inter-segment information exchange. A supplementary CNN branch with a Selective-Kernel (SK) Block further refines the segmentation by blending local and global contextual information. Tested on four benchmark datasets (ISIC16/17/18 and PH2), ASP-VMUNet demonstrates superior performance in skin lesion segmentation, validated by comprehensive ablation studies. This approach not only advances medical image segmentation but also highlights the benefits of hybrid architectures in medical imaging technology. Our code is available at https://github.com/BaoBao0926/ASP-VMUNet/tree/main.

Xinjie Sun, Boxiong Wei, Yalong Jiang et al.

Thyroid nodule segmentation in ultrasound images is crucial for accurate diagnosis and treatment planning. However, existing methods face challenges in segmentation accuracy, interpretability, and generalization, which hinder their performance. This letter proposes a novel framework, CLIP-TNseg, to address these issues by integrating a multimodal large model with a neural network architecture. CLIP-TNseg consists of two main branches: the Coarse-grained Branch, which extracts high-level semantic features from a frozen CLIP model, and the Fine-grained Branch, which captures fine-grained features using U-Net style residual blocks. These features are fused and processed by the prediction head to generate precise segmentation maps. CLIP-TNseg leverages the Coarse-grained Branch to enhance semantic understanding through textual and high-level visual features, while the Fine-grained Branch refines spatial details, enabling precise and robust segmentation. Extensive experiments on public and our newly collected datasets demonstrate its competitive performance. Our code and the original dataset are available at https://github.com/jayxjsun/CLIP-TNseg.

Yan Luo, Yongkang Wong, Mohan Kankanhalli et al.

A key challenge for machine intelligence is to learn new visual concepts without forgetting the previously acquired knowledge. Continual learning is aimed towards addressing this challenge. However, there is a gap between existing supervised continual learning and human-like intelligence, where human is able to learn from both labeled and unlabeled data. How unlabeled data affects learning and catastrophic forgetting in the continual learning process remains unknown. To explore these issues, we formulate a new semi-supervised continual learning method, which can be generically applied to existing continual learning models. Specifically, a novel gradient learner learns from labeled data to predict gradients on unlabeled data. Hence, the unlabeled data could fit into the supervised continual learning method. Different from conventional semi-supervised settings, we do not hypothesize that the underlying classes, which are associated to the unlabeled data, are known to the learning process. In other words, the unlabeled data could be very distinct from the labeled data. We evaluate the proposed method on mainstream continual learning, adversarial continual learning, and semi-supervised learning tasks. The proposed method achieves state-of-the-art performance on classification accuracy and backward transfer in the continual learning setting while achieving desired performance on classification accuracy in the semi-supervised learning setting. This implies that the unlabeled images can enhance the generalizability of continual learning models on the predictive ability on unseen data and significantly alleviate catastrophic forgetting. The code is available at \url{https://github.com/luoyan407/grad_prediction.git}.

Yan Luo, Yongkang Wong, Mohan S. Kankanhalli et al.

The learning process of deep learning methods usually updates the model's parameters in multiple iterations. Each iteration can be viewed as the first-order approximation of Taylor's series expansion. The remainder, which consists of higher-order terms, is usually ignored in the learning process for simplicity. This learning scheme empowers various multimedia based applications, such as image retrieval, recommendation system, and video search. Generally, multimedia data (e.g., images) are semantics-rich and high-dimensional, hence the remainders of approximations are possibly non-zero. In this work, we consider the remainder to be informative and study how it affects the learning process. To this end, we propose a new learning approach, namely gradient adjustment learning (GAL), to leverage the knowledge learned from the past training iterations to adjust vanilla gradients, such that the remainders are minimized and the approximations are improved. The proposed GAL is model- and optimizer-agnostic, and is easy to adapt to the standard learning framework. It is evaluated on three tasks, i.e., image classification, object detection, and regression, with state-of-the-art models and optimizers. The experiments show that the proposed GAL consistently enhances the evaluated models, whereas the ablation studies validate various aspects of the proposed GAL. The code is available at \url{https://github.com/luoyan407/gradient_adjustment.git}.

Yan Luo, Yongkang Wong, Mohan S. Kankanhalli et al.

Understanding the trustworthiness of a prediction yielded by a classifier is critical for the safe and effective use of AI models. Prior efforts have been proven to be reliable on small-scale datasets. In this work, we study the problem of predicting trustworthiness on real-world large-scale datasets, where the task is more challenging due to high-dimensional features, diverse visual concepts, and large-scale samples. In such a setting, we observe that the trustworthiness predictors trained with prior-art loss functions, i.e., the cross entropy loss, focal loss, and true class probability confidence loss, are prone to view both correct predictions and incorrect predictions to be trustworthy. The reasons are two-fold. Firstly, correct predictions are generally dominant over incorrect predictions. Secondly, due to the data complexity, it is challenging to differentiate the incorrect predictions from the correct ones on real-world large-scale datasets. To improve the generalizability of trustworthiness predictors, we propose a novel steep slope loss to separate the features w.r.t. correct predictions from the ones w.r.t. incorrect predictions by two slide-like curves that oppose each other. The proposed loss is evaluated with two representative deep learning models, i.e., Vision Transformer and ResNet, as trustworthiness predictors. We conduct comprehensive experiments and analyses on ImageNet, which show that the proposed loss effectively improves the generalizability of trustworthiness predictors. The code and pre-trained trustworthiness predictors for reproducibility are available at https://github.com/luoyan407/predict_trustworthiness.

Qi Zhao, Binghao Liu, Shuchang Lyu et al.

Few-shot segmentation focuses on the generalization of models to segment unseen object with limited annotated samples. However, existing approaches still face two main challenges. First, huge feature distinction between support and query images causes knowledge transferring barrier, which harms the segmentation performance. Second, limited support prototypes cannot adequately represent features of support objects, hard to guide high-quality query segmentation. To deal with the above two issues, we propose self-distillation embedded supervised affinity attention model to improve the performance of few-shot segmentation task. Specifically, the self-distillation guided prototype module uses self-distillation to align the features of support and query. The supervised affinity attention module generates high-quality query attention map to provide sufficient object information. Extensive experiments prove that our model significantly improves the performance compared to existing methods. Comprehensive ablation experiments and visualization studies also show the significant effect of our method on few-shot segmentation task. On COCO-20i dataset, we achieve new state-of-the-art results. Training code and pretrained models are available at https://github.com/cv516Buaa/SD-AANet.

Shi Chen, Ming Jiang, Jinhui Yang et al.

While attention has been an increasingly popular component in deep neural networks to both interpret and boost performance of models, little work has examined how attention progresses to accomplish a task and whether it is reasonable. In this work, we propose an Attention with Reasoning capability (AiR) framework that uses attention to understand and improve the process leading to task outcomes. We first define an evaluation metric based on a sequence of atomic reasoning operations, enabling quantitative measurement of attention that considers the reasoning process. We then collect human eye-tracking and answer correctness data, and analyze various machine and human attentions on their reasoning capability and how they impact task performance. Furthermore, we propose a supervision method to jointly and progressively optimize attention, reasoning, and task performance so that models learn to look at regions of interests by following a reasoning process. We demonstrate the effectiveness of the proposed framework in analyzing and modeling attention with better reasoning capability and task performance. The code and data are available at https://github.com/szzexpoi/AiR

Xianyu Chen, Ming Jiang, Qi Zhao

Few-shot object detection aims at detecting objects with few annotated examples, which remains a challenging research problem yet to be explored. Recent studies have shown the effectiveness of self-learned top-down attention mechanisms in object detection and other vision tasks. The top-down attention, however, is less effective at improving the performance of few-shot detectors. Due to the insufficient training data, object detectors cannot effectively generate attention maps for few-shot examples. To improve the performance and interpretability of few-shot object detectors, we propose an attentive few-shot object detection network (AttFDNet) that takes the advantages of both top-down and bottom-up attention. Being task-agnostic, the bottom-up attention serves as a prior that helps detect and localize naturally salient objects. We further address specific challenges in few-shot object detection by introducing two novel loss terms and a hybrid few-shot learning strategy. Experimental results and visualization demonstrate the complementary nature of the two types of attention and their roles in few-shot object detection. Codes are available at https://github.com/chenxy99/AttFDNet.

Yan Luo, Yongkang Wong, Mohan S. Kankanhalli et al.

Benefiting from deep learning research and large-scale datasets, saliency prediction has achieved significant success in the past decade. However, it still remains challenging to predict saliency maps on images in new domains that lack sufficient data for data-hungry models. To solve this problem, we propose a few-shot transfer learning paradigm for saliency prediction, which enables efficient transfer of knowledge learned from the existing large-scale saliency datasets to a target domain with limited labeled examples. Specifically, very few target domain examples are used as the reference to train a model with a source domain dataset such that the training process can converge to a local minimum in favor of the target domain. Then, the learned model is further fine-tuned with the reference. The proposed framework is gradient-based and model-agnostic. We conduct comprehensive experiments and ablation study on various source domain and target domain pairs. The results show that the proposed framework achieves a significant performance improvement. The code is publicly available at \url{https://github.com/luoyan407/n-reference}.

Yan Luo, Yongkang Wong, Mohan S. Kankanhalli et al.

One of the well-known challenges in computer vision tasks is the visual diversity of images, which could result in an agreement or disagreement between the learned knowledge and the visual content exhibited by the current observation. In this work, we first define such an agreement in a concepts learning process as congruency. Formally, given a particular task and sufficiently large dataset, the congruency issue occurs in the learning process whereby the task-specific semantics in the training data are highly varying. We propose a Direction Concentration Learning (DCL) method to improve congruency in the learning process, where enhancing congruency influences the convergence path to be less circuitous. The experimental results show that the proposed DCL method generalizes to state-of-the-art models and optimizers, as well as improves the performances of saliency prediction task, continual learning task, and classification task. Moreover, it helps mitigate the catastrophic forgetting problem in the continual learning task. The code is publicly available at https://github.com/luoyan407/congruency.