Hui Li

Hui Li, Tianyang Xu, Xiao-Jun Wu et al.

Deep learning based fusion methods have been achieving promising performance in image fusion tasks. This is attributed to the network architecture that plays a very important role in the fusion process. However, in general, it is hard to specify a good fusion architecture, and consequently, the design of fusion networks is still a black art, rather than science. To address this problem, we formulate the fusion task mathematically, and establish a connection between its optimal solution and the network architecture that can implement it. This approach leads to a novel method proposed in the paper of constructing a lightweight fusion network. It avoids the time-consuming empirical network design by a trial-and-test strategy. In particular we adopt a learnable representation approach to the fusion task, in which the construction of the fusion network architecture is guided by the optimisation algorithm producing the learnable model. The low-rank representation (LRR) objective is the foundation of our learnable model. The matrix multiplications, which are at the heart of the solution are transformed into convolutional operations, and the iterative process of optimisation is replaced by a special feed-forward network. Based on this novel network architecture, an end-to-end lightweight fusion network is constructed to fuse infrared and visible light images. Its successful training is facilitated by a detail-to-semantic information loss function proposed to preserve the image details and to enhance the salient features of the source images. Our experiments show that the proposed fusion network exhibits better fusion performance than the state-of-the-art fusion methods on public datasets. Interestingly, our network requires a fewer training parameters than other existing methods. The codes are available at https://github.com/hli1221/imagefusion-LRRNet

Aaron Blakeman, Aarti Basant, Abhinav Khattar et al. · nvidia

As inference-time scaling becomes critical for enhanced reasoning capabilities, it is increasingly becoming important to build models that are efficient to infer. We introduce Nemotron-H, a family of 8B and 56B/47B hybrid Mamba-Transformer models designed to reduce inference cost for a given accuracy level. To achieve this goal, we replace the majority of self-attention layers in the common Transformer model architecture with Mamba layers that perform constant computation and require constant memory per generated token. We show that Nemotron-H models offer either better or on-par accuracy compared to other similarly-sized state-of-the-art open-sourced Transformer models (e.g., Qwen-2.5-7B/72B and Llama-3.1-8B/70B), while being up to 3$\times$ faster at inference. To further increase inference speed and reduce the memory required at inference time, we created Nemotron-H-47B-Base from the 56B model using a new compression via pruning and distillation technique called MiniPuzzle. Nemotron-H-47B-Base achieves similar accuracy to the 56B model, but is 20% faster to infer. In addition, we introduce an FP8-based training recipe and show that it can achieve on par results with BF16-based training. This recipe is used to train the 56B model. We are releasing Nemotron-H base model checkpoints with support in Hugging Face and NeMo.

Xiaomin Fang, Fan Wang, Lihang Liu et al. · baidu

AI-based protein structure prediction pipelines, such as AlphaFold2, have achieved near-experimental accuracy. These advanced pipelines mainly rely on Multiple Sequence Alignments (MSAs) as inputs to learn the co-evolution information from the homologous sequences. Nonetheless, searching MSAs from protein databases is time-consuming, usually taking dozens of minutes. Consequently, we attempt to explore the limits of fast protein structure prediction by using only primary sequences of proteins. HelixFold-Single is proposed to combine a large-scale protein language model with the superior geometric learning capability of AlphaFold2. Our proposed method, HelixFold-Single, first pre-trains a large-scale protein language model (PLM) with thousands of millions of primary sequences utilizing the self-supervised learning paradigm, which will be used as an alternative to MSAs for learning the co-evolution information. Then, by combining the pre-trained PLM and the essential components of AlphaFold2, we obtain an end-to-end differentiable model to predict the 3D coordinates of atoms from only the primary sequence. HelixFold-Single is validated in datasets CASP14 and CAMEO, achieving competitive accuracy with the MSA-based methods on the targets with large homologous families. Furthermore, HelixFold-Single consumes much less time than the mainstream pipelines for protein structure prediction, demonstrating its potential in tasks requiring many predictions. The code of HelixFold-Single is available at https://github.com/PaddlePaddle/PaddleHelix/tree/dev/apps/protein_folding/helixfold-single, and we also provide stable web services on https://paddlehelix.baidu.com/app/drug/protein-single/forecast.

Hui Li, Yangfan Gao, Junlin Shang et al.

Audio tokenizers serve as the discrete interface between continuous audio and Audio Language Models (ALMs), but existing tokenizers often struggle to support both understanding and generation. Reconstruction-oriented codecs preserve acoustic fidelity but lack rich semantics, while semantic-aware tokenizers typically rely on separate semantic and acoustic streams, introducing redundancy or misalignment. We propose \textbf{EntangleCodec}, a unified discrete audio tokenizer that learns caption-aligned semantic-acoustic representations before quantization. By aligning audio with rich captions rather than ASR transcripts, EntangleCodec captures linguistic content, speaker identity, emotion, prosody, and acoustic scenes within a compact token stream. A flow-matching diffusion decoder further enables high-quality reconstruction across speech, music, and general audio. EntangleCodec achieves reconstruction quality competitive with specialized codecs, outperforms all codec-based baselines on audio understanding by up to \textbf{+7.4\%} on MMAR, and supports both TTS and TTA generation in a unified framework. Furthermore, EntangleCodec-based audio language models demonstrate strong scaling behavior: even at \textit{0.6B} parameters, the model surpasses specialized continuous-representation LLMs with over \textit{13B} parameters across three benchmarks using \textbf{22$\times$} fewer parameters; scaling to \textit{8B} further establishes new state-of-the-art results on MMAR, highlighting that representation quality is as critical as model scale in audio language modeling. Code and model weights are available at https://github.com/luckyerr/EntangleCodec.

Ziyao Guo, Haonan Yan, Hui Li et al.

Previous knowledge distillation methods have shown their impressive performance on model compression tasks, however, it is hard to explain how the knowledge they transferred helps to improve the performance of the student network. In this work, we focus on proposing a knowledge distillation method that has both high interpretability and competitive performance. We first revisit the structure of mainstream CNN models and reveal that possessing the capacity of identifying class discriminative regions of input is critical for CNN to perform classification. Furthermore, we demonstrate that this capacity can be obtained and enhanced by transferring class activation maps. Based on our findings, we propose class attention transfer based knowledge distillation (CAT-KD). Different from previous KD methods, we explore and present several properties of the knowledge transferred by our method, which not only improve the interpretability of CAT-KD but also contribute to a better understanding of CNN. While having high interpretability, CAT-KD achieves state-of-the-art performance on multiple benchmarks. Code is available at: https://github.com/GzyAftermath/CAT-KD.

Yinsong Xu, Wei Jing, Liuxin Zhang et al.

Understanding long-form egocentric videos remains challenging for multimodal large language models (MLLMs) due to limited context length and insufficient grounding of fine-grained visual details. The recently proposed HD-EPIC benchmark highlights these limitations: even strong long-context models achieve relatively low performance across diverse video question answering tasks. In this paper, we propose a unified framework that decouples long-video reasoning into two complementary forms of evidence: semantic evidence and visual evidence. Semantic evidence captures global procedural structure through a coarse-to-fine extraction pipeline, while object-centric visual evidence preserves fine-grained grounding through bounding boxes and visual embeddings. During inference, we formulate reasoning as a query-conditioned evidence retrieval and integration process, dynamically selecting relevant information from both sources. Our approach achieves competitive performance in the HD-EPIC-VQA Challenge across multiple task categories. More broadly, our results demonstrate that explicitly structuring, retrieving, and integrating semantic and visual evidence is critical for effective long-video understanding with MLLMs.

Jianyang Zhai, Xiawu Zheng, Chang-Dong Wang et al.

Pre-trained language models (PLMs) have demonstrated strong performance in sequential recommendation (SR), which are utilized to extract general knowledge. However, existing methods still lack domain knowledge and struggle to capture users' fine-grained preferences. Meanwhile, many traditional SR methods improve this issue by integrating side information while suffering from information loss. To summarize, we believe that a good recommendation system should utilize both general and domain knowledge simultaneously. Therefore, we introduce an external knowledge base and propose Knowledge Prompt-tuning for Sequential Recommendation (\textbf{KP4SR}). Specifically, we construct a set of relationship templates and transform a structured knowledge graph (KG) into knowledge prompts to solve the problem of the semantic gap. However, knowledge prompts disrupt the original data structure and introduce a significant amount of noise. We further construct a knowledge tree and propose a knowledge tree mask, which restores the data structure in a mask matrix form, thus mitigating the noise problem. We evaluate KP4SR on three real-world datasets, and experimental results show that our approach outperforms state-of-the-art methods on multiple evaluation metrics. Specifically, compared with PLM-based methods, our method improves NDCG@5 and HR@5 by \textcolor{red}{40.65\%} and \textcolor{red}{36.42\%} on the books dataset, \textcolor{red}{11.17\%} and \textcolor{red}{11.47\%} on the music dataset, and \textcolor{red}{22.17\%} and \textcolor{red}{19.14\%} on the movies dataset, respectively. Our code is publicly available at the link: \href{https://github.com/zhaijianyang/KP4SR}{\textcolor{blue}{https://github.com/zhaijianyang/KP4SR}.}

Yanlin Wang, Wanjun Zhong, Yanxian Huang et al.

In recent years, Large Language Models (LLMs) have achieved remarkable success and have been widely used in various downstream tasks, especially in the tasks of the software engineering (SE) field. We find that many studies combining LLMs with SE have employed the concept of agents either explicitly or implicitly. However, there is a lack of an in-depth survey to sort out the development context of existing works, analyze how existing works combine the LLM-based agent technologies to optimize various tasks, and clarify the framework of LLM-based agents in SE. In this paper, we conduct the first survey of the studies on combining LLM-based agents with SE and present a framework of LLM-based agents in SE which includes three key modules: perception, memory, and action. We also summarize the current challenges in combining the two fields and propose future opportunities in response to existing challenges. We maintain a GitHub repository of the related papers at: https://github.com/DeepSoftwareAnalytics/Awesome-Agent4SE.

Meifang Zeng, Ke Li, Bingchuan Jiang et al.

In shilling attacks, an adversarial party injects a few fake user profiles into a Recommender System (RS) so that the target item can be promoted or demoted. Although much effort has been devoted to developing shilling attack methods, we find that existing approaches are still far from practical. In this paper, we analyze the properties a practical shilling attack method should have and propose a new concept of Cross-system Attack. With the idea of Cross-system Attack, we design a Practical Cross-system Shilling Attack (PC-Attack) framework that requires little information about the victim RS model and the target RS data for conducting attacks. PC-Attack is trained to capture graph topology knowledge from public RS data in a self-supervised manner. Then, it is fine-tuned on a small portion of target data that is easy to access to construct fake profiles. Extensive experiments have demonstrated the superiority of PC-Attack over state-of-the-art baselines. Our implementation of PC-Attack is available at https://github.com/KDEGroup/PC-Attack.

Zhishe Wang, Yanlin Chen, Wenyu Shao et al.

The existing deep learning fusion methods mainly concentrate on the convolutional neural networks, and few attempts are made with transformer. Meanwhile, the convolutional operation is a content-independent interaction between the image and convolution kernel, which may lose some important contexts and further limit fusion performance. Towards this end, we present a simple and strong fusion baseline for infrared and visible images, namely\textit{ Residual Swin Transformer Fusion Network}, termed as SwinFuse. Our SwinFuse includes three parts: the global feature extraction, fusion layer and feature reconstruction. In particular, we build a fully attentional feature encoding backbone to model the long-range dependency, which is a pure transformer network and has a stronger representation ability compared with the convolutional neural networks. Moreover, we design a novel feature fusion strategy based on $L_{1}$-norm for sequence matrices, and measure the corresponding activity levels from row and column vector dimensions, which can well retain competitive infrared brightness and distinct visible details. Finally, we testify our SwinFuse with nine state-of-the-art traditional and deep learning methods on three different datasets through subjective observations and objective comparisons, and the experimental results manifest that the proposed SwinFuse obtains surprising fusion performance with strong generalization ability and competitive computational efficiency. The code will be available at https://github.com/Zhishe-Wang/SwinFuse.

Jingkai Wang, Jue Gong, Zheng Chen et al.

This paper provides a review of the NTIRE 2026 challenge on real-world face restoration, highlighting the proposed solutions and the resulting outcomes. The challenge focuses on generating natural and realistic outputs while maintaining identity consistency. Its goal is to advance state-of-the-art solutions for perceptual quality and realism, without imposing constraints on computational resources or training data. Performance is evaluated using a weighted image quality assessment (IQA) score and employs the AdaFace model as an identity checker. The competition attracted 96 registrants, with 10 teams submitting valid models; ultimately, 9 teams achieved valid scores in the final ranking. This collaborative effort advances the performance of real-world face restoration while offering an in-depth overview of the latest trends in the field.

Hui Li, Jinghan Jia, Shijun Liang et al.

Although deep learning (DL) has gained much popularity for accelerated magnetic resonance imaging (MRI), recent studies have shown that DL-based MRI reconstruction models could be oversensitive to tiny input perturbations (that are called 'adversarial perturbations'), which cause unstable, low-quality reconstructed images. This raises the question of how to design robust DL methods for MRI reconstruction. To address this problem, we propose a novel image reconstruction framework, termed SMOOTHED UNROLLING (SMUG), which advances a deep unrolling-based MRI reconstruction model using a randomized smoothing (RS)-based robust learning operation. RS, which improves the tolerance of a model against input noises, has been widely used in the design of adversarial defense for image classification. Yet, we find that the conventional design that applies RS to the entire DL process is ineffective for MRI reconstruction. We show that SMUG addresses the above issue by customizing the RS operation based on the unrolling architecture of the DL-based MRI reconstruction model. Compared to the vanilla RS approach and several variants of SMUG, we show that SMUG improves the robustness of MRI reconstruction with respect to a diverse set of perturbation sources, including perturbations to the input measurements, different measurement sampling rates, and different unrolling steps. Code for SMUG will be available at https://github.com/LGM70/SMUG.

Chen Lin, Si Chen, Meifang Zeng et al.

Due to the pivotal role of Recommender Systems (RS) in guiding customers towards the purchase, there is a natural motivation for unscrupulous parties to spoof RS for profits. In this paper, we study Shilling Attack where an adversarial party injects a number of fake user profiles for improper purposes. Conventional Shilling Attack approaches lack attack transferability (i.e., attacks are not effective on some victim RS models) and/or attack invisibility (i.e., injected profiles can be easily detected). To overcome these issues, we present Leg-UP, a novel attack model based on the Generative Adversarial Network. Leg-UP learns user behavior patterns from real users in the sampled ``templates'' and constructs fake user profiles. To simulate real users, the generator in Leg-UP directly outputs discrete ratings. To enhance attack transferability, the parameters of the generator are optimized by maximizing the attack performance on a surrogate RS model. To improve attack invisibility, Leg-UP adopts a discriminator to guide the generator to generate undetectable fake user profiles. Experiments on benchmarks have shown that Leg-UP exceeds state-of-the-art Shilling Attack methods on a wide range of victim RS models. The source code of our work is available at: https://github.com/XMUDM/ShillingAttack.

Jin Liang, Yuchen Yang, Anran Zhang et al.

The photographs captured by digital cameras usually suffer from over or under exposure problems. For image exposure enhancement, the tasks of Single-Exposure Correction (SEC) and Multi-Exposure Fusion (MEF) are widely studied in the image processing community. However, current SEC or MEF methods are developed under different motivations and thus ignore the internal correlation between SEC and MEF, making it difficult to process arbitrary-length sequences with improper exposures. Besides, the MEF methods usually fail at estimating the exposure of a sequence containing only under-exposed or over-exposed images. To alleviate these problems, in this paper, we develop a novel Fusion-Correction Network (FCNet) to tackle an arbitrary-length (including one) image sequence with improper exposures. This is achieved by fusing and correcting an image sequence by Laplacian Pyramid (LP) image decomposition. In each LP level, the low-frequency base component of the input image sequence is fed into a Fusion block and a Correction block sequentially for consecutive exposure estimation, implemented by alternative exposure fusion and correction. The exposure-corrected image in current LP level is upsampled and fused with the high-frequency detail components of the input image sequence in the next LP level, to output the base component for the Fusion and Correction blocks in next LP level. Experiments on the benchmark dataset demonstrate that our FCNet is effective on arbitrary-length exposure estimation, including both SEC and MEF. The code is publicly released at https://github.com/NKUJinLiang/FCNet.

Jinpeng Shi, Hui Li, Tianle Liu et al.

Transformers have achieved remarkable results in single-image super-resolution (SR). However, the challenge of balancing model performance and complexity has hindered their application in lightweight SR (LSR). To tackle this challenge, we propose an efficient striped window transformer (ESWT). We revisit the normalization layer in the transformer and design a concise and efficient transformer structure to build the ESWT. Furthermore, we introduce a striped window mechanism to model long-term dependencies more efficiently. To fully exploit the potential of the ESWT, we propose a novel flexible window training strategy that can improve the performance of the ESWT without additional cost. Extensive experiments show that ESWT outperforms state-of-the-art LSR transformers, and achieves a better trade-off between model performance and complexity. The ESWT requires fewer parameters, incurs faster inference, smaller FLOPs, and less memory consumption, making it a promising solution for LSR.

Zheng Chen, Kai Liu, Jingkai Wang et al.

This paper presents the NTIRE 2026 image super-resolution ($\times$4) challenge, one of the associated competitions of the NTIRE 2026 Workshop at CVPR 2026. The challenge aims to reconstruct high-resolution (HR) images from low-resolution (LR) inputs generated through bicubic downsampling with a $\times$4 scaling factor. The objective is to develop effective super-resolution solutions and analyze recent advances in the field. To reflect the evolving objectives of image super-resolution, the challenge includes two tracks: (1) a restoration track, which emphasizes pixel-wise fidelity and ranks submissions based on PSNR; and (2) a perceptual track, which focuses on visual realism and evaluates results using a perceptual score. A total of 194 participants registered for the challenge, with 31 teams submitting valid entries. This report summarizes the challenge design, datasets, evaluation protocol, main results, and methods of participating teams. The challenge provides a unified benchmark and offers insights into current progress and future directions in image super-resolution.

Hui Li, Yanxin Li

Understanding and predicting the evolution of across complex systems remains a fundamental challenge due to the absence of unified and computationally testable frameworks. Here we propose the Recursive Hierarchical Network(RHN), conceptualizing evolution as recursive encapsulation along a trajectory of node $\to$ module $\to$ system $\to$ new node, governed by gradual accumulation and abrupt transition. Theoretically, we formalize and prove the law of functional evolution, revealing an irreversible progression from structure-dominated to regulation-dominated to intelligence-dominated stages. Empirically, we operationalize functional levels and align life, cosmic, informational, and social systems onto this scale. The resulting trajectories are strictly monotonic and exhibit strong cross-system similarity, with high pairwise cosine similarities and robust stage resonance. We locate current system states and project future transitions. RHN provides a mathematically rigorous, multi-scale framework for reconstructing and predicting system evolution, offering theoretical guidance for designing next-generation intelligent systems.

Hui Li, Zihao Li, Rui Ma et al.

Class activation map (CAM) has been widely studied for visual explanation of the internal working mechanism of convolutional neural networks. The key of existing CAM-based methods is to compute effective weights to combine activation maps in the target convolution layer. Existing gradient and score based weighting schemes have shown superiority in ensuring either the discriminability or faithfulness of the CAM, but they normally cannot excel in both properties. In this paper, we propose a novel CAM weighting scheme, named FD-CAM, to improve both the faithfulness and discriminability of the CAM-based CNN visual explanation. First, we improve the faithfulness and discriminability of the score-based weights by performing a grouped channel switching operation. Specifically, for each channel, we compute its similarity group and switch the group of channels on or off simultaneously to compute changes in the class prediction score as the weights. Then, we combine the improved score-based weights with the conventional gradient-based weights so that the discriminability of the final CAM can be further improved. We perform extensive comparisons with the state-of-the-art CAM algorithms. The quantitative and qualitative results show our FD-CAM can produce more faithful and more discriminative visual explanations of the CNNs. We also conduct experiments to verify the effectiveness of the proposed grouped channel switching and weight combination scheme on improving the results. Our code is available at https://github.com/crishhh1998/FD-CAM.

Changhao Jiang, Jiahao Chen, Zhenghao Xiang et al.

Recent commercial systems such as Suno demonstrate strong capabilities in long-form song generation, while academic research remains largely non-reproducible due to the lack of publicly available training data, hindering fair comparison and progress. To this end, we release a fully open-source system for long-form song generation with fine-grained style conditioning, including a licensed synthetic dataset, training and evaluation pipelines, and Muse, an easy-to-deploy song generation model. The dataset consists of 116k fully licensed synthetic songs with automatically generated lyrics and style descriptions paired with audio synthesized by SunoV5. We train Muse via single-stage supervised finetuning of a Qwen-based language model extended with discrete audio tokens using MuCodec, without task-specific losses, auxiliary objectives, or additional architectural components. Our evaluations find that although Muse is trained with a modest data scale and model size, it achieves competitive performance on phoneme error rate, text--music style similarity, and audio aesthetic quality, while enabling controllable segment-level generation across different musical structures. All data, model weights, and training and evaluation pipelines will be publicly released, paving the way for continued progress in controllable long-form song generation research. The project repository is available at https://github.com/yuhui1038/Muse.

Zequan Xu, Qihang Sun, Shaofeng Hu et al.

The rise of the click farm business using Multi-purpose Messaging Mobile Apps (MMMAs) tempts cybercriminals to perpetrate crowdsourcing frauds that cause financial losses to click farm workers. In this paper, we propose a novel contrastive multi-view learning method named CMT for crowdsourcing fraud detection over the heterogeneous temporal graph (HTG) of MMMA. CMT captures both heterogeneity and dynamics of HTG and generates high-quality representations for crowdsourcing fraud detection in a self-supervised manner. We deploy CMT to detect crowdsourcing frauds on an industry-size HTG of a representative MMMA WeChat and it significantly outperforms other methods. CMT also shows promising results for fraud detection on a large-scale public financial HTG, indicating that it can be applied in other graph anomaly detection tasks. We provide our implementation at https://github.com/KDEGroup/CMT.

Tao Chen, Ze Lin, Hui Li et al.

Given the long textual product information and the product image, Multi-modal Product Summarization (MPS) aims to increase customers' desire to purchase by highlighting product characteristics with a short textual summary. Existing MPS methods can produce promising results. Nevertheless, they still 1) lack end-to-end product summarization, 2) lack multi-grained multi-modal modeling, and 3) lack multi-modal attribute modeling. To improve MPS, we propose an end-to-end multi-grained multi-modal attribute-aware product summarization method (MMAPS) for generating high-quality product summaries in e-commerce. MMAPS jointly models product attributes and generates product summaries. We design several multi-grained multi-modal tasks to better guide the multi-modal learning of MMAPS. Furthermore, we model product attributes based on both text and image modalities so that multi-modal product characteristics can be manifested in the generated summaries. Extensive experiments on a real large-scale Chinese e-commence dataset demonstrate that our model outperforms state-of-the-art product summarization methods w.r.t. several summarization metrics. Our code is publicly available at: https://github.com/KDEGroup/MMAPS.

Yunsheng Tian, Karl D. D. Willis, Bassel Al Omari et al.

The automated assembly of complex products requires a system that can automatically plan a physically feasible sequence of actions for assembling many parts together. In this paper, we present ASAP, a physics-based planning approach for automatically generating such a sequence for general-shaped assemblies. ASAP accounts for gravity to design a sequence where each sub-assembly is physically stable with a limited number of parts being held and a support surface. We apply efficient tree search algorithms to reduce the combinatorial complexity of determining such an assembly sequence. The search can be guided by either geometric heuristics or graph neural networks trained on data with simulation labels. Finally, we show the superior performance of ASAP at generating physically realistic assembly sequence plans on a large dataset of hundreds of complex product assemblies. We further demonstrate the applicability of ASAP on both simulation and real-world robotic setups. Project website: asap.csail.mit.edu

Chunyang Cheng, Tianyang Xu, Xiao-Jun Wu et al.

Evaluation is essential in image fusion research, yet most existing metrics are directly borrowed from other vision tasks without proper adaptation. These traditional metrics, often based on complex image transformations, not only fail to capture the true quality of the fusion results but also are computationally demanding. To address these issues, we propose a unified evaluation framework specifically tailored for image fusion. At its core is a lightweight network designed efficiently to approximate widely used metrics, following a divide-and-conquer strategy. Unlike conventional approaches that directly assess similarity between fused and source images, we first decompose the fusion result into infrared and visible components. The evaluation model is then used to measure the degree of information preservation in these separated components, effectively disentangling the fusion evaluation process. During training, we incorporate a contrastive learning strategy and inform our evaluation model by perceptual scene assessment provided by a large language model. Last, we propose the first consistency evaluation framework, which measures the alignment between image fusion metrics and human visual perception, using both independent no-reference scores and downstream tasks performance as objective references. Extensive experiments show that our learning-based evaluation paradigm delivers both superior efficiency (up to 1,000 times faster) and greater consistency across a range of standard image fusion benchmarks. Our code will be publicly available at https://github.com/AWCXV/EvaNet.

Hanchen Yang, Wengen Li, Shuyu Wang et al.

With the rapid amassing of spatial-temporal (ST) ocean data, many spatial-temporal data mining (STDM) studies have been conducted to address various oceanic issues, including climate forecasting and disaster warning. Compared with typical ST data (e.g., traffic data), ST ocean data is more complicated but with unique characteristics, e.g., diverse regionality and high sparsity. These characteristics make it difficult to design and train STDM models on ST ocean data. To the best of our knowledge, a comprehensive survey of existing studies remains missing in the literature, which hinders not only computer scientists from identifying the research issues in ocean data mining but also ocean scientists to apply advanced STDM techniques. In this paper, we provide a comprehensive survey of existing STDM studies for ocean science. Concretely, we first review the widely-used ST ocean datasets and highlight their unique characteristics. Then, typical ST ocean data quality enhancement techniques are explored. Next, we classify existing STDM studies in ocean science into four types of tasks, i.e., prediction, event detection, pattern mining, and anomaly detection, and elaborate on the techniques for these tasks. Finally, promising research opportunities are discussed. This survey can help scientists from both computer science and ocean science better understand the fundamental concepts, key techniques, and open challenges of STDM for ocean science.

Hui Li, Mingjie Sun, Jimin Xiao et al.

Referring Expression Segmentation (RES), which is aimed at localizing and segmenting the target according to the given language expression, has drawn increasing attention. Existing methods jointly consider the localization and segmentation steps, which rely on the fused visual and linguistic features for both steps. We argue that the conflict between the purpose of identifying an object and generating a mask limits the RES performance. To solve this problem, we propose a parallel position-kernel-segmentation pipeline to better isolate and then interact the localization and segmentation steps. In our pipeline, linguistic information will not directly contaminate the visual feature for segmentation. Specifically, the localization step localizes the target object in the image based on the referring expression, and then the visual kernel obtained from the localization step guides the segmentation step. This pipeline also enables us to train RES in a weakly-supervised way, where the pixel-level segmentation labels are replaced by click annotations on center and corner points. The position head is fully-supervised and trained with the click annotations as supervision, and the segmentation head is trained with weakly-supervised segmentation losses. To validate our framework on a weakly-supervised setting, we annotated three RES benchmark datasets (RefCOCO, RefCOCO+ and RefCOCOg) with click annotations.Our method is simple but surprisingly effective, outperforming all previous state-of-the-art RES methods on fully- and weakly-supervised settings by a large margin. The benchmark code and datasets will be released.

Yanglin Deng, Tianyang Xu, Chunyang Cheng et al.

Infrared and visible image fusion(IVIF) combines complementary modalities while preserving natural textures and salient thermal signatures. Existing solutions predominantly rely on extensive sets of rigidly aligned image pairs for training. However, acquiring such data is often impractical due to the costly and labour-intensive alignment process. Besides, maintaining a rigid pairing setting during training restricts the volume of cross-modal relationships, thereby limiting generalisation performance. To this end, this work challenges the necessity of Strictly Paired Training Paradigm (SPTP) by systematically investigating UnPaired and Arbitrarily Paired Training Paradigms (UPTP and APTP) for high-performance IVIF. We establish a theoretical objective of APTP, reflecting the complementary nature between UPTP and SPTP. More importantly, we develop a practical framework capable of significantly enriching cross-modal relationships even with severely limited and unaligned training data. To validate our propositions, three end-to-end lightweight baselines, alongside a set of innovative loss functions, are designed to cover three classic frameworks (CNN, Transformer, GAN). Comprehensive experiments demonstrate that the proposed APTP and UPTP are feasible and capable of training models on a severely limited and content-inconsistent infrared and visible dataset, achieving performance comparable to that of a dataset 100$\times$ larger in SPTP. This finding fundamentally alleviates the cost and difficulty of data collection while enhancing model robustness from the data perspective, delivering a feasible solution for IVIF studies. The code is available at \href{https://github.com/yanglinDeng/IVIF_unpair}{\textcolor{blue}{https://github.com/yanglinDeng/IVIF\_unpair}}.

Ziyao Guo, Kai Wang, George Cazenavette et al.

The ultimate goal of Dataset Distillation is to synthesize a small synthetic dataset such that a model trained on this synthetic set will perform equally well as a model trained on the full, real dataset. Until now, no method of Dataset Distillation has reached this completely lossless goal, in part due to the fact that previous methods only remain effective when the total number of synthetic samples is extremely small. Since only so much information can be contained in such a small number of samples, it seems that to achieve truly loss dataset distillation, we must develop a distillation method that remains effective as the size of the synthetic dataset grows. In this work, we present such an algorithm and elucidate why existing methods fail to generate larger, high-quality synthetic sets. Current state-of-the-art methods rely on trajectory-matching, or optimizing the synthetic data to induce similar long-term training dynamics as the real data. We empirically find that the training stage of the trajectories we choose to match (i.e., early or late) greatly affects the effectiveness of the distilled dataset. Specifically, early trajectories (where the teacher network learns easy patterns) work well for a low-cardinality synthetic set since there are fewer examples wherein to distribute the necessary information. Conversely, late trajectories (where the teacher network learns hard patterns) provide better signals for larger synthetic sets since there are now enough samples to represent the necessary complex patterns. Based on our findings, we propose to align the difficulty of the generated patterns with the size of the synthetic dataset. In doing so, we successfully scale trajectory matching-based methods to larger synthetic datasets, achieving lossless dataset distillation for the very first time. Code and distilled datasets are available at https://gzyaftermath.github.io/DATM.

Weijia Dou, Hui Li, Jiahao Cui et al.

Streaming video generation models typically rely on temporal-centric memory, which organizes historical context as raw frames, chunk segments, or unclustered tokens. This organization frequently leads to identity drift and semantic inconsistency when entities exit the frame or during interactive prompt transitions. To address these limitations, we propose SlotMemory, an object-centric Key-Value memory mechanism for streaming video diffusion. Our approach shifts the memory abstraction from "when" an event occurred to "what" is being represented by decomposing the transformer's key-value manifold into discrete, reusable semantic slots. By utilizing these slots as routing addresses to index and store high-fidelity key-value tokens, we enable entity-level persistence and prompt-aware retrieval across long horizons. Evaluated on 60-second interactive narratives using the Wan2.1-T2V-1.3B backbone, SlotMemory achieves a state-of-the-art quality score of 81.61 and a 22.8 percent relative improvement in dynamic consistency over the strongest existing streaming baseline. Our results demonstrate that structured semantic representation, rather than raw temporal capacity, is the essential primitive for persistent long-form video synthesis. Our codes and checkpoints are available at https://tj12323.github.io/SlotMemory/.

DeepSeek-AI, Daya Guo, Dejian Yang et al. · stanford, tsinghua

We introduce our first-generation reasoning models, DeepSeek-R1-Zero and DeepSeek-R1. DeepSeek-R1-Zero, a model trained via large-scale reinforcement learning (RL) without supervised fine-tuning (SFT) as a preliminary step, demonstrates remarkable reasoning capabilities. Through RL, DeepSeek-R1-Zero naturally emerges with numerous powerful and intriguing reasoning behaviors. However, it encounters challenges such as poor readability, and language mixing. To address these issues and further enhance reasoning performance, we introduce DeepSeek-R1, which incorporates multi-stage training and cold-start data before RL. DeepSeek-R1 achieves performance comparable to OpenAI-o1-1217 on reasoning tasks. To support the research community, we open-source DeepSeek-R1-Zero, DeepSeek-R1, and six dense models (1.5B, 7B, 8B, 14B, 32B, 70B) distilled from DeepSeek-R1 based on Qwen and Llama.

Zequan Xu, Lianyun Li, Hui Li et al.

Nowadays, Multi-purpose Messaging Mobile App (MMMA) has become increasingly prevalent. MMMAs attract fraudsters and some cybercriminals provide support for frauds via black market accounts (BMAs). Compared to fraudsters, BMAs are not directly involved in frauds and are more difficult to detect. This paper illustrates our BMA detection system SGRL (Self-supervised Graph Representation Learning) used in WeChat, a representative MMMA with over a billion users. We tailor Graph Neural Network and Graph Self-supervised Learning in SGRL for BMA detection. The workflow of SGRL contains a pretraining phase that utilizes structural information, node attribute information and available human knowledge, and a lightweight detection phase. In offline experiments, SGRL outperforms state-of-the-art methods by 16.06%-58.17% on offline evaluation measures. We deploy SGRL in the online environment to detect BMAs on the billion-scale WeChat graph, and it exceeds the alternative by 7.27% on the online evaluation measure. In conclusion, SGRL can alleviate label reliance, generalize well to unseen data, and effectively detect BMAs in WeChat.

DeepSeek-AI, Aixin Liu, Bei Feng et al. · pku

We present DeepSeek-V2, a strong Mixture-of-Experts (MoE) language model characterized by economical training and efficient inference. It comprises 236B total parameters, of which 21B are activated for each token, and supports a context length of 128K tokens. DeepSeek-V2 adopts innovative architectures including Multi-head Latent Attention (MLA) and DeepSeekMoE. MLA guarantees efficient inference through significantly compressing the Key-Value (KV) cache into a latent vector, while DeepSeekMoE enables training strong models at an economical cost through sparse computation. Compared with DeepSeek 67B, DeepSeek-V2 achieves significantly stronger performance, and meanwhile saves 42.5% of training costs, reduces the KV cache by 93.3%, and boosts the maximum generation throughput to 5.76 times. We pretrain DeepSeek-V2 on a high-quality and multi-source corpus consisting of 8.1T tokens, and further perform Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) to fully unlock its potential. Evaluation results show that, even with only 21B activated parameters, DeepSeek-V2 and its chat versions still achieve top-tier performance among open-source models.

Hui Li, Zidong Guo, Seon-Min Rhee et al.

Accurate facial landmarks are essential prerequisites for many tasks related to human faces. In this paper, an accurate facial landmark detector is proposed based on cascaded transformers. We formulate facial landmark detection as a coordinate regression task such that the model can be trained end-to-end. With self-attention in transformers, our model can inherently exploit the structured relationships between landmarks, which would benefit landmark detection under challenging conditions such as large pose and occlusion. During cascaded refinement, our model is able to extract the most relevant image features around the target landmark for coordinate prediction, based on deformable attention mechanism, thus bringing more accurate alignment. In addition, we propose a novel decoder that refines image features and landmark positions simultaneously. With few parameter increasing, the detection performance improves further. Our model achieves new state-of-the-art performance on several standard facial landmark detection benchmarks, and shows good generalization ability in cross-dataset evaluation.

Yifei Xu, Jingguo Ge, Haina Tang et al.

In the AIOps (Artificial Intelligence for IT Operations) era, accurately forecasting system states is crucial. In microservices systems, this task encounters the challenge of dynamic and complex spatio-temporal relationships among microservice instances, primarily due to dynamic deployments, diverse call paths, and cascading effects among instances. Current time-series forecasting methods, which focus mainly on intrinsic patterns, are insufficient in environments where spatial relationships are critical. Similarly, spatio-temporal graph approaches often neglect the nature of temporal trend, concentrating mostly on message passing between nodes. Moreover, current research in microservices domain frequently underestimates the importance of network metrics and topological structures in capturing the evolving dynamics of systems. This paper introduces STMformer, a model tailored for forecasting system states in microservices environments, capable of handling multi-node and multivariate time series. Our method leverages dynamic network connection data and topological information to assist in modeling the intricate spatio-temporal relationships within the system. Additionally, we integrate the PatchCrossAttention module to compute the impact of cascading effects globally. We have developed a dataset based on a microservices system and conducted comprehensive experiments with STMformer against leading methods. In both short-term and long-term forecasting tasks, our model consistently achieved a 8.6% reduction in MAE(Mean Absolute Error) and a 2.2% reduction in MSE (Mean Squared Error). The source code is available at https://github.com/xuyifeiiie/STMformer.

Jianan Li, Shenwang Jiang, Liqiang Song et al.

The Five-hundred-meter Aperture Spherical radio Telescope (FAST) is the world's largest single-dish radio telescope. Its large reflecting surface achieves unprecedented sensitivity but is prone to damage, such as dents and holes, caused by naturally-occurring falling objects. Hence, the timely and accurate detection of surface defects is crucial for FAST's stable operation. Conventional manual inspection involves human inspectors climbing up and examining the large surface visually, a time-consuming and potentially unreliable process. To accelerate the inspection process and increase its accuracy, this work makes the first step towards automating the inspection of FAST by integrating deep-learning techniques with drone technology. First, a drone flies over the surface along a predetermined route. Since surface defects significantly vary in scale and show high inter-class similarity, directly applying existing deep detectors to detect defects on the drone imagery is highly prone to missing and misidentifying defects. As a remedy, we introduce cross-fusion, a dedicated plug-in operation for deep detectors that enables the adaptive fusion of multi-level features in a point-wise selective fashion, depending on local defect patterns. Consequently, strong semantics and fine-grained details are dynamically fused at different positions to support the accurate detection of defects of various scales and types. Our AI-powered drone-based automated inspection is time-efficient, reliable, and has good accessibility, which guarantees the long-term and stable operation of FAST.

Hao Cheng, Jinhao Duan, Hui Li et al.

Recently, there has been a surge of interest and attention in Transformer-based structures, such as Vision Transformer (ViT) and Vision Multilayer Perceptron (VMLP). Compared with the previous convolution-based structures, the Transformer-based structure under investigation showcases a comparable or superior performance under its distinctive attention-based input token mixer strategy. Introducing adversarial examples as a robustness consideration has had a profound and detrimental impact on the performance of well-established convolution-based structures. This inherent vulnerability to adversarial attacks has also been demonstrated in Transformer-based structures. In this paper, our emphasis lies on investigating the intrinsic robustness of the structure rather than introducing novel defense measures against adversarial attacks. To address the susceptibility to robustness issues, we employ a rational structure design approach to mitigate such vulnerabilities. Specifically, we enhance the adversarial robustness of the structure by increasing the proportion of high-frequency structural robust biases. As a result, we introduce a novel structure called Robust Bias Transformer-based Structure (RBFormer) that shows robust superiority compared to several existing baseline structures. Through a series of extensive experiments, RBFormer outperforms the original structures by a significant margin, achieving an impressive improvement of +16.12% and +5.04% across different evaluation criteria on CIFAR-10 and ImageNet-1k, respectively.

Chenyu Xue, Fan Wang, Yuanzhuo Zhu et al.

Deploying reliable deep learning techniques in interdisciplinary applications needs learned models to output accurate and (even more importantly) explainable predictions. Existing approaches typically explicate network outputs in a post-hoc fashion, under an implicit assumption that faithful explanations come from accurate predictions/classifications. We have an opposite claim that explanations boost (or even determine) classification. That is, end-to-end learning of explanation factors to augment discriminative representation extraction could be a more intuitive strategy to inversely assure fine-grained explainability, e.g., in those neuroimaging and neuroscience studies with high-dimensional data containing noisy, redundant, and task-irrelevant information. In this paper, we propose such an explainable geometric deep network dubbed as NeuroExplainer, with applications to uncover altered infant cortical development patterns associated with preterm birth. Given fundamental cortical attributes as network input, our NeuroExplainer adopts a hierarchical attention-decoding framework to learn fine-grained attentions and respective discriminative representations to accurately recognize preterm infants from term-born infants at term-equivalent age. NeuroExplainer learns the hierarchical attention-decoding modules under subject-level weak supervision coupled with targeted regularizers deduced from domain knowledge regarding brain development. These prior-guided constraints implicitly maximizes the explainability metrics (i.e., fidelity, sparsity, and stability) in network training, driving the learned network to output detailed explanations and accurate classifications. Experimental results on the public dHCP benchmark suggest that NeuroExplainer led to quantitatively reliable explanation results that are qualitatively consistent with representative neuroimaging studies.

Letian Fu, Huang Huang, Gaurav Datta et al.

We explore how to enhance next-token prediction models to perform in-context imitation learning on a real robot, where the robot executes new tasks by interpreting contextual information provided during the input phase, without updating its underlying policy parameters. We propose In-Context Robot Transformer (ICRT), a causal transformer that performs autoregressive prediction on sensorimotor trajectories without relying on any linguistic data or reward function. This formulation enables flexible and training-free execution of new tasks at test time, achieved by prompting the model with sensorimotor trajectories of the new task composing of image observations, actions and states tuples, collected through human teleoperation. Experiments with a Franka Emika robot demonstrate that the ICRT can adapt to new tasks specified by prompts, even in environment configurations that differ from both the prompt and the training data. In a multitask environment setup, ICRT significantly outperforms current state-of-the-art next-token prediction models in robotics on generalizing to unseen tasks. Code, checkpoints and data are available on https://icrt.dev/

Zichen Yu, Changyong Shu, Jiajun Deng et al.

Given the capability of mitigating the long-tail deficiencies and intricate-shaped absence prevalent in 3D object detection, occupancy prediction has become a pivotal component in autonomous driving systems. However, the procession of three-dimensional voxel-level representations inevitably introduces large overhead in both memory and computation, obstructing the deployment of to-date occupancy prediction approaches. In contrast to the trend of making the model larger and more complicated, we argue that a desirable framework should be deployment-friendly to diverse chips while maintaining high precision. To this end, we propose a plug-and-play paradigm, namely FlashOCC, to consolidate rapid and memory-efficient occupancy prediction while maintaining high precision. Particularly, our FlashOCC makes two improvements based on the contemporary voxel-level occupancy prediction approaches. Firstly, the features are kept in the BEV, enabling the employment of efficient 2D convolutional layers for feature extraction. Secondly, a channel-to-height transformation is introduced to lift the output logits from the BEV into the 3D space. We apply the FlashOCC to diverse occupancy prediction baselines on the challenging Occ3D-nuScenes benchmarks and conduct extensive experiments to validate the effectiveness. The results substantiate the superiority of our plug-and-play paradigm over previous state-of-the-art methods in terms of precision, runtime efficiency, and memory costs, demonstrating its potential for deployment. The code will be made available.

Zhengjia Zhong, Shuyan Ke, Zaizhou Lin et al.

Vector quantization is a fundamental tool for compressing high-dimensional embeddings, yet existing multi-codebook methods rely on static codebooks that limit expressiveness under heterogeneous data geometry. While recent dynamic quantizers like QINCo adapt codebooks to individual inputs and improve expressiveness, their strict sequential dependencies create decoding bottlenecks. We propose Residual Quantization via Mixture of Experts (RQ-MoE), a framework combining a two-level MoE with dual-stream quantization to enable input-dependent codebook adaptation for efficient vector quantization. RQ-MoE enables dynamic codebook construction and decouples instruction from quantization, facilitating parallel decoding. Theoretically, we show that standard Residual Quantization and QINCo can be recovered as constrained special cases of RQ-MoE, and derive a guideline for setting expert dimensionality in RQ-MoE. Extensive experiments show that RQ-MoE achieves state-of-the-art or on-par performance in reconstruction and retrieval, while providing 6x-14x faster decoding than prior vector quantization methods. The implementation is available at https://github.com/KDEGroup/RQ-MoE.

Yuhan Ding, Fukun Yin, Jiayuan Fan et al.

Recent advances in implicit neural representations have achieved impressive results by sampling and fusing individual points along sampling rays in the sampling space. However, due to the explosively growing sampling space, finely representing and synthesizing detailed textures remains a challenge for unbounded large-scale outdoor scenes. To alleviate the dilemma of using individual points to perceive the entire colossal space, we explore learning the surface distribution of the scene to provide structural priors and reduce the samplable space and propose a Point Diffusion implicit Function, PDF, for large-scale scene neural representation. The core of our method is a large-scale point cloud super-resolution diffusion module that enhances the sparse point cloud reconstructed from several training images into a dense point cloud as an explicit prior. Then in the rendering stage, only sampling points with prior points within the sampling radius are retained. That is, the sampling space is reduced from the unbounded space to the scene surface. Meanwhile, to fill in the background of the scene that cannot be provided by point clouds, the region sampling based on Mip-NeRF 360 is employed to model the background representation. Expensive experiments have demonstrated the effectiveness of our method for large-scale scene novel view synthesis, which outperforms relevant state-of-the-art baselines.

Caiyuan Zheng, Hui Li, Seon-Min Rhee et al.

Scene text recognition (STR) attracts much attention over the years because of its wide application. Most methods train STR model in a fully supervised manner which requires large amounts of labeled data. Although synthetic data contributes a lot to STR, it suffers from the real-tosynthetic domain gap that restricts model performance. In this work, we aim to boost STR models by leveraging both synthetic data and the numerous real unlabeled images, exempting human annotation cost thoroughly. A robust consistency regularization based semi-supervised framework is proposed for STR, which can effectively solve the instability issue due to domain inconsistency between synthetic and real images. A character-level consistency regularization is designed to mitigate the misalignment between characters in sequence recognition. Extensive experiments on standard text recognition benchmarks demonstrate the effectiveness of the proposed method. It can steadily improve existing STR models, and boost an STR model to achieve new state-of-the-art results. To our best knowledge, this is the first consistency regularization based framework that applies successfully to STR.

Yining Wang, Xumeng Gong, Shaochuan Li et al.

In the field of antibody engineering, an essential task is to design a novel antibody whose paratopes bind to a specific antigen with correct epitopes. Understanding antibody structure and its paratope can facilitate a mechanistic understanding of its function. Therefore, antibody structure prediction from its sequence alone has always been a highly valuable problem for de novo antibody design. AlphaFold2, a breakthrough in the field of structural biology, provides a solution to predict protein structure based on protein sequences and computationally expensive coevolutionary multiple sequence alignments (MSAs). However, the computational efficiency and undesirable prediction accuracy of antibodies, especially on the complementarity-determining regions (CDRs) of antibodies limit their applications in the industrially high-throughput drug design. To learn an informative representation of antibodies, we employed a deep antibody language model (ALM) on curated sequences from the observed antibody space database via a transformer model. We also developed a novel model named xTrimoABFold to predict antibody structure from antibody sequence based on the pretrained ALM as well as efficient evoformers and structural modules. The model was trained end-to-end on the antibody structures in PDB by minimizing the ensemble loss of domain-specific focal loss on CDR and the frame-aligned point loss. xTrimoABFold outperforms AlphaFold2 and other protein language model based SOTAs, e.g., OmegaFold, HelixFold-Single, and IgFold with a large significant margin (30+\% improvement on RMSD) while performing 151 times faster than AlphaFold2. To the best of our knowledge, xTrimoABFold achieved state-of-the-art antibody structure prediction. Its improvement in both accuracy and efficiency makes it a valuable tool for de novo antibody design and could make further improvements in immuno-theory.

Wang Bin, Ao Yang, Kedan Li et al.

In the domain of software security testing, Directed Grey-Box Fuzzing (DGF) has garnered widespread attention for its efficient target localization and excellent detection performance. However, existing approaches measure only the physical distance between seed execution paths and target locations, overlooking logical relationships among code segments. This omission can yield redundant or misleading guidance in complex binaries, weakening DGF's real-world effectiveness. To address this, we introduce \textbf{attention distance}, a novel metric that leverages a large language model's contextual analysis to compute attention scores between code elements and reveal their intrinsic connections. Under the same AFLGo configuration -- without altering any fuzzing components other than the distance metric -- replacing physical distances with attention distances across 38 real vulnerability reproduction experiments delivers a \textbf{3.43$\times$} average increase in testing efficiency over the traditional method. Compared to state-of-the-art directed fuzzers DAFL and WindRanger, our approach achieves \textbf{2.89$\times$} and \textbf{7.13$\times$} improvements, respectively. To further validate the generalizability of attention distance, we integrate it into DAFL and WindRanger, where it also consistently enhances their original performance. All related code and datasets are publicly available at https://github.com/TheBinKing/Attention\_Distance.git.

Zhenxi Zhang, Yitao Zhuang, Yao Pu et al.

Anatomical structure masks are widely adopted in radiotherapy dose prediction, as they provide explicit geometric constraints that facilitate structure-dose coupling. However, conventional manual delineation of these masks requires precise annotation of structure boundaries relevant to radiotherapy, which is time-consuming and labor-intensive. To address these limitations, we propose a scribble-guided dose prediction framework that relies solely on anatomical structures annotated with sparse scribbles. Specifically, we design a Scribble Completion Module (SCM) to generate dense anatomical masks by propagating sparse scribble labels to semantically similar voxels. During the propagation process, a supervoxel-based regularization is introduced to preserve geometric boundary consistency to ensure anatomical plausibility. Furthermore, we propose a Structure-Guided Dose Generation Module (SGDGM) to strengthen the correspondence between sparse structural cues and dose distribution. The completed dense masks derived from scribbles serve as structural guidance to condition dose prediction, forming a scribble-mask-dose learning pipeline under sparse annotation. Experiments on the GDP-HMM dataset demonstrate that ScribbleDose achieves competitive dose prediction performance using only sparse structural annotations. The source code and reannotated scribble annotations are publicly available at https://github.com/iCherishxixixi/ScribbleDose.

Shuquan Lian, Juncheng Liu, Yazhe Chen et al.

Prior representative ReAct-style approaches in autonomous Software Engineering (SWE) typically lack the explicit System-2 reasoning required for deep analysis and handling complex edge cases. While recent reasoning models demonstrate the potential of extended Chain-of-Thought (CoT), applying them to the multi-turn SWE task creates a fundamental dilemma: retaining full reasoning history leads to context explosion and ``Lost-in-the-Middle'' degradation, while discarding it would force the agent to redundantly re-reason at every step. To address these challenges, we propose SWE-AGILE, a novel software agent framework designed to bridge the gap between reasoning depth, efficiency, and context constraints. SWE-AGILE introduces a Dynamic Reasoning Context strategy, maintaining a ``sliding window'' of detailed reasoning for immediate continuity to prevent redundant re-analyzing, while compressing historical reasoning content into concise Reasoning Digests. Empirically, SWE-AGILE sets a new standard for 7B-8B models on SWE-Bench-Verified using only 2.2k trajectories and 896 tasks. Code is available at https://github.com/KDEGroup/SWE-AGILE.

Yaqian Zhang, Ruyi Yu, Tianyi Li et al.

Industrial recommender systems increasingly leverage lifelong user behavior histories and rich multi-modal content to capture evolving user preferences. However, effectively integrating multi-modal features into lifelong interest modeling remains challenging due to the inherent misalignment between multi-modal and collaborative spaces. Existing paradigms typically rely on separate modeling of multi-modal sequence and behavior sequence, and late fusion to alleviate the modality gap, which results in coarse-grained multi-modal representation and limited integration. In this paper, we propose SIREN, a unified multi-granularity semantic interaction framework for multi-modal lifelong user interest modeling. In the General Search Unit stage, we introduce two alternative retrieval strategies: multi-modal similarity-based soft retrieval for retrieval effectiveness, and Semantic ID (SemID)-based hard retrieval for efficient industrial serving. For the Exact Search Unit stage, we explicitly incorporate target-aware relevance via coarse similarity buckets and fine-grained prefix-encoded SemIDs, enabling unified interaction with collaborative ID features within the target-conditioned transformer architecture. Extensive experiments on the offline dataset demonstrate that SIREN achieves a state-of-the-art GAUC. Online A/B tests further demonstrate consistent GMV gains across multiple production scenarios, including +2.28% in Weixin Moments, +3.87% in Weixin Official Accounts, and +1.61% in Weixin Channels. From July 2025, SIREN has been fully launched for full-traffic serving in Tencent's advertising platform.

Jiabin Liu, Zihao Zhou, Jialei Yan et al.

Precise three-dimensional (3D) reconstruction of wave free surfaces and associated velocity fields is essential for developing a comprehensive understanding of ocean physics. To address the high computational cost of dense visual reconstruction in long-term ocean wave observation tasks and the challenges introduced by persistent visual occlusions, we propose an wave free surface visual reconstruction neural network, which is designed as an attention-augmented pyramid architecture tailored to the multi-scale and temporally continuous characteristics of wave motions. Using physics-based constraints, we perform time-resolved reconstruction of nonlinear 3D velocity fields from the evolving free-surface boundary. Experiments under real-sea conditions demonstrate millimetre-level wave elevation prediction in the central region, dominant-frequency errors below 0.01 Hz, precise estimation of high-frequency spectral power laws, and high-fidelity 3D reconstruction of nonlinear velocity fields, while enabling dense reconstruction of two million points in only 1.35 s. Built on a stereo-vision dataset, the model outperforms conventional visual reconstruction approaches and maintains strong generalization in occluded conditions, owing to its global multi-scale attention and its learned encoding of wave propagation dynamics.

Wei-Dong Qiao, Yang Xu, Hui Li

The weight-sharing mechanism of convolutional kernels ensures translation-equivariance of convolution neural networks (CNNs). Recently, rotation-equivariance has been investigated. However, research on scale-equivariance or simultaneous scale-rotation-equivariance is insufficient. This study proposes a Lie group-CNN, which can keep scale-rotation-equivariance for image classification tasks. The Lie group-CNN includes a lifting module, a series of group convolution modules, a global pooling layer, and a classification layer. The lifting module transfers the input image from Euclidean space to Lie group space, and the group convolution is parameterized through a fully connected network using Lie-algebra of Lie-group elements as inputs to achieve scale-rotation-equivariance. The Lie group SIM(2) is utilized to establish the Lie group-CNN with scale-rotation-equivariance. Scale-rotation-equivariance of Lie group-CNN is verified and achieves the best recognition accuracy on the blood cell dataset (97.50%) and the HAM10000 dataset (77.90%) superior to Lie algebra convolution network, dilation convolution, spatial transformer network, and scale-equivariant steerable network. In addition, the generalization ability of the Lie group-CNN on SIM(2) on rotation-equivariance is verified on rotated-MNIST and rotated-CIFAR10, and the robustness of the network is verified on SO(2) and SE(2). Therefore, the Lie group-CNN can successfully extract geometric features and performs equivariant recognition on images with rotation and scale transformations.

Shuangchen Zhao, Changde Du, Hui Li et al.

Deep Neural Networks (DNNs) have demonstrated exceptional recognition capabilities in traditional computer vision (CV) tasks. However, existing CV models often suffer a significant decrease in accuracy when confronted with out-of-distribution (OOD) data. In contrast to these DNN models, human can maintain a consistently low error rate when facing OOD scenes, partly attributed to the rich prior cognitive knowledge stored in the human brain. Previous OOD generalization researches only focus on the single modal, overlooking the advantages of multimodal learning method. In this paper, we utilize the multimodal learning method to improve the OOD generalization and propose a novel Brain-machine Fusion Learning (BMFL) framework. We adopt the cross-attention mechanism to fuse the visual knowledge from CV model and prior cognitive knowledge from the human brain. Specially, we employ a pre-trained visual neural encoding model to predict the functional Magnetic Resonance Imaging (fMRI) from visual features which eliminates the need for the fMRI data collection and pre-processing, effectively reduces the workload associated with conventional BMFL methods. Furthermore, we construct a brain transformer to facilitate the extraction of knowledge inside the fMRI data. Moreover, we introduce the Pearson correlation coefficient maximization regularization method into the training process, which improves the fusion capability with better constrains. Our model outperforms the DINOv2 and baseline models on the ImageNet-1k validation dataset as well as six curated OOD datasets, showcasing its superior performance in diverse scenarios.

Jiaying Wu, Can Gao, Jinglu Hu et al.

Few-shot learning aims to identify novel categories from only a handful of labeled samples, where prototypes estimated from scarce data are often biased and generalize poorly. Semantic-based methods alleviate this by introducing coarse class-level information, but they are mostly applied on the support side, leaving query representations unchanged. In this paper, we present PMCE, a Probabilistic few-shot framework that leverages Multi-granularity semantics with Caption-guided Enhancement. PMCE constructs a nonparametric knowledge bank that stores visual statistics for each category as well as CLIP-encoded class name embeddings of the base classes. At meta-test time, the most relevant base classes are retrieved based on the similarities of class name embeddings for each novel category. These statistics are then aggregated into category-specific prior information and fused with the support set prototypes via a simple MAP update. Simultaneously, a frozen BLIP captioner provides label-free instance-level image descriptions, and a lightweight enhancer trained on base classes optimizes both support prototypes and query features under an inductive protocol with a consistency regularization to stabilize noisy captions. Experiments on four benchmarks show that PMCE consistently improves over strong baselines, achieving up to 7.71% absolute gain over the strongest semantic competitor on MiniImageNet in the 1-shot setting. Our code is available at https://anonymous.4open.science/r/PMCE-275D