Min Yang

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.

Yufeng Wang, Cong Xu, Min Yang et al.

Although Physics-Informed Neural Networks (PINNs) have been successfully applied in a wide variety of science and engineering fields, they can fail to accurately predict the underlying solution in slightly challenging convection-diffusion-reaction problems. In this paper, we investigate the reason of this failure from a domain distribution perspective, and identify that learning multi-scale fields simultaneously makes the network unable to advance its training and easily get stuck in poor local minima. We show that the widespread experience of sampling more collocation points in high-loss layer regions hardly help optimize and may even worsen the results. These findings motivate the development of a novel curriculum learning method that encourages neural networks to prioritize learning on easier non-layer regions while downplaying learning on harder layer regions. The proposed method helps PINNs automatically adjust the learning emphasis and thereby facilitate the optimization procedure. Numerical results on typical benchmark equations show that the proposed curriculum learning approach mitigates the failure modes of PINNs and can produce accurate results for very sharp boundary and interior layers. Our work reveals that for equations whose solutions have large scale differences, paying less attention to high-loss regions can be an effective strategy for learning them accurately.

Sunbowen Lee, Shiwen Ni, Chi Wei et al.

Safety alignment mechanism are essential for preventing large language models (LLMs) from generating harmful information or unethical content. However, cleverly crafted prompts can bypass these safety measures without accessing the model's internal parameters, a phenomenon known as black-box jailbreak. Existing heuristic black-box attack methods, such as genetic algorithms, suffer from limited effectiveness due to their inherent randomness, while recent reinforcement learning (RL) based methods often lack robust and informative reward signals. To address these challenges, we propose a novel black-box jailbreak method leveraging RL, which optimizes prompt generation by analyzing the embedding proximity between benign and malicious prompts. This approach ensures that the rewritten prompts closely align with the intent of the original prompts while enhancing the attack's effectiveness. Furthermore, we introduce a comprehensive jailbreak evaluation framework incorporating keywords, intent matching, and answer validation to provide a more rigorous and holistic assessment of jailbreak success. Experimental results show the superiority of our approach, achieving state-of-the-art (SOTA) performance on several prominent open and closed-source LLMs, including Qwen2.5-7B-Instruct, Llama3.1-8B-Instruct, and GPT-4o-0806. Our method sets a new benchmark in jailbreak attack effectiveness, highlighting potential vulnerabilities in LLMs. The codebase for this work is available at https://github.com/Aegis1863/xJailbreak.

Run Luo, Ting-En Lin, Haonan Zhang et al.

Recent advancements in omnimodal learning have significantly improved understanding and generation across images, text, and speech, yet these developments remain predominantly confined to proprietary models. The lack of high-quality omnimodal datasets and the challenges of real-time emotional speech synthesis have notably hindered progress in open-source research. To address these limitations, we introduce \name, a two-stage training framework that integrates omnimodal alignment and speech generation to develop a state-of-the-art omnimodal large language model. In the alignment phase, a pre-trained speech model undergoes further training on text-image tasks, enabling (near) zero-shot generalization from vision to speech, outperforming models trained on tri-modal datasets. In the speech generation phase, a lightweight decoder is trained on speech tasks with direct preference optimization, enabling real-time emotional speech synthesis with high fidelity. Experiments show that \name surpasses state-of-the-art models across omnimodal, vision-language, and speech-language benchmarks. It achieves a 4-point absolute improvement on OmniBench over the leading open-source model VITA, despite using 5x fewer training samples and a smaller model size (7B vs. 7x8B). Additionally, \name achieves real-time speech generation with <1s latency at non-autoregressive mode, reducing inference time by 5x compared to autoregressive methods, and improves emotion classification accuracy by 7.7\%

Dingwei Chen, Ziqiang Liu, Feiteng Fang et al.

Large Language Models (LLMs) demonstrate remarkable capabilities in text understanding and generation. However, their tendency to produce factually inconsistent outputs, commonly referred to as ''hallucinations'', remains a critical challenge. Existing approaches, such as retrieval-based and inference-time correction methods, primarily address this issue at the input or output level, often overlooking the intrinsic information refinement process and the role of premature layers. Meanwhile, alignment- and fine-tuning-based methods are resource-intensive. In this paper, we propose PLI (Premature Layers Interpolation), a novel, training-free, and plug-and-play intervention designed to enhance factuality. PLI mitigates hallucinations by inserting premature layers formed through mathematical interpolation with adjacent layers. Inspired by stable diffusion and sampling steps, PLI extends the depth of information processing and transmission in LLMs, improving factual coherence. Experiments on four publicly available datasets demonstrate that PLI effectively reduces hallucinations while outperforming existing baselines in most cases. Further analysis suggests that the success of layer interpolation is closely linked to LLMs' internal mechanisms. To promote reproducibility, we will release our code and data upon acceptance.

Feiteng Fang, Ting-En Lin, Yuchuan Wu et al.

Role-Playing Language Agents (RPLAs) aim to simulate characters for realistic and engaging human-computer interactions. However, traditional reward models often struggle with scalability and adapting to subjective conversational preferences. We propose ChARM, a Character-based Act-adaptive Reward Model, addressing these challenges through two innovations: (1) an act-adaptive margin that significantly enhances learning efficiency and generalizability, and (2) a self-evolution mechanism leveraging large-scale unlabeled data to improve training coverage. Additionally, we introduce RoleplayPref, the first large-scale preference dataset specifically for RPLAs, featuring 1,108 characters, 13 subcategories, and 16,888 bilingual dialogues, alongside RoleplayEval, a dedicated evaluation benchmark. Experimental results show a 13% improvement over the conventional Bradley-Terry model in preference rankings. Furthermore, applying ChARM-generated rewards to preference learning techniques (e.g., direct preference optimization) achieves state-of-the-art results on CharacterEval and RoleplayEval. Code and dataset are available at https://github.com/calubkk/ChARM.

Changyi Li, Jiayi Wang, Xudong Pan et al.

Large Reasoning Models (LRMs) leverage transparent reasoning traces, known as Chain-of-Thoughts (CoTs), to break down complex problems into intermediate steps and derive final answers. However, these reasoning traces introduce unique safety challenges: harmful content can be embedded in intermediate steps even when final answers appear benign. Existing moderation tools, designed to handle generated answers, struggle to effectively detect hidden risks within CoTs. To address these challenges, we introduce ReasoningShield, a lightweight yet robust framework for moderating CoTs in LRMs. Our key contributions include: (1) formalizing the task of CoT moderation with a multi-level taxonomy of 10 risk categories across 3 safety levels, (2) creating the first CoT moderation benchmark which contains 9.2K pairs of queries and reasoning traces, including a 7K-sample training set annotated via a human-AI framework and a rigorously curated 2.2K human-annotated test set, and (3) developing a two-stage training strategy that combines stepwise risk analysis and contrastive learning to enhance robustness. Experiments show that ReasoningShield achieves state-of-the-art performance, outperforming task-specific tools like LlamaGuard-4 by 35.6% and general-purpose commercial models like GPT-4o by 15.8% on benchmarks, while also generalizing effectively across diverse reasoning paradigms, tasks, and unseen scenarios. All resources are released at https://github.com/CosmosYi/ReasoningShield.

Ancheng Xu, Di Yang, Renhao Li et al.

Traditional in-person psychological counseling remains primarily niche, often chosen by individuals with psychological issues, while online automated counseling offers a potential solution for those hesitant to seek help due to feelings of shame. Cognitive Behavioral Therapy (CBT) is an essential and widely used approach in psychological counseling. The advent of large language models (LLMs) and agent technology enables automatic CBT diagnosis and treatment. However, current LLM-based CBT systems use agents with a fixed structure, limiting their self-optimization capabilities, or providing hollow, unhelpful suggestions due to redundant response patterns. In this work, we utilize Quora-like and YiXinLi single-round consultation models to build a general agent framework that generates high-quality responses for single-turn psychological consultation scenarios. We use a bilingual dataset to evaluate the quality of single-response consultations generated by each framework. Then, we incorporate dynamic routing and supervisory mechanisms inspired by real psychological counseling to construct a CBT-oriented autonomous multi-agent framework, demonstrating its general applicability. Experimental results indicate that AutoCBT can provide higher-quality automated psychological counseling services.

Yupeng Qi, Ziyu Lyu, Min Yang et al.

As large language models (LLMs) are increasingly applied across various domains, enhancing safety while maintaining the helpfulness of LLMs has become a critical challenge. Recent studies solve this problem through safety-constrained online preference optimization or safety-constrained offline preference optimization. However, the safety-constrained online methods often suffer from excessive safety, which might reduce helpfulness, while the safety-constrained offline methods perform poorly in adaptively balancing safety and helpfulness. To address these limitations, we propose MidPO, a \textbf{\underline{Mi}}xture of Experts (MoE) framework for safety-helpfulness \textbf{\underline{d}}ual \textbf{\underline{P}}reference \textbf{\underline{O}}ptimization. Firstly, MidPO devises single-preference enhanced direct preference optimization approach to transform the base model into two independent experts, termed safety and helpfulness experts, and fine-tunes the two independent experts for optimal safety or helpfulness performance. Secondly, to achieve an effective balance between safety and helpfulness, MidPO incorporates the two experts into the MoE framework and designs a dynamic routing mechanism to allocate contributions from each expert adaptively. We conduct quantitative and qualitative experiments on three popular datasets to demonstrate the proposed MidPO significantly outperforms state-of-the-art approaches in both safety and helpfulness. The code and models will be released.

Zhishuo Zhang, Chengxiang Tan, Xueyan Zhao et al.

Cross-lingual and cross-domain knowledge alignment without sufficient external resources is a fundamental and crucial task for fusing irregular data. As the element-wise fusion process aiming to discover equivalent objects from different knowledge graphs (KGs), entity alignment (EA) has been attracting great interest from industry and academic research recent years. Most of existing EA methods usually explore the correlation between entities and relations through neighbor nodes, structural information and external resources. However, the complex intrinsic interactions among triple elements and role information are rarely modeled in these methods, which may lead to the inadequate illustration for triple. In addition, external resources are usually unavailable in some scenarios especially cross-lingual and cross-domain applications, which reflects the little scalability of these methods. To tackle the above insufficiency, a novel universal EA framework (OTIEA) based on ontology pair and role enhancement mechanism via triple-aware attention is proposed in this paper without introducing external resources. Specifically, an ontology-enhanced triple encoder is designed via mining intrinsic correlations and ontology pair information instead of independent elements. In addition, the EA-oriented representations can be obtained in triple-aware entity decoder by fusing role diversity. Finally, a bidirectional iterative alignment strategy is deployed to expand seed entity pairs. The experimental results on three real-world datasets show that our framework achieves a competitive performance compared with baselines.