Domain Generalization (DG) seeks to develop a versatile model capable of performing effectively on unseen target domains. Notably, recent advances in pre-trained Visual Foundation Models (VFMs), such as CLIP, have demonstrated considerable potential in enhancing the generalization capabilities of deep learning models. Despite the increasing attention toward VFM-based domain prompt tuning within DG, the effective design of prompts capable of disentangling invariant features across diverse domains remains a critical challenge. In this paper, we propose addressing this challenge by leveraging the controllable and flexible language prompt of the VFM. Noting that the text modality of VFMs is naturally easier to disentangle, we introduce a novel framework for text feature-guided visual prompt tuning. This framework first automatically disentangles the text prompt using a large language model (LLM) and then learns domain-invariant visual representation guided by the disentangled text feature. However, relying solely on language to guide visual feature disentanglement has limitations, as visual features can sometimes be too complex or nuanced to be fully captured by descriptive text. To address this, we introduce Worst Explicit Representation Alignment (WERA), which extends text-guided visual prompts by incorporating an additional set of abstract prompts. These prompts enhance source domain diversity through stylized image augmentations, while alignment constraints ensure that visual representations remain consistent across both the original and augmented distributions. Experiments conducted on major DG datasets, including PACS, VLCS, OfficeHome, DomainNet, and TerraInc, demonstrate that our proposed method outperforms state-of-the-art DG methods.

Graphical Abstracts (GAs) play a crucial role in visually conveying the key findings of scientific papers. While recent research has increasingly incorporated visual materials such as Figure 1 as de facto GAs, their potential to enhance scientific communication remains largely unexplored. Moreover, designing effective GAs requires advanced visualization skills, creating a barrier to their widespread adoption. To tackle these challenges, we introduce SciGA-145k, a large-scale dataset comprising approximately 145,000 scientific papers and 1.14 million figures, explicitly designed for supporting GA selection and recommendation as well as facilitating research in automated GA generation. As a preliminary step toward GA design support, we define two tasks: 1) Intra-GA recommendation, which identifies figures within a given paper that are well-suited to serve as GAs, and 2) Inter-GA recommendation, which retrieves GAs from other papers to inspire the creation of new GAs. We provide reasonable baseline models for these tasks. Furthermore, we propose Confidence Adjusted top-1 ground truth Ratio (CAR), a novel recommendation metric that offers a fine-grained analysis of model behavior. CAR addresses limitations in traditional ranking-based metrics by considering cases where multiple figures within a paper, beyond the explicitly labeled GA, may also serve as GAs. By unifying these tasks and metrics, our SciGA-145k establishes a foundation for advancing visual scientific communication while contributing to the development of AI for Science.

Process Reinforcement Learning~(PRL) has demonstrated considerable potential in enhancing the reasoning capabilities of Large Language Models~(LLMs). However, introducing additional process reward models incurs substantial computational overhead, and there is no unified theoretical framework for process-level advantage estimation. To bridge this gap, we propose \textbf{S}elf-Guided \textbf{P}rocess \textbf{R}eward \textbf{O}ptimization~(\textbf{SPRO}), a novel framework that enables process-aware RL through two key innovations: (1) we first theoretically demonstrate that process rewards can be derived intrinsically from the policy model itself, and (2) we introduce well-defined cumulative process rewards and \textbf{M}asked \textbf{S}tep \textbf{A}dvantage (\textbf{MSA}), which facilitates rigorous step-wise action advantage estimation within shared-prompt sampling groups. Our experimental results demonstrate that SPRO outperforms vaniila GRPO with 3.4x higher training efficiency and a 17.5\% test accuracy improvement. Furthermore, SPRO maintains a stable and elevated policy entropy throughout training while reducing the average response length by approximately $1/3$, evidencing sufficient exploration and prevention of reward hacking. Notably, SPRO incurs no additional computational overhead compared to outcome-supervised RL methods such as GRPO, which benefit industrial implementation.

This paper explores how older adults, particularly aging migrants in urban China, can engage AI-assisted co-creation to express personal narratives that are often fragmented, underrepresented, or difficult to verbalize. Through a pilot workshop combining oral storytelling and the symbolic reconstruction of Hanzi, participants shared memories of migration and recreated new character forms using Xiaozhuan glyphs, suggested by the Large Language Model (LLM), together with physical materials. Supported by human facilitation and a soft AI presence, participants transformed lived experience into visual and tactile expressions without requiring digital literacy. This approach offers new perspectives on human-AI collaboration and aging by repositioning AI not as a content producer but as a supportive mechanism, and by supporting narrative agency within sociotechnical systems.

Despite the critical role of reward models (RMs) in reinforcement learning from human feedback (RLHF), current state-of-the-art open RMs perform poorly on most existing evaluation benchmarks, failing to capture the spectrum of nuanced and sophisticated human preferences. Even approaches that incorporate advanced training techniques have not yielded meaningful performance improvements. We hypothesize that this brittleness stems primarily from limitations in preference datasets, which are often narrowly scoped, synthetically labeled, or lack rigorous quality control. To address these challenges, we present a large-scale preference dataset comprising 40 million preference pairs, named SynPref-40M. To enable data curation at scale, we design a human-AI synergistic two-stage pipeline that leverages the complementary strengths of human annotation quality and AI scalability. In this pipeline, humans provide verified annotations, while large language models perform automatic curation based on human guidance. Training on this preference mixture, we introduce Skywork-Reward-V2, a suite of eight reward models ranging from 0.6B to 8B parameters, trained on a carefully curated subset of 26 million preference pairs from SynPref-40M. We demonstrate that Skywork-Reward-V2 is versatile across a wide range of capabilities, including alignment with human preferences, objective correctness, safety, resistance to stylistic biases, and best-of-N scaling, achieving state-of-the-art performance across seven major reward model benchmarks. Ablation studies confirm that the effectiveness of our approach stems not only from data scale but also from high-quality curation. The Skywork-Reward-V2 series represents substantial progress in open reward models, highlighting the untapped potential of existing preference datasets and demonstrating how human-AI curation synergy can unlock significantly higher data quality.

The paradigm of large language models in natural language processing (NLP) has also shown promise in modeling biological languages, including proteins, RNA, and DNA. Both the auto-regressive generation paradigm and evaluation metrics have been transferred from NLP to biological sequence modeling. However, the intrinsic structural correlations in natural and biological languages differ fundamentally. Therefore, we revisit the notion of language in biological systems to better understand how NLP successes can be effectively translated to biological domains. By treating the 3D structure of biomolecules as the semantic content of a sentence and accounting for the strong correlations between residues or bases, we highlight the importance of structural evaluation and demonstrate the applicability of the auto-regressive paradigm in biological language modeling. Code can be found at \href{https://github.com/zjuKeLiu/RiFold}{github.com/zjuKeLiu/RiFold}

Large language models (LLMs) excel in complex tasks through advanced prompting techniques like Chain-of-Thought (CoT) and Tree-of-Thought (ToT), but their reliance on manually crafted, task-specific prompts limits adaptability and efficiency. We introduce Mixture of Reasoning (MoR), a training framework that embeds diverse reasoning strategies into LLMs for autonomous, task-adaptive reasoning without external prompt engineering. MoR has two phases: Thought Generation, creating reasoning chain templates with models like GPT-4o, and SFT Dataset Construction, pairing templates with benchmark datasets for supervised fine-tuning. Our experiments show that MoR significantly enhances performance, with MoR150 achieving 0.730 (2.2% improvement) using CoT prompting and 0.734 (13.5% improvement) compared to baselines. MoR eliminates the need for task-specific prompts, offering a generalizable solution for robust reasoning across diverse tasks.

Every publicly traded U.S. company files an annual 10-K report containing critical insights into financial health and risk. We propose Tiny eXplainable Risk Assessor (TinyXRA), a lightweight and explainable transformer-based model that automatically assesses company risk from these reports. Unlike prior work that relies solely on the standard deviation of excess returns (adjusted for the Fama-French model), which indiscriminately penalizes both upside and downside risk, TinyXRA incorporates skewness, kurtosis, and the Sortino ratio for more comprehensive risk assessment. We leverage TinyBERT as our encoder to efficiently process lengthy financial documents, coupled with a novel dynamic, attention-based word cloud mechanism that provides intuitive risk visualization while filtering irrelevant terms. This lightweight design ensures scalable deployment across diverse computing environments with real-time processing capabilities for thousands of financial documents which is essential for production systems with constrained computational resources. We employ triplet loss for risk quartile classification, improving over pairwise loss approaches in existing literature by capturing both the direction and magnitude of risk differences. Our TinyXRA achieves state-of-the-art predictive accuracy across seven test years on a dataset spanning 2013-2024, while providing transparent and interpretable risk assessments. We conduct comprehensive ablation studies to evaluate our contributions and assess model explanations both quantitatively by systematically removing highly attended words and sentences, and qualitatively by examining explanation coherence. The paper concludes with findings, practical implications, limitations, and future research directions.

Turn-taking prediction models are essential components in spoken dialogue systems and conversational robots. Recent approaches leverage transformer-based architectures to predict speech activity continuously and in real-time. In this study, we propose a novel model that enables turn-taking prediction to be dynamically controlled via textual prompts. This approach allows intuitive and explicit control through instructions such as "faster" or "calmer" adapting dynamically to conversational partners and contexts. The proposed model builds upon a transformer-based voice activity projection (VAP) model, incorporating textual prompt embeddings into both channel-wise transformers and a cross-channel transformer. We evaluated the feasibility of our approach using over 950 hours of human-human spoken dialogue data. Since textual prompt data for the proposed approach was not available in existing datasets, we utilized a large language model (LLM) to generate synthetic prompt sentences. Experimental results demonstrated that the proposed model improved prediction accuracy and effectively varied turn-taking timing behaviors according to the textual prompts.

The recent development and wider accessibility of LLMs have spurred discussions about how they can be used in survey research, including classifying open-ended survey responses. Due to their linguistic capacities, it is possible that LLMs are an efficient alternative to time-consuming manual coding and the pre-training of supervised machine learning models. As most existing research on this topic has focused on English-language responses relating to non-complex topics or on single LLMs, it is unclear whether its findings generalize and how the quality of these classifications compares to established methods. In this study, we investigate to what extent different LLMs can be used to code open-ended survey responses in other contexts, using German data on reasons for survey participation as an example. We compare several state-of-the-art LLMs and several prompting approaches, and evaluate the LLMs' performance by using human expert codings. Overall performance differs greatly between LLMs, and only a fine-tuned LLM achieves satisfactory levels of predictive performance. Performance differences between prompting approaches are conditional on the LLM used. Finally, LLMs' unequal classification performance across different categories of reasons for survey participation results in different categorical distributions when not using fine-tuning. We discuss the implications of these findings, both for methodological research on coding open-ended responses and for their substantive analysis, and for practitioners processing or substantively analyzing such data. Finally, we highlight the many trade-offs researchers need to consider when choosing automated methods for open-ended response classification in the age of LLMs. In doing so, our study contributes to the growing body of research about the conditions under which LLMs can be efficiently, accurately, and reliably leveraged in survey research.