❌

Normal view

Received — 8 April 2026 ⏭

Agile Deliberation: Concept Deliberation for Subjective Visual Classification

✇cs.AI, q-bio.NC updates on arXiv.org
By: Leijie Wang · Otilia Stretcu · Wei Qiao · Thomas Denby · Krishnamurthy Viswanathan · Enming Luo · Chun-Ta Lu · Tushar Dogra · Ranjay Krishna · Ariel Fuxman
7 April 2026 at 12:00
arXiv:2512.10821v2 Announce Type: replace Abstract: From content moderation to content curation, applications requiring vision classifiers for visual concepts are rapidly expanding. Existing human-in-the-loop approaches typically assume users begin with a clear, stable concept understanding to be able to provide high-quality supervision. In reality, users often start with a vague idea and must iteratively refine it through "concept deliberation", a practice we uncovered through structured interviews with content moderation experts. We operationalize the common strategies in deliberation used by real content moderators into a human-in-the-loop framework called "Agile Deliberation" that explicitly supports evolving and subjective concepts. The system supports users in defining the concept for themselves by exposing them to borderline cases. The system does this with two deliberation stages: (1) concept scoping, which decomposes the initial concept into a structured hierarchy of sub-concepts, and (2) concept iteration, which surfaces semantically borderline examples for user reflection and feedback to iteratively align an image classifier with the user's evolving intent. Since concept deliberation is inherently subjective and interactive, we painstakingly evaluate the framework through 18 user sessions, each 1.5h long, rather than standard benchmarking datasets. We find that Agile Deliberation achieves 7.5% higher F1 scores than automated decomposition baselines and more than 3% higher than manual deliberation, while participants reported clearer conceptual understanding and lower cognitive effort.
Received — 4 April 2026 ⏭

Efficient Reasoning with Balanced Thinking

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yulin Li · Tengyao Tu · Li Ding · Junjie Wang · Huiling Zhen · Yixin Chen · Yong Li · Zhuotao Tian
3 April 2026 at 12:00
arXiv:2603.12372v3 Announce Type: replace Abstract: Large Reasoning Models (LRMs) have shown remarkable reasoning capabilities, yet they often suffer from overthinking, expending redundant computational steps on simple problems, or underthinking, failing to explore sufficient reasoning paths despite inherent capabilities. These issues lead to inefficiencies and potential inaccuracies, limiting practical deployment in resource-constrained settings. Existing methods to mitigate overthinking, such as suppressing reflective keywords or adjusting reasoning length, may inadvertently induce underthinking, compromising accuracy. Therefore, we propose ReBalance, a training-free framework that achieves efficient reasoning with balanced thinking. ReBalance leverages confidence as a continuous indicator of reasoning dynamics, identifying overthinking through high confidence variance and underthinking via consistent overconfidence. By aggregating hidden states from a small-scale dataset into reasoning mode prototypes, we compute a steering vector to guide LRMs' reasoning trajectories. A dynamic control function modulates this vector's strength and direction based on real-time confidence, pruning redundancy during overthinking, and promoting exploration during underthinking. Extensive experiments conducted on four models ranging from 0.5B to 32B, and across nine benchmarks in math reasoning, general question answering, and coding tasks demonstrate that ReBalance effectively reduces output redundancy while improving accuracy, offering a general, training-free, and plug-and-play strategy for efficient and robust LRM deployment. Project page and code are available at https://rebalance-ai.github.io .

DIVER: A Multi-Stage Approach for Reasoning-intensive Information Retrieval

✇cs.AI, q-bio.NC updates on arXiv.org
By: Duolin Sun · Meixiu Long · Dan Yang · Junjie Wang · Yecheng Luo · Yue Shen · Jian Wang · Hualei Zhou · Chunxiao Guo · Peng Wei · Jiahai Wang · Jinjie Gu
3 April 2026 at 12:00
arXiv:2508.07995v5 Announce Type: replace-cross Abstract: Retrieval-augmented generation has achieved strong performance on knowledge-intensive tasks where query-document relevance can be identified through direct lexical or semantic matches. However, many real-world queries involve abstract reasoning, analogical thinking, or multi-step inference, which existing retrievers often struggle to capture. To address this challenge, we present DIVER, a retrieval pipeline designed for reasoning-intensive information retrieval. It consists of four components. The document preprocessing stage enhances readability and preserves content by cleaning noisy texts and segmenting long documents. The query expansion stage leverages large language models to iteratively refine user queries with explicit reasoning and evidence from retrieved documents. The retrieval stage employs a model fine-tuned on synthetic data spanning medical and mathematical domains, along with hard negatives, enabling effective handling of reasoning-intensive queries. Finally, the reranking stage combines pointwise and listwise strategies to produce both fine-grained and globally consistent rankings. On the BRIGHT benchmark, DIVER achieves state-of-the-art nDCG@10 scores of 46.8 overall and 31.9 on original queries, consistently outperforming competitive reasoning-aware models. These results demonstrate the effectiveness of reasoning-aware retrieval strategies in complex real-world tasks.

Contextual Distributionally Robust Optimization with Causal and Continuous Structure: An Interpretable and Tractable Approach

✇cs.AI, q-bio.NC updates on arXiv.org
By: Fenglin Zhang · Jie Wang
3 April 2026 at 12:00
arXiv:2601.11016v2 Announce Type: replace-cross Abstract: In this paper, we introduce a framework for contextual distributionally robust optimization (DRO) that considers the causal and continuous structure of the underlying distribution by developing interpretable and tractable decision rules that prescribe decisions using covariates. We first introduce the causal Sinkhorn discrepancy (CSD), an entropy-regularized causal Wasserstein distance that encourages continuous transport plans while preserving the causal consistency. We then formulate a contextual DRO model with a CSD-based ambiguity set, termed Causal Sinkhorn DRO (Causal-SDRO), and derive its strong dual reformulation where the worst-case distribution is characterized as a mixture of Gibbs distributions. To solve the corresponding infinite-dimensional policy optimization, we propose the Soft Regression Forest (SRF) decision rule, which approximates optimal policies within arbitrary measurable function spaces. The SRF preserves the interpretability of classical decision trees while being fully parametric, differentiable, and Lipschitz smooth, enabling intrinsic interpretation from both global and local perspectives. To solve the Causal-SDRO with parametric decision rules, we develop an efficient stochastic compositional gradient algorithm that converges to an $\varepsilon$-stationary point at a rate of $O(\varepsilon^{-4})$, matching the convergence rate of standard stochastic gradient descent. Finally, we validate our method through numerical experiments on synthetic and real-world datasets, demonstrating its superior performance and interpretability.
Received — 2 April 2026 ⏭

IMPACT: Influence Modeling for Open-Set Time Series Anomaly Detection

✇cs.AI, q-bio.NC updates on arXiv.org
By: Xiaohui Zhou · Yijie Wang · Hongzuo Xu · Weixuan Liang · Xiaoli Li · Guansong Pang
1 April 2026 at 12:00
arXiv:2603.29183v1 Announce Type: cross Abstract: Open-set anomaly detection (OSAD) is an emerging paradigm designed to utilize limited labeled data from anomaly classes seen in training to identify both seen and unseen anomalies during testing. Current approaches rely on simple augmentation methods to generate pseudo anomalies that replicate unseen anomalies. Despite being promising in image data, these methods are found to be ineffective in time series data due to the failure to preserve its sequential nature, resulting in trivial or unrealistic anomaly patterns. They are further plagued when the training data is contaminated with unlabeled anomalies. This work introduces $\textbf{IMPACT}$, a novel framework that leverages $\underline{\textbf{i}}$nfluence $\underline{\textbf{m}}$odeling for o$\underline{\textbf{p}}$en-set time series $\underline{\textbf{a}}$nomaly dete$\underline{\textbf{ct}}$ion, to tackle these challenges. The key insight is to $\textbf{i)}$ learn an influence function that can accurately estimate the impact of individual training samples on the modeling, and then $\textbf{ii)}$ leverage these influence scores to generate semantically divergent yet realistic unseen anomalies for time series while repurposing high-influential samples as supervised anomalies for anomaly decontamination. Extensive experiments show that IMPACT significantly outperforms existing state-of-the-art methods, showing superior accuracy under varying OSAD settings and contamination rates.

UniAI-GraphRAG: Synergizing Ontology-Guided Extraction, Multi-Dimensional Clustering, and Dual-Channel Fusion for Robust Multi-Hop Reasoning

✇cs.AI, q-bio.NC updates on arXiv.org
By: Jie Wang · Honghua Huang · Xi Ge · Jianhui Su · Wen Liu · Shiguo Lian
1 April 2026 at 12:00
arXiv:2603.25152v2 Announce Type: replace Abstract: Retrieval-Augmented Generation (RAG) systems face significant challenges in complex reasoning, multi-hop queries, and domain-specific QA. While existing GraphRAG frameworks have made progress in structural knowledge organization, they still have limitations in cross-industry adaptability, community report integrity, and retrieval performance. This paper proposes UniAI-GraphRAG, an enhanced framework built upon open-source GraphRAG. The framework introduces three core innovations: (1) Ontology-Guided Knowledge Extraction that uses predefined Schema to guide LLMs in accurately identifying domain-specific entities and relations; (2) Multi-Dimensional Community Clustering Strategy that improves community completeness through alignment completion, attribute-based clustering, and multi-hop relationship clustering; (3) Dual-Channel Graph Retrieval Fusion that balances QA accuracy and performance through hybrid graph and community retrieval. Evaluation results on MultiHopRAG benchmark show that UniAI-GraphRAG outperforms mainstream open source solutions (e.g.LightRAG) in comprehensive F1 scores, particularly in inference and temporal queries. The code is available at https://github.com/UnicomAI/wanwu/tree/main/rag/rag_open_source/rag_core/graph.

Correction: The multifunctional RNA helicase DDX39A drives glioblastoma progression by modulating WISP1 alternative splicing that induces an immunosuppressive macrophage polarization

✇Oncogene - Issue - nature.com science feeds
By: Yan Zhang · Zhiwei Xue · Naibin Zhang · Yuehua Zhu · YanZhao Wu · Meilin Lv · Zhihan Zhang · Feiyu Mu · Wenchen Xing · Ziyi Tang · Chunjie Wang · Zhiyi Xue · Wenjing Zhou · Xiaofei Liu · Xingang Li · Rolf Bjerkvig · Bin Huang · Mingzhi Han · Jian Wang · Donghai Wang
1 April 2026 at 08:00

Oncogene, Published online: 01 April 2026; doi:10.1038/s41388-026-03756-2

Correction: The multifunctional RNA helicase DDX39A drives glioblastoma progression by modulating WISP1 alternative splicing that induces an immunosuppressive macrophage polarization

STAT3-mediated transactivation of NOVA2 promotes lung adenocarcinoma metastasis by splicing SMAD4

✇Oncogene - Issue - nature.com science feeds
By: Shengjie Wang · Xin Tong · Runfeng Sun · Xinyi Xia · Donglai Chen · Zhao Wang · Hao Shi · Chenzhuo Wu · Xiaoxiao Guo · Die Hu · Ersuo Jin · Hong-Tao Zhang
1 April 2026 at 08:00

Oncogene, Published online: 01 April 2026; doi:10.1038/s41388-026-03752-6

STAT3-mediated transactivation of NOVA2 promotes lung adenocarcinoma metastasis by splicing SMAD4

Nanoscale transfer-printed full-colour ultrahigh-resolution quantum dot LEDs

✇Nature - Issue - nature.com science feeds
By: Lihua Lin · Jie Wang · Hailong Hu · Haolin Luo · Yanbin Liu · Xingjie Yang · Jingnan Su · De’er Li · Zhongwei Xu · Chengyu Luo · Yongshen Yu · Tailiang Guo · Fushan Li
1 April 2026 at 08:00

Nature, Published online: 01 April 2026; doi:10.1038/s41586-026-10333-w

A dual-action force dynamics strategy using a hard silicon template as a nanoimprinting stamp combined with inverted transfer printing is described for the manufacture of high-performance full-colour ultrahigh-resolution quantum dot light-emitting diodes (LEDs) for active-matrix displays, while revealing electric-field reconstruction in nanoscale arrays and introducing dielectric matching to mitigate field concentration and performance degradation.
  • ✇MRD
  • Circulating Tumor DNA in Cholangiocarcinoma: A Precision Oncology Roadmap Haixing Wei · Jie Wang · Qing Wu · Mengbin Qin
    Cancer Manag Res. 2026 Feb 6;18:574678. doi: 10.2147/CMAR.S574678. eCollection 2026.ABSTRACTCholangiocarcinoma (CCA) is a rare but aggressive malignancy with a rising global incidence and few therapeutic options for advanced disease. In recent decades, precision oncology for CCA has advanced rapidly, particularly through the development of targeted therapies for patients with actionable genetic alterations. These therapies have markedly prolonged survival and improved other clinical outcomes amo
     
  • 🏷️
  • ↗️

Circulating Tumor DNA in Cholangiocarcinoma: A Precision Oncology Roadmap

✇MRD
By: Haixing Wei · Jie Wang · Qing Wu · Mengbin Qin
26 March 2026 at 18:00

Cancer Manag Res. 2026 Feb 6;18:574678. doi: 10.2147/CMAR.S574678. eCollection 2026.

ABSTRACT

Cholangiocarcinoma (CCA) is a rare but aggressive malignancy with a rising global incidence and few therapeutic options for advanced disease. In recent decades, precision oncology for CCA has advanced rapidly, particularly through the development of targeted therapies for patients with actionable genetic alterations. These therapies have markedly prolonged survival and improved other clinical outcomes among patients with unresectable, advanced CCA. The implementation of precision oncology largely depends on detecting genetic mutations to guide patient selection and treatment, using tumor tissue biopsies or liquid biopsies, including circulating tumor DNA (ctDNA) from blood or bile. As a minimally invasive biomarker, ctDNA shows great promise for transforming the clinical management of CCA. This review provides a comprehensive overview of the roles of ctDNA in CCA, including early detection, prognostic stratification, minimal residual disease assessment, recurrence monitoring, therapeutic target identification, and treatment response evaluation. A synthesis of existing studies indicates that bile-derived ctDNA shows superior sensitivity compared with blood-based ctDNA in capturing the genetic profiles and heterogeneity of CCA. We also propose an integrative framework that illustrates how ctDNA profiling can inform diagnosis, treatment, and surveillance across the disease continuum. Because research on ctDNA in CCA remains in its infancy, we discuss current challenges and outline future directions for translating these findings into clinical practice. Collectively, the evidence positions ctDNA-particularly bile-derived ctDNA-as a dynamic tool for real-time genomic profiling, sensitive residual disease detection, and therapy monitoring. This integrative framework provides a roadmap for translating these capabilities into clinical practice, with the potential to enable earlier, more personalized interventions and improve outcomes for patients with CCA.

PMID:41883993 | PMC:PMC13012645 | DOI:10.2147/CMAR.S574678

Received — 26 March 2026 ⏭

WIST: Web-Grounded Iterative Self-Play Tree for Domain-Targeted Reasoning Improvement

✇cs.AI, q-bio.NC updates on arXiv.org
By: Fangyuan Li · Pengfei Li · Shijie Wang · Junqi Gao · Jianxing Liu · Biqing Qi · Yuqiang Li
25 March 2026 at 12:00
arXiv:2603.22352v1 Announce Type: cross Abstract: Recent progress in reinforcement learning with verifiable rewards (RLVR) offers a practical path to self-improvement of language models, but existing methods face a key trade-off: endogenous self-play can drift over iterations, while corpus-grounded approaches rely on curated data environments. We present \textbf{WIST}, a \textbf{W}eb-grounded \textbf{I}terative \textbf{S}elf-play \textbf{T}ree framework for domain-targeted reasoning improvement that learns directly from the open web without requiring any pre-arranged domain corpus. WIST incrementally expands a domain tree for exploration, and retrieves and cleans path-consistent web corpus to construct a controllable training environment. It then performs Challenger--Solver self-play with verifiable rewards, and feeds learnability signals back to update node posteriors and guide subsequent exploration through an adaptive curriculum. Across four backbones, WIST consistently improves over the base models and typically outperforms both purely endogenous self-evolution and corpus-grounded self-play baselines, with the Overall gains reaching \textbf{+9.8} (\textit{Qwen3-4B-Base}) and \textbf{+9.7} (\textit{OctoThinker-8B}). WIST is also domain-steerable, improving \textit{Qwen3-8B-Base} by \textbf{+14.79} in medicine and \textit{Qwen3-4B-Base} by \textbf{+5.28} on PhyBench. Ablations further confirm the importance of WIST's key components for stable open-web learning. Our Code is available at https://github.com/lfy-123/WIST.

MARS: toward more efficient multi-agent collaboration for LLM reasoning

✇cs.AI, q-bio.NC updates on arXiv.org
By: Xiao Wang · Jia Wang · Yijie Wang · Pengtao Dang · Sha Cao · Chi Zhang
25 March 2026 at 12:00
arXiv:2509.20502v2 Announce Type: replace-cross Abstract: Large language models (LLMs) have achieved impressive results in natural language understanding, yet their reasoning capabilities remain limited when operating as single agents. Multi-Agent Debate (MAD) has been proposed to address this limitation by enabling collaborative reasoning among multiple models in a round-table debate manner. While effective, MAD introduces substantial computational overhead due to the number of agents involved and the frequent communication required. In this paper, we propose MARS (Multi-Agent Review System), a role-based collaboration framework inspired by the review process. In MARS, an author agent generates an initial solution, reviewer agents provide decisions and comments independently, and a meta-reviewer integrates the feedback to make the final decision and guide further revision. This design enhances reasoning quality while avoiding costly reviewer-to-reviewer interactions, thereby controlling token consumption and inference time. We compared MARS with both MAD and other state-of-the-art reasoning strategies across multiple benchmarks. Extensive experiments with different LLMs show that MARS matches the accuracy of MAD while reducing both token usage and inference time by approximately 50\%. Code is available at https://github.com/xwang97/MARS.

MOON2.0: Dynamic Modality-balanced Multimodal Representation Learning for E-commerce Product Understanding

✇cs.AI, q-bio.NC updates on arXiv.org
By: Zhanheng Nie · Chenghan Fu · Daoze Zhang · Junxian Wu · Wanxian Guan · Pengjie Wang · Jian Xu · Bo Zheng
25 March 2026 at 12:00
arXiv:2511.12449v2 Announce Type: replace-cross Abstract: Recent Multimodal Large Language Models (MLLMs) have significantly advanced e-commerce product understanding. However, they still face three challenges: (i) the modality imbalance induced by modality mixed training; (ii) underutilization of the intrinsic alignment relationships among visual and textual information within a product; and (iii) limited handling of noise in e-commerce multimodal data. To address these, we propose MOON2.0, a dynamic modality-balanced MultimOdal representation learning framework for e-commerce prOduct uNderstanding. It comprises: (1) a Modality-driven Mixture-of-Experts (MoE) that adaptively processes input samples by their modality composition, enabling Multimodal Joint Learning to mitigate the modality imbalance; (2) a Dual-level Alignment method to better leverage semantic alignment properties inside individual products; and (3) an MLLM-based Image-text Co-augmentation strategy that integrates textual enrichment with visual expansion, coupled with Dynamic Sample Filtering to improve training data quality. We further release MBE2.0, a co-augmented Multimodal representation Benchmark for E-commerce representation learning and evaluation at https://huggingface.co/datasets/ZHNie/MBE2.0. Experiments show that MOON2.0 delivers state-of-the-art zero-shot performance on MBE2.0 and multiple public datasets. Furthermore, attention-based heatmap visualization provides qualitative evidence of improved multimodal alignment of MOON2.0.
Received — 17 March 2026 ⏭

Efficient Reasoning with Balanced Thinking

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yulin Li · Tengyao Tu · Li Ding · Junjie Wang · Huiling Zhen · Yixin Chen · Yong Li · Zhuotao Tian
16 March 2026 at 12:00
arXiv:2603.12372v1 Announce Type: new Abstract: Large Reasoning Models (LRMs) have shown remarkable reasoning capabilities, yet they often suffer from overthinking, expending redundant computational steps on simple problems, or underthinking, failing to explore sufficient reasoning paths despite inherent capabilities. These issues lead to inefficiencies and potential inaccuracies, limiting practical deployment in resource-constrained settings. Existing methods to mitigate overthinking, such as suppressing reflective keywords or adjusting reasoning length, may inadvertently induce underthinking, compromising accuracy. Therefore, we propose ReBalance, a training-free framework that achieves efficient reasoning with balanced thinking. ReBalance leverages confidence as a continuous indicator of reasoning dynamics, identifying overthinking through high confidence variance and underthinking via consistent overconfidence. By aggregating hidden states from a small-scale dataset into reasoning mode prototypes, we compute a steering vector to guide LRMs' reasoning trajectories. A dynamic control function modulates this vector's strength and direction based on real-time confidence, pruning redundancy during overthinking, and promoting exploration during underthinking. Extensive experiments conducted on four models ranging from 0.5B to 32B, and across nine benchmarks in math reasoning, general question answering, and coding tasks demonstrate that ReBalance effectively reduces output redundancy while improving accuracy, offering a general, training-free, and plug-and-play strategy for efficient and robust LRM deployment. Code is available at https://github.com/yu-lin-li/ReBalance .
Received — 11 March 2026 ⏭

Chart Deep Research in LVLMs via Parallel Relative Policy Optimization

✇cs.AI, q-bio.NC updates on arXiv.org
By: Jiajin Tang · Gaoyang · Wenjie Wang · Sibei Yang · Xing Chen
10 March 2026 at 12:00
arXiv:2603.06677v1 Announce Type: cross Abstract: With the rapid advancement of data science, charts have evolved from simple numerical presentation tools to essential instruments for insight discovery and decision-making support. However, current chart data intelligence exhibits significant limitations in deep research capabilities, with existing methods predominantly addressing shallow tasks such as visual recognition or factual question-answering, rather than the complex reasoning and high-level data analysis that deep research requires. This limitation stems from two primary technical bottlenecks: at the training level, existing post-training techniques exhibit deficiencies in handling multi-dimensional reward signal interference and heterogeneous data gradient conflicts, preventing models from achieving balanced development across multiple capability dimensions; at the evaluation level, current methods remain limited to factual retrieval and basic computation, failing to assess end-to-end analytic reasoning and other deep research capabilities. To address the training challenge, we propose PRPO, which performs parallel optimization across reward dimensions and capability partitioning across data types, effectively disentangling conflicts between heterogeneous data and multi-dimensional reward signals while ensuring optimization stability. For the evaluation challenge, we construct MCDR-Bench based on the ``error uniqueness principle," transforming subjective generation assessment into objective error identification through controllable error injection, enabling quantifiable evaluation of deep research capabilities. Experimental validation confirms that the proposed PRPO and MCDR-Bench jointly establish a unified framework that systematically advances chart deep research through enhanced collaborative training and objective evaluation.

UnSCAR: Universal, Scalable, Controllable, and Adaptable Image Restoration

✇cs.AI, q-bio.NC updates on arXiv.org
By: Debabrata Mandal · Soumitri Chattopadhyay · Yujie Wang · Marc Niethammer · Praneeth Chakravarthula
10 March 2026 at 12:00
arXiv:2603.07406v1 Announce Type: cross Abstract: Universal image restoration aims to recover clean images from arbitrary real-world degradations using a single inference model. Despite significant progress, existing all-in-one restoration networks do not scale to multiple degradations. As the number of degradations increases, training becomes unstable, models grow excessively large, and performance drops across both seen and unseen domains. In this work, we show that scaling universal restoration is fundamentally limited by interference across degradations during joint learning, leading to catastrophic task forgetting. To address this challenge, we introduce a unified inference pipeline with a multi-branch mixture-of-experts architecture that decomposes restoration knowledge across specialized task-adaptable experts. Our approach enables scalable learning (over sixteen degradations), adapts and generalizes robustly to unseen domains, and supports user-controllable restoration across degradations. Beyond achieving superior performance across benchmarks, this work establishes a new design paradigm for scalable and controllable universal image restoration.

EasyInsert: A Data-Efficient and Generalizable Insertion Policy

✇cs.AI, q-bio.NC updates on arXiv.org
By: Guanghe Li · Junming Zhao · Shengjie Wang · Yang Gao
10 March 2026 at 12:00
arXiv:2505.16187v2 Announce Type: replace-cross Abstract: Robotic insertion is a highly challenging task that requires exceptional precision in cluttered environments. Existing methods often have poor generalization capabilities. They typically function in restricted and structured environments, and frequently fail when the plug and socket are far apart, when the scene is densely cluttered, or when handling novel objects. They also rely on strong assumptions such as access to CAD models or a digital twin in simulation. To address these limitations, we propose EasyInsert. Inspired by human intuition, it formulates insertion as a delta-pose regression problem, which unlocks an efficient, highly scalable data collection pipeline with minimal human labor to train an end-to-end visual policy. During execution, the visual policy predicts the relative pose between plug and socket to drive a multi-phase, coarse-to-fine insertion process. EasyInsert demonstrates strong zero-shot generalization capability for unseen objects in cluttered environments, robustly handling cases with significant initial pose deviations. In real-world experiments, by leveraging just 1 hour of human teleoperation data to bootstrap a large-scale automated data collection process, EasyInsert achieves an over 90% success rate in zero-shot insertion for 13 out of 15 unseen novel objects, including challenging objects like Type-C cables, HDMI cables, and Ethernet cables. Furthermore, requiring only a single manual reset, EasyInsert allows for fast adaptation to novel test objects through automated data collection and fine-tuning, achieving an over 90% success rate across all 15 objects.
Received — 5 March 2026 ⏭

Can Large Language Models Derive New Knowledge? A Dynamic Benchmark for Biological Knowledge Discovery

✇cs.AI, q-bio.NC updates on arXiv.org
By: Chaoqun Yang · Xinyu Lin · Shulin Li · Wenjie Wang · Ruihan Guo · Fuli Feng · Tat-Seng Chua
5 March 2026 at 13:00
arXiv:2603.03322v1 Announce Type: cross Abstract: Recent advancements in Large Language Model (LLM) agents have demonstrated remarkable potential in automatic knowledge discovery. However, rigorously evaluating an AI's capacity for knowledge discovery remains a critical challenge. Existing benchmarks predominantly rely on static datasets, leading to inevitable data contamination where models have likely seen the evaluation knowledge during training. Furthermore, the rapid release cycles of modern LLMs render static benchmarks quickly outdated, failing to assess the ability to discover truly new knowledge. To address these limitations, we propose DBench-Bio, a dynamic and fully automated benchmark designed to evaluate AI's biological knowledge discovery ability. DBench-Bio employs a three-stage pipeline: (1) data acquisition of rigorous, authoritative paper abstracts; (2) QA extraction utilizing LLMs to synthesize scientific hypothesis questions and corresponding discovery answers; and (3) QA filter to ensure quality based on relevance, clarity, and centrality. We instantiate this pipeline to construct a monthly-updated benchmark covering 12 biomedical sub-domains. Extensive evaluations of SOTA models reveal current limitations in discovering new knowledge. Our work provides the first dynamic, automatic framework for assessing the new knowledge discovery capabilities of AI systems, establishing a living, evolving resource for AI research community to catalyze the development of knowledge discovery.

Silent Sabotage During Fine-Tuning: Few-Shot Rationale Poisoning of Compact Medical LLMs

✇cs.AI, q-bio.NC updates on arXiv.org
By: Jingyuan Xie · Wenjie Wang · Ji Wu · Jiandong Gao
4 March 2026 at 13:00
arXiv:2603.02262v1 Announce Type: cross Abstract: Supervised fine-tuning (SFT) is essential for the development of medical large language models (LLMs), yet prior poisoning studies have mainly focused on the detectable backdoor attacks. We propose a novel poisoning attack targeting the reasoning process of medical LLMs during SFT. Unlike backdoor attacks, our method injects poisoned rationales into few-shot training data, leading to stealthy degradation of model performance on targeted medical topics. Results showed that knowledge overwriting was ineffective, while rationale poisoning caused significant decline on the accuracy of the target subject, as long as no correct samples of the same subject appear in the dataset. A minimum number and ratio of poisoned samples was needed to carry out an effective and stealthy attack, which was more efficient and accurate than catastrophic forgetting. We demonstrate though this study the risk of SFT-stage poisoning, hoping to spur more studies of defense in the sensitive medical domain.

Skywork-Reward-V2: Scaling Preference Data Curation via Human-AI Synergy

✇cs.AI, q-bio.NC updates on arXiv.org
By: Chris Yuhao Liu · Liang Zeng · Yuzhen Xiao · Jujie He · Jiacai Liu · Chaojie Wang · Rui Yan · Wei Shen · Fuxiang Zhang · Jiacheng Xu · Yang Liu · Yahui Zhou
4 March 2026 at 13:00
arXiv:2507.01352v3 Announce Type: replace-cross Abstract: 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 nuanced human preferences. 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 SynPref-40M, a large-scale preference dataset comprising 40 million preference pairs. 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 LLMs 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. These reward models achieve state-of-the-art performance across seven major reward model benchmarks, outperform generative reward models, and demonstrate strong downstream performance. Ablation studies confirm that effectiveness stems not only from data scale but also from high-quality curation. The Skywork-Reward-V2 series represents substantial progress in open reward models, demonstrating how human-AI curation synergy can unlock significantly higher data quality.
❌