❌

Normal view

Received — 4 April 2026 ⏭

RIFT: A RubrIc Failure Mode Taxonomy and Automated Diagnostics

✇cs.AI, q-bio.NC updates on arXiv.org
By: Zhengyang Qi · Charles Dickens · Derek Pham · Amanda Dsouza · Armin Parchami · Frederic Sala · Paroma Varma
3 April 2026 at 12:00
arXiv:2604.01375v1 Announce Type: new Abstract: Rubric-based evaluation is widely used in LLM benchmarks and training pipelines for open-ended, less verifiable tasks. While prior work has demonstrated the effectiveness of rubrics using downstream signals such as reinforcement learning outcomes, there remains no principled way to diagnose rubric quality issues from such aggregated or downstream signals alone. To address this gap, we introduce RIFT: RubrIc Failure mode Taxonomy, a taxonomy for systematically characterizing failure modes in rubric composition and design. RIFT consists of eight failure modes organized into three high-level categories: Reliability Failures, Content Validity Failures, and Consequential Validity Failures. RIFT is developed using grounded theory by iteratively annotating rubrics drawn from five diverse benchmarks spanning general instruction following, code generation, creative writing, and expert-level deep research, until no new failure modes are identified. We evaluate the consistency of the taxonomy by measuring agreement among independent human annotators, observing fair agreement overall (87% pairwise agreement and 0.64 average Cohen's kappa). Finally, to support scalable diagnosis, we propose automated rubric quality metrics and show that they align with human failure-mode annotations, achieving up to 0.86 F1.
Received — 26 March 2026 ⏭

Energy-Aware Reinforcement Learning for Robotic Manipulation of Articulated Components in Infrastructure Operation and Maintenance

✇cs.AI, q-bio.NC updates on arXiv.org
By: Xiaowen Tao · Yinuo Wang · Haitao Ding · Yuanyang Qi · Ziyu Song
25 March 2026 at 12:00
arXiv:2602.12288v3 Announce Type: replace-cross Abstract: With the growth of intelligent civil infrastructure and smart cities, operation and maintenance (O&M) increasingly requires safe, efficient, and energy-conscious robotic manipulation of articulated components, including access doors, service drawers, and pipeline valves. However, existing robotic approaches either focus primarily on grasping or target object-specific articulated manipulation, and they rarely incorporate explicit actuation energy into multi-objective optimisation, which limits their scalability and suitability for long-term deployment in real O&M settings. Therefore, this paper proposes an articulation-agnostic and energy-aware reinforcement learning framework for robotic manipulation in intelligent infrastructure O&M. The method combines part-guided 3D perception, weighted point sampling, and PointNet-based encoding to obtain a compact geometric representation that generalises across heterogeneous articulated objects. Manipulation is formulated as a Constrained Markov Decision Process (CMDP), in which actuation energy is explicitly modelled and regulated via a Lagrangian-based constrained Soft Actor-Critic scheme. The policy is trained end-to-end under this CMDP formulation, enabling effective articulated-object operation while satisfying a long-horizon energy budget. Experiments on representative O&M tasks demonstrate 16%-30% reductions in energy consumption, 16%-32% fewer steps to success, and consistently high success rates, indicating a scalable and sustainable solution for infrastructure O&M manipulation.

PAK4 functions as an immune suppressor by reprogramming the phosphatidylcholine metabolism of CD8 + T cells within the glioblastoma tumor microenvironment

✇Oncogene - Issue - nature.com science feeds
By: Lin Yao · Heyang Gao · Zuopeng Su · Guozheng Zhao · Liyan Tang · Shuo Feng · Yuting Ma · Xiao Zhang · Ming Feng · Jialiang Qian · Yanyan Li · Ting Sun · Jiangang Liu · Hao Wang · Youxin Zhou
23 March 2026 at 08:00

Oncogene, Published online: 23 March 2026; doi:10.1038/s41388-026-03734-8

PAK4 functions as an immune suppressor by reprogramming the phosphatidylcholine metabolism of CD8 + T cells within the glioblastoma tumor microenvironment

Mechanisms of Xinwei Tang in stress-induced gastric dysmotility: evidence from rat and In Vitro models

✇Omics in Gastric
By: Man Tian · Honghui Xiao · Yingbing Mei · Xinyu Zhang · Zhengliang Qi
19 March 2026 at 18:00

In Vitro Cell Dev Biol Anim. 2026 Mar 18. doi: 10.1007/s11626-026-01151-5. Online ahead of print.

ABSTRACT

Stress is a key trigger of gastric dysmotility, partly via mitochondrial dysfunction and disordered gut-brain hormonal signaling. Xinwei Tang (XWT) is a multi-herb formula used empirically for upper gastrointestinal symptoms, but its mechanisms remain unclear. This study aimed to determine whether XWT alleviates water-immersion restraint stress (WIRS)-induced gastric dysmotility and to delineate underlying mitochondrial and metabolic pathways using integrated in vivo, in vitro and multi-omics approaches. Male rats underwent 7-d WIRS and received vehicle, domperidone (3 mg/kg) or XWT (3, 6, 12 g/kg). Gastric emptying, serum motilin/gastrin, oxidative stress indices and PINK1/Parkin-LC3/p62 proteins were assessed, and H₂O₂-injured GES-1 cells were treated with XWT-medicated serum. Gastric antra from MOD and XWT-H rats were analyzed by RNA-seq and DIA proteomics (n = 3/group). WIRS reduced gastric emptying by roughly half and lowered motilin/gastrin, increased ROS/MDA and disrupted PINK1/Parkin-LC3/p62 profiles; XWT dose-dependently reversed these changes, with XWT-H approximating domperidone. Omics revealed XWT-associated downregulation of inflammatory/protease and acute-phase genes/proteins and enrichment of oxidative phosphorylation, tricarboxylic-acid cycle and other metabolic pathways, without global activation of canonical autophagy/mitophagy gene sets. These preclinical data indicate that XWT ameliorates stress-induced gastric dysmotility via mitochondria- and metabolism-centred protection with selective tuning of mitophagy-related proteins.

PMID:41851413 | DOI:10.1007/s11626-026-01151-5

Received — 17 March 2026 ⏭

When Drafts Evolve: Speculative Decoding Meets Online Learning

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yu-Yang Qian · Hao-Cong Wu · Yichao Fu · Hao Zhang · Peng Zhao
16 March 2026 at 12:00
arXiv:2603.12617v1 Announce Type: cross Abstract: Speculative decoding has emerged as a widely adopted paradigm for accelerating large language model inference, where a lightweight draft model rapidly generates candidate tokens that are then verified in parallel by a larger target model. However, due to limited model capacity, drafts often struggle to approximate the target distribution, resulting in shorter acceptance lengths and diminished speedup. A key yet under-explored observation is that speculative decoding inherently provides verification feedback that quantifies the deviation between the draft and target models at no additional cost. This process naturally forms an iterative "draft commits-feedback provides-draft adapts" evolving loop, which precisely matches the online learning paradigm. Motivated by this connection, we propose OnlineSpec, a unified framework that systematically leverages interactive feedback to continuously evolve draft models. Grounded in dynamic regret minimization, we establish a formal link between online learning performance and speculative system's acceleration rate, and develop novel algorithms via modern online learning techniques, including optimistic online learning that adaptively reuses historical gradients as predictive update hints, and online ensemble learning that dynamically maintains multiple draft models. Our algorithms are equipped with theoretical justifications and improved acceleration rates, achieving up to 24% speedup over seven benchmarks and three foundation models.
Received — 11 March 2026 ⏭

FuzzingRL: Reinforcement Fuzz-Testing for Revealing VLM Failures

✇cs.AI, q-bio.NC updates on arXiv.org
By: Jiajun Xu · Jiageng Mao · Ang Qi · Weiduo Yuan · Alexander Romanus · Helen Xia · Vitor Campagnolo Guizilini · Yue Wang
10 March 2026 at 12:00
arXiv:2603.06600v1 Announce Type: cross Abstract: Vision Language Models (VLMs) are prone to errors, and identifying where these errors occur is critical for ensuring the reliability and safety of AI systems. In this paper, we propose an approach that automatically generates questions designed to deliberately induce incorrect responses from VLMs, thereby revealing their vulnerabilities. The core of this approach lies in fuzz testing and reinforcement finetuning: we transform a single input query into a large set of diverse variants through vision and language fuzzing. Based on the fuzzing outcomes, the question generator is further instructed by adversarial reinforcement fine-tuning to produce increasingly challenging queries that trigger model failures. With this approach, we can consistently drive down a target VLM's answer accuracy -- for example, the accuracy of Qwen2.5-VL-32B on our generated questions drops from 86.58\% to 65.53\% in four RL iterations. Moreover, a fuzzing policy trained against a single target VLM transfers to multiple other VLMs, producing challenging queries that degrade their performance as well.

Enhancing Consistency of Werewolf AI through Dialogue Summarization and Persona Information

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yoshiki Tanaka · Takumasa Kaneko · Hiroki Onozeki · Natsumi Ezure · Ryuichi Uehara · Zhiyang Qi · Tomoya Higuchi · Ryutaro Asahara · Michimasa Inaba
10 March 2026 at 12:00
arXiv:2603.07111v1 Announce Type: cross Abstract: The Werewolf Game is a communication game where players' reasoning and discussion skills are essential. In this study, we present a Werewolf AI agent developed for the AIWolfDial 2024 shared task, co-hosted with the 17th INLG. In recent years, large language models like ChatGPT have garnered attention for their exceptional response generation and reasoning capabilities. We thus develop the LLM-based agents for the Werewolf Game. This study aims to enhance the consistency of the agent's utterances by utilizing dialogue summaries generated by LLMs and manually designed personas and utterance examples. By analyzing self-match game logs, we demonstrate that the agent's utterances are contextually consistent and that the character, including tone, is maintained throughout the game.
Received — 5 March 2026 ⏭

MIRAGE: Knowledge Graph-Guided Cross-Cohort MRI Synthesis for Alzheimer's Disease Prediction

✇cs.AI, q-bio.NC updates on arXiv.org
By: Guanchen Wu · Zhe Huang · Yuzhang Xie · Runze Yan · Akul Chopra · Deqiang Qiu · Xiao Hu · Fei Wang · Carl Yang
4 March 2026 at 13:00
arXiv:2603.02434v1 Announce Type: cross Abstract: Reliable Alzheimer's disease (AD) diagnosis increasingly relies on multimodal assessments combining structural Magnetic Resonance Imaging (MRI) and Electronic Health Records (EHR). However, deploying these models is bottlenecked by modality missingness, as MRI scans are expensive and frequently unavailable in many patient cohorts. Furthermore, synthesizing de novo 3D anatomical scans from sparse, high-dimensional tabular records is technically challenging and poses severe clinical risks. To address this, we introduce MIRAGE, a novel framework that reframes the missing-MRI problem as an anatomy-guided cross-modal latent distillation task. First, MIRAGE leverages a Biomedical Knowledge Graph (KG) and Graph Attention Networks to map heterogeneous EHR variables into a unified embedding space that can be propagated from cohorts with real MRIs to cohorts without them. To bridge the semantic gap and enforce physical spatial awareness, we employ a frozen pre-trained 3D U-Net decoder strictly as an auxiliary regularization engine. Supported by a novel cohort-aggregated skip feature compensation strategy, this decoder acts as a rigorous structural penalty, forcing 1D latent representations to encode biologically plausible, macro-level pathological semantics. By exclusively utilizing this distilled "diagnostic-surrogate" representation during inference, MIRAGE completely bypasses computationally expensive 3D voxel reconstruction. Experiments demonstrate that our framework successfully bridges the missing-modality gap, improving the AD classification rate by 13% compared to unimodal baselines in cohorts without real MRIs.

MedXIAOHE: A Comprehensive Recipe for Building Medical MLLMs

✇cs.AI, q-bio.NC updates on arXiv.org
By: Baorong Shi · Bo Cui · Boyuan Jiang · Deli Yu · Fang Qian · Haihua Yang · Huichao Wang · Jiale Chen · Jianfei Pan · Jieqiong Cao · Jinghao Lin · Kai Wu · Lin Yang · Shengsheng Yao · Tao Chen · Xiaojun Xiao · Xiaozhong Ji · Xu Wang · Yijun He · Zhixiong Yang
4 March 2026 at 13:00
arXiv:2602.12705v3 Announce Type: replace-cross Abstract: We present MedXIAOHE, a medical vision-language foundation model designed to advance general-purpose medical understanding and reasoning in real-world clinical applications. MedXIAOHE achieves state-of-the-art performance across diverse medical benchmarks and surpasses leading closed-source multimodal systems on multiple capabilities. To achieve this, we propose an entity-aware continual pretraining framework that organizes heterogeneous medical corpora to broaden knowledge coverage and reduce long-tail gaps (e.g., rare diseases). For medical expert-level reasoning and interaction, MedXIAOHE incorporates diverse medical reasoning patterns via reinforcement learning and tool-augmented agentic training, enabling multi-step diagnostic reasoning with verifiable decision traces. To improve reliability in real-world use, MedXIAOHE integrates user-preference rubrics, evidence-grounded reasoning, and low-hallucination long-form report generation, with improved adherence to medical instructions. We release this report to document our practical design choices, scaling insights, and evaluation framework, hoping to inspire further research.
Received — 17 February 2026 ⏭

Deep Dense Exploration for LLM Reinforcement Learning via Pivot-Driven Resampling

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yiran Guo · Zhongjian Qiao · Yingqi Xie · Jie Liu · Dan Ye · Ruiqing Zhang · Shuang Qiu · Lijie Xu
17 February 2026 at 13:00
arXiv:2602.14169v1 Announce Type: cross Abstract: Effective exploration is a key challenge in reinforcement learning for large language models: discovering high-quality trajectories within a limited sampling budget from the vast natural language sequence space. Existing methods face notable limitations: GRPO samples exclusively from the root, saturating high-probability trajectories while leaving deep, error-prone states under-explored. Tree-based methods blindly disperse budgets across trivial or unrecoverable states, causing sampling dilution that fails to uncover rare correct suffixes and destabilizes local baselines. To address this, we propose Deep Dense Exploration (DDE), a strategy that focuses exploration on $\textit{pivots}$-deep, recoverable states within unsuccessful trajectories. We instantiate DDE with DEEP-GRPO, which introduces three key innovations: (1) a lightweight data-driven utility function that automatically balances recoverability and depth bias to identify pivot states; (2) local dense resampling at each pivot to increase the probability of discovering correct subsequent trajectories; and (3) a dual-stream optimization objective that decouples global policy learning from local corrective updates. Experiments on mathematical reasoning benchmarks demonstrate that our method consistently outperforms GRPO, tree-based methods, and other strong baselines.

MedXIAOHE: A Comprehensive Recipe for Building Medical MLLMs

✇cs.AI, q-bio.NC updates on arXiv.org
By: Baorong Shi · Bo Cui · Boyuan Jiang · Deli Yu · Fang Qian · Haihua Yang · Huichao Wang · Jiale Chen · Jianfei Pan · Jieqiong Cao · Jinghao Lin · Kai Wu · Lin Yang · Shengsheng Yao · Tao Chen · Xiaojun Xiao · Xiaozhong Ji · Xu Wang · Yijun He · Zhixiong Yang
17 February 2026 at 13:00
arXiv:2602.12705v2 Announce Type: replace-cross Abstract: We present MedXIAOHE, a medical vision-language foundation model designed to advance general-purpose medical understanding and reasoning in real-world clinical applications. MedXIAOHE achieves state-of-the-art performance across diverse medical benchmarks and surpasses leading closed-source multimodal systems on multiple capabilities. To achieve this, we propose an entity-aware continual pretraining framework that organizes heterogeneous medical corpora to broaden knowledge coverage and reduce long-tail gaps (e.g., rare diseases). For medical expert-level reasoning and interaction, MedXIAOHE incorporates diverse medical reasoning patterns via reinforcement learning and tool-augmented agentic training, enabling multi-step diagnostic reasoning with verifiable decision traces. To improve reliability in real-world use, MedXIAOHE integrates user-preference rubrics, evidence-grounded reasoning, and low-hallucination long-form report generation, with improved adherence to medical instructions. We release this report to document our practical design choices, scaling insights, and evaluation framework, hoping to inspire further research.
There are no more articles
❌