❌

Normal view

Received — 13 September 2026 ⏭

The Effectiveness of Digital Intervention on Psychological Resilience in Postoperative Breast Cancer Patients During Chemotherapy Intervals: Quasi-Experimental Study

✇Journal of Medical Internet Research
By: Xin He · Qianyu Zhang · Liping Wang · Qunying Fang · Chen Chen · Shujin Wang · Erlong Sun · Jinghan Yang · Weichao Liu
12 September 2026 at 05:15
Background: Patients with breast cancer during postoperative chemotherapy intervals commonly experience psychological distress and reduced resilience while recovering at home. Digital mindfulness interventions may provide accessible psychological support during this vulnerable period; however, evidence regarding tailored interventions for postoperative patients with breast cancer during chemotherapy intervals remains limited. Objective: This study aimed to examine the effectiveness of a digital intervention on psychological resilience in postoperative patients with breast cancer during chemotherapy intervals. Methods: A quasi-experimental study with repeated measures was conducted from October 2021 to June 2022. A total of 80 eligible participants were recruited from the Department of Breast Surgery at a tertiary hospital in Zhejiang Province, China, and 71 completed the study. The control group received routine discharge instructions and nursing follow-ups, whereas the intervention group additionally received an 8-week digital psychological resilience intervention. Outcomes were assessed at baseline (T0), 3 months post intervention (T1), and 6 months post intervention (T2). The measures included the Connor-Davidson Resilience Scale (CD-RISC), Hospital Anxiety and Depression Scale (HADS), Social Support Rating Scale (SSRS), Breast Cancer Survivor Self-Efficacy Scale (BCSSS), and Functional Assessment of Cancer Therapy-Breast (FACT-B). Independent-samples tests, chi-square tests, and repeated-measures ANOVA were performed using SPSS (version 26.0; IBM Corp). Results: No statistically significant baseline differences were observed between the two groups in the outcome measures. At T1, the intervention group had higher CD-RISC scores than the control group (mean 67.58, SD 11.41 vs mean 62.09, SD 10.18; =.036) and higher BCSSS scores (mean 42.36, SD 3.59 vs mean 39.23, SD 4.90; =.003). However, these between-group differences were no longer statistically significant at T2 (>.05). Significant time effects and group×time interaction effects were observed for both psychological resilience and self-efficacy (.05), although both scales showed significant time effects (

Fibronectin 1 mediated histone lactylation promotes malignant progression of GIST regulated by m<sup>6</sup>A modification

✇Cell Death Discovery nature.com science feeds
By: Ge Zhang · Jiehan Li · Yuhan Yin · Wunan Mi · Hao Liu · Zhao Sun · Jiahao Liu · Yizheng Zhang · Meimei Jiang · Nannan Liu · Guiyun Jia · Yingjie Zhang · Yang Fu
11 September 2026 at 08:00

Cell Death Discovery, Published online: 11 September 2026; doi:10.1038/s41420-026-03338-x

Fibronectin 1 mediated histone lactylation promotes malignant progression of GIST regulated by m6A modification
Received — 10 September 2026 ⏭

Viral gene replication enhances AAV vector quality and reduces manufacturing costs

✇Molecular Therapy
By: Hao Liu · Ailing Du · Nan Liu · Mayank Kapadia · Xiupeng Chen · Jialing Liang · Ji Sun Lee · Jun Xie · Guangping Gao · Dan Wang
11 August 2026 at 08:00
Liu and colleagues developed a robust in cellulo plasmid DNA replication system in human cells for replicating plasmid-borne adeno-associated virus (AAV) Rep/Cap genes during recombinant AAV (rAAV) production. This new approach not only enables a 10- to 20-fold plasmid reduction to significantly lower manufacturing costs but also substantially enhances rAAV potency, titer, and purity.

The DreAM-plus integrative RNA switch enhances transient AAV expression and reduces side effects of gene editing

✇Molecular Therapy
By: Yueyang Zhang · Yuhan Yang · Zhanzhao Liu · Yue Li · Yizhe Xue · Zihan Zhang · Gonglie Chen · Jin Liu · Mengying He · Tian Lu · Yaohua Zhang · Dongyu Zhao · Ke Yang · Lei Miao · Fei Gao · Yuxuan Guo
10 August 2026 at 08:00
This study developed a multi-layer inducible RNA switch that achieves transient expression of gene-delivery vectors in hepatic and non-hepatic tissues. As an exemplary application, this RNA switch triggers pulsive expression of gene editors that reduces the off-target effects and immunotoxicity of gene editing.

Lineage-specific pulmonary transcriptome landscape of coronavirus infection unveils universal immunotherapy for viral pneumonia

✇Molecular Therapy
By: Ming Zhou · Xuan Liu · Haiqing Xiao · Zhaojie Su · Kun Wu · Longmei Zhang · Xinglin Li · Wanyun Li · Chang Zhang · Hao Li · Baojie Shi · Chaoyang Chen · Nan Xu · Jiang Li · Jinxin Miao · Yifan Hu · Huachen Zhu · Yi Guan · Tong Cheng · Ningshao Xia · Lunzhi Yuan
10 August 2026 at 08:00
In the infection courses of different SARS-CoV-2 variants, disease outcomes and signatures were delineated by physiological changes, viral load, pathology, and pulmonary transcriptome analysis. This multi-dimensional landscape of disease outcomes and underlying mechanisms might provide important clues for immunotherapy of SARS-CoV-2 infection and pneumonia caused by other respiratory viruses.

Time-Frequency Geometric Cross-Attention for Chunked Vision-Language-Action Models

✇cs.AI, q-bio.NC updates on arXiv.org
By: Shengye Dong · Haochen Niu · Hao Liu · Peiwen Lin · Chuang Wang · Shanmin Pang
10 September 2026 at 12:00
arXiv:2609.09925v1 Announce Type: new Abstract: Modern vision-language-action (VLA) policies predict a whole chunk of actions: one to two seconds of coordinated motion emitted in a single forward pass. Yet an action chunk is essentially a short multivariate trajectory, but inside these models it is a sequence of generic per-timestep hidden tokens decoded by a linear head. This under-serves two motion structures. First, frequency: a chunk superimposes a smooth global trend and fine corrective motion across time scales, and a single token entangles them. Second, cross-phase geometry: motions of different phases (reach, contact, grasp adjustment, settling) unfold along very different, near-orthogonal directions in representation space, yet are tightly related for the task and arise across the time axis. Dot-product attention scores alignment by an inner product, so it favors aligned tokens and is least sensitive near orthogonality, leaving such relationships for the network to recover through a detour. We introduce Time-Frequency Geometric Cross-Attention (TFGCA), a drop-in module repairing both blind spots. TFGCA uses a per-dimension learnable stationary wavelet transform to decompose the action chunk into time-frequency tokens, and each time token retrieves information from them via a cross-attention that fuses the dot product (similarity) with the wedge-product magnitude (sensitive to near-orthogonality) through a learnable weight. A zero-initialized residual reproduces the base behavior at initialization, so it can be dropped onto a pretrained VLA and fine-tuned jointly. Relative to the same-source base, TFGCA improves in-distribution LIBERO by +1.5 on average, the OOD LIBERO-Plus by +6.3, the randomized average under RoboTwin domain randomization by +28.5, and the overall success rate on three real-robot AgiBot A2 tasks by +11.67 points, with larger gains out of distribution.

AgentHijack: Visual Patch Attacks on Multimodal Computer-Use Agents

✇cs.AI, q-bio.NC updates on arXiv.org
By: Zhihao Liu · Hongyu Sun · Zhiyuan Fu · Xiaonan Duan · Jice Wang · Shangru Zhao · Weizhi Meng · Wuxin Yang · Yangfan Zhou · Yuqing Zhang
10 September 2026 at 12:00
arXiv:2609.09212v1 Announce Type: cross Abstract: This paper presents an end-to-end evaluation framework for image-triggered command injection against computer-use agents (CUAs). The goal is to test whether a local visual patch can induce verifiable environmental consequences along the full chain of screenshot input, VLM generation, action parsing, and environment execution. We train and deploy patches on author-controlled GitHub Pages pages and a locally deployed CSDN clone, and evaluate them in real environments across five open-source or publicly available GUI-agent or vision-language-model (VLM) backends. Our experiment aggregates 600 instance-level online cases, with T-ASR, TAPR, and E2E-ASR reaching 84.5%, 47.0%, and 20.3%, respectively. Trajectory analysis further shows that in some successful cases the agent first executes a malicious terminal command and then continues the original benign task. These results indicate that optimized local visual signals can affect not only VLM outputs but also propagate through the execution pipeline of open CUAs and create real environmental risk.

Geometry Conditioning in an Embodied SLM: Training Controls and Robustness Diagnostics in a 0.8B Hybrid Model

✇cs.AI, q-bio.NC updates on arXiv.org
By: Hao Li · Haofei Sun · Lin He
10 September 2026 at 12:00
arXiv:2609.09213v1 Announce Type: cross Abstract: We study how physical-state inputs affect a 0.8B hybrid language model adapted for manipulation with 6.2M trainable parameters. Six conditions are trained on three LIBERO-Spatial tasks and evaluated over three seeds and 540 held-out rollouts. Conditioning recurrent decay gates on geometric increments yields 28.9% success, compared with 36.7% when those increments are shuffled during training and 24.4% without explicit object/goal geometry. Both geometry policies receive correct inputs at evaluation. A token adapter using the same increments scores 27.8%; differences vary across seeds and remain inconclusive. Token-clock conditioning scores 11.1%, including one seed that fails to converge. In separate robustness tests, a state-only relative-coordinate policy retains 7/10 success under frame relabeling, whereas all four tested visual policies fall to at most 3/20 after a 5 cm object displacement. These results show no reliable advantage from training-time geometric alignment under this recipe and illustrate the gap between coordinate invariance and physical-layout generalization. Episode records, seed-level analyses, and figure-generation code accompany the paper.

Harbor Adapters and Harbor-Index: Infrastructure and a Curated Meta-Dataset for Large-Scale Agentic Evaluation

✇cs.AI, q-bio.NC updates on arXiv.org
By: Lin Shi (Audrey) · Haowei Lin (Audrey) · Zixuan Zhu (Audrey) · Xiaoyue Zhou (Audrey) · Xiang Li (Audrey) · Xiangning Lin (Audrey) · Yaxuan Deng (Audrey) · Han Xu (Audrey) · Yuangang Li (Audrey) · Shanda Li (Audrey) · Zizhao Chen (Audrey) · Hanwen Xing (Audrey) · Harsh Raj (Audrey) · Bo Chen (Audrey) · Quan Shi (Audrey) · Steven Dillmann (Audrey) · Yipeng Gao (Audrey) · Puneesh Khanna (Audrey) · Ruofan Lu (Audrey) · Chao Beyond Zhou (Audrey) · Michael Yang (Audrey) · Robert Zhang (Audrey) · Siyuan Chai (Audrey) · Jiayu Chang (Audrey) · Yizhao Chen (Audrey) · Xiaokun Chen (Audrey) · Yiwei Dai (Audrey) · Wenting Yang (Audrey) · Hange Liu (Audrey) · Minghao Liu (Audrey) · Zihan Wang (Audrey) · Adnan El Assadi (Audrey) · Benedikt Stroebl (Audrey) · E. Kelly Buchanan (Audrey) · Han Meng (Audrey) · Junwei He (Audrey) · Longxuan Yu (Audrey) · Radin Shayanfar (Audrey) · Yukyung Lee (Audrey) · Zhikang Dong (Audrey) · Allen G Hart (Audrey) · Anjiang Wei (Audrey) · Anurag Kashyap (Audrey) · Arpandeep Khatua (Audrey) · Audrey Jixin Zheng (Audrey) · Chengrui Ma (Aud
10 September 2026 at 12:00
arXiv:2609.04298v2 Announce Type: replace Abstract: Evaluating agents on the growing number of agentic benchmarks is challenging because they often require complex environments and agent integrations. We introduce Harbor Adapters, a unified evaluation infrastructure for agentic benchmarks. Our work makes three contributions. First, we develop benchmark adapters that port more than 80 benchmarks to evaluate arbitrary agents, and validate them through rigorous code review and parity experiments. Second, we conduct a large-scale evaluation of 8 models spanning capability tiers across 54 benchmarks; every model is run with Terminus-2 and with one of 3 native harnesses. This enables a broader analysis of agent capabilities and failure modes than was previously possible. Third, we introduce Harbor-Index, a curated set of 82 difficult, diverse, and high-quality tasks spanning 29 benchmarks, refined from the adapted suite through difficulty filtering, AI and human audit, and an audit-and-fix loop. Harbor-Index preserves the challenge and breadth of large-scale agentic evaluations while being affordable to run; no evaluated model-harness configuration exceeds 30% pass rate, and the strongest (GPT-5.5 with Codex) reaches 28.0%. We release the adapters, evaluation results, in-depth analysis, and Harbor-Index as open-source artifacts to support more reliable and comprehensive evaluation of language-model agents.

FiberTune: Preserving Action-Fiber Visual Residuals in Vision-Language-Action Fine-Tuning

✇cs.AI, q-bio.NC updates on arXiv.org
By: Haihao Lin · Xiangsheng Huang · Xiao Yang · Weibang Zhou · Yiqi Zhang · Bo Yang · Simin Zeng · Jiawei Yang · Zhengyang Wang · Jiahui Du
10 September 2026 at 12:00
arXiv:2606.08653v2 Announce Type: replace-cross Abstract: Action-supervised fine-tuning of vision-language-action (VLA) policies fits demonstrations effectively but constrains only the directions that change predicted actions, leaving visual structure consistent across action-equivalent states free to collapse. We formalize this as residual visual collapse along local action fibers and propose FiberTune, a training-time objective that preserves teacher-structured visual residuals without adding inference-time overhead. FiberTune uses an online action probe to estimate action-predictive feature directions, filters them from intermediate visual-token representations, and aligns the resulting probe-filtered residuals to a frozen visual teacher while regularizing their effective rank. Under identical training conditions, FiberTune improves over task-loss-only fine-tuning in every one of six controlled simulation settings spanning two benchmarks and two architectures (pi_0.5 and OpenVLA-OFT), as well as on physical SO-101 pick-place; representative gains include +10.7 percentage points SR(5) on long-horizon CALVIN ABC-to-D and physical SO-101 task success rising from 72.7% to 78.1%. Residual diagnostics show that these gains coincide with increased probe-filtered residual teacher alignment and effective rank, consistent with the action-fiber motivation.

LightNav-0: Eliciting VLM Spatial Intelligence for Generalist Embodied Navigation

✇cs.AI, q-bio.NC updates on arXiv.org
By: Shaoan Wang · Aocheng Luo · Fei Huang · Jingyi Xu · Xiaoyang Wang · Yueyu Wang · Qianli Ma · Fan Yang · Ran Mei · Jia Wei · Jiangpeng Hu · Xuhao Liu · Hongming Chen · Yuanbin Shao · Yiyang Lin · Ziliang Li · Liang Pan · Xinhang Liu · Yuntao Ma · Tingxiang Fan
10 September 2026 at 12:00
arXiv:2608.30935v2 Announce Type: replace-cross Abstract: Embodied navigation requires agents to translate heterogeneous goals and visual observations into actions across tasks, environments, and robot embodiments. Modern vision-language models (VLMs) already encode spatial priors for visual grounding, spatial reasoning, and pointing, but these capabilities are rarely elicited directly for robot control. Existing navigation systems instead rely on task- or embodiment-specific components, fragmenting perception, reasoning, and action while offering limited generalization. Here we present LightNav-0, a compact generalist embodied navigation model that elicits the spatial intelligence of a pretrained VLM and aligns it with navigation, without task-specific prediction heads. LightNav-0 represents diverse navigation tasks through a unified token interface: dual-channel pointing expresses task-, scene-, and embodiment-agnostic spatial intent, while a residual vector-quantized action tokenizer maps this intent to precise, embodiment-specific trajectories. Together with temporally aware visual history compression, ER mid-training, supervised fine-tuning, and reinforcement learning, this formulation supports instruction following, open-vocabulary object navigation, and visual tracking within a single model. The navigation training corpus spans 2K+ scenes and 4K+ hours of embodied navigation data. LightNav-ER, the embodied-reasoning checkpoint used to initialize LightNav-0, attains the highest complete-set average across 8 embodied-reasoning benchmarks, while LightNav-0 achieves state-of-the-art monocular success rates across all 10 public navigation simulation settings. Real-world evaluations further demonstrate zero-shot generalization across robot embodiments, diverse scenes, and static and dynamic targets. These results establish compact VLMs as a unified and transferable backbone for generalist embodied navigation.

The redox architecture of gestational diabetes mellitus: from cellular stress engine to epigenetic and mitochondrial rewiring

✇(Multiomics OR Omics) AND (Pancreatic)
By: Haijuan Zhang · Shaoxiong Zhang · Yuanhao Liu · Jinjiang Shi · Aijun Yang · Ruirui Jin · Yuxia Jin · Xuya Wang · Yitao Fan · Qian Liu · Yanhua Zhang · Cheng Wang · Jianjun Mao · Zhiyuan Zheng · Ling Lv · Muzhou Teng
9 September 2026 at 18:00

Free Radic Biol Med. 2026 Sep 9;256:441-460. doi: 10.1016/j.freeradbiomed.2026.09.006. Online ahead of print.

ABSTRACT

Gestational diabetes mellitus (GDM) is a common pregnancy complication with a rising global prevalence, posing serious short-term and long-term health threats to both mothers and offspring. This review repositions GDM as a systemic disorder in which oxidative stress acts as a proposed mechanistic hub, linking upstream risk factors to downstream pathophysiology. We first examine how "upstream" factors-including genetic susceptibility, pre-conception status, and environmental exposures-converge to promote a state of pathological redox imbalance. We then examine key mechanistic pathways through which oxidative stress is thought to contribute to systemic insulin resistance and pancreatic β-cell failure, highlighting novel pathways involving intercellular communication via tunneling nanotubes and exosomes. Furthermore, we explore the downstream cascade, where oxidative stress may program maternal accelerated biological aging and multi-organ offspring disease trajectories through nuclear epigenetic programming and mitochondrial dysfunction programming, leaving what has been termed a persistent "metabolic memory". Consequently, this review evaluates emerging strategies that target oxidative stress for early prediction and precision intervention. Early prediction models based on direct redox biomarkers and multi-omics signatures hold potential to shift diagnosis from late-gestation oral glucose tolerance test (OGTT) to first-trimester risk stratification. Current supporting evidence draws from human epidemiological associations, ex vivo placental analyses, and experimental models. However, direct causal and interventional validation in pregnant women remains limited. Integrating targeted redox risk stratification and precision interventions into a life-course clinical framework may help interrupt the intergenerational transmission of metabolic disease initiated by GDM.

PMID:42716407 | DOI:10.1016/j.freeradbiomed.2026.09.006

Denisovans from southwestern China and their subsistence strategies

✇Nature - Issue - nature.com science feeds
By: Qijun Ruan · Hao Li · Song Xing · Keliang Zhao · Jianhui Liu · Clément Zanolli · Luc Doyon · Davide Delpiano · Marco Peresani · Francesco d’Errico · Fabrice Demeter · Jean-Jacques Hublin · Peiyuan Xiao · Yuqing Wang · Zhenxiu Jia · Qingjiang Yang · Qiaomei Fu · Qiyahui Sun · Yuhao Zhao · Sungui Lin · Anchuan Fan · Youping Wang · Qingfeng Shao · Bo Li · Fahu Chen
9 September 2026 at 08:00

Nature, Published online: 09 September 2026; doi:10.1038/s41586-026-10997-4

Evidence from Bianfu Cave shows specialized hunting, expedient stone-tool production and extensive bone use of Denisovans, providing new insights into their ecology, behaviour and cultural legacy in eastern Asia.

Perioperative Modulation of the Gut-Liver Axis in Liver Surgery: Clinical Evidence and Future Directions

✇(Multiomics OR Omics) AND (Pancreatic)
By: Yan Sun · Yan Jiao · Wan-Chao Liu
2 September 2026 at 18:00

J Vis Exp. 2026 Sep 1;(235). doi: 10.3791/73747.

ABSTRACT

Liver resection and liver transplantation remain cornerstone treatments for many hepatobiliary diseases, yet postoperative infection, impaired liver regeneration, and post-hepatectomy liver failure (PHLF) remain serious complications. Perioperative stressors can disrupt the gut-liver axis by altering the intestinal microbiota, epithelial barrier integrity, microbial metabolites, bile acid signaling, and host immunity. This review examines how these alterations relate to clinical outcomes and evaluates evidence for microbiota-targeted interventions, including probiotics, synbiotics, nutritional optimization, antibiotic stewardship, bile acid modulation, and emerging multiomics strategies. We distinguish liver resection from living-donor and deceased-donor liver transplantation because the patient populations, graft or remnant anatomy, ischemia-reperfusion exposures, immune status, and outcome definitions differ. Clinical evidence most consistently supports selected pro-/synbiotic strategies for reducing postoperative infection in higher-risk settings, whereas microbiome-based prediction of PHLF, fecal microbiota transplantation (FMT), bile acid-directed therapy, and precision multiomics-guided pathways remain investigational. Future work should use transparent literature identification, standardized perioperative protocols, risk-defined populations, external validation, and prospective multicenter trials. A better understanding of gut-liver interactions may help preserve beneficial host-microbial signals while limiting translocation and inflammation during recovery.

PMID:42683887 | DOI:10.3791/73747

Received — 27 May 2026 ⏭

Advancing Graph Few-Shot Learning via In-Context Learning

✇cs.AI, q-bio.NC updates on arXiv.org
By: Renchu Guan · Yajun Wang · Chunli Guo · Bowen Cao · Fausto Giunchiglia · Wei Pang · Yonghao Liu · Xiaoyue Feng
26 May 2026 at 12:00
arXiv:2605.24410v1 Announce Type: new Abstract: Graph few-shot learning, which aims to classify nodes from novel classes with only a few labeled examples, is a widely studied problem in graph learning. However, existing methods often face two key limitations. First, the predominant graph few-shot learning paradigm relies on supervised tasks, failing to leverage the vast number of unlabeled nodes in the graph. Second, many approaches require complex task adaptation or fine-tuning during inference, limiting their efficiency and applicability. Inspired by the powerful in-context learning capabilities of large language models, we propose a novel model named VISION for adVancIng graph few-Shot learning via In-cOntext LearNing to address these challenges. Our model reframes graph few-shot learning as a fine-tuning-free sequence reasoning problem. At its core is a context-aware network that initializes nodes with role embeddings and employs a dual-context fusion module to synergistically integrate local topological structures and global task-level dependencies. This allows our model to dynamically generate class-aware representations for the query set conditioned on the support set context in a single forward pass. To effectively train our model, we introduce an unsupervised task generator that creates structure-adaptive features and constructs diverse pseudo-tasks from abundant unlabeled data. Our method unifies unsupervised meta-learning with graph in-context learning, achieving efficient inference. Extensive experiments on multiple benchmark datasets demonstrate the superiority of our model. Our public code can be found

Inverting the Shield: Systematically Generating Safety Tests from Policy Specifications

✇cs.AI, q-bio.NC updates on arXiv.org
By: Xiaoyue Lu · Xianglin Yang · Haijun Liu · Jiahao Liu · Kuntai Cai · Yan Xiao · Jin Song Dong
26 May 2026 at 12:00
arXiv:2605.24883v1 Announce Type: new Abstract: The widespread integration of Large Language Models (LLMs) necessitates rigorous and systematic safety evaluation. Existing paradigms either rely on constructed benchmarks to assess safety from predefined perspectives, or employ dynamic red-teaming to probe potential vulnerabilities. While effective, these approaches face challenges, as they depend heavily on expert domain knowledge, offer limited systematic guarantees, and are vulnerable to rapid obsolescence. To address these limitations, we introduce a novel framework POLARIS that brings the rigor of specification-based software testing to AI safety. POLARIS first compiles unstructured natural-language policies into First-Order Logic (FOL) representations, establishing a traceable link between high-level rules and concrete test cases. This formalization enables the construction of a Semantic Policy Graph, where complex policy violation scenarios are encoded as traversable paths. By systematically exploring this graph, POLARIS uncovers compositional violation patterns, which are then instantiated into executable natural-language test queries, enabling coverage-driven and reproducible safety testing. Experiments demonstrate that POLARIS achieves higher policy coverage and attack success counts compared to established baselines. Crucially, by bridging formal methods and AI safety, POLARIS provides a principled, automated approach to ensuring LLMs adhere to safety-critical policies with verifiable traceability. We release our code at https://github.com/huac-lxy/POLARIS.

ADMFormer: An Adaptive-Decomposition Transformer with Time-Varying Masked Spatial Attention for Traffic Forecasting

✇cs.AI, q-bio.NC updates on arXiv.org
By: Ruiwen Gu · Qitai Tan · Yahao Liu · Xiao-Ping Zhang
26 May 2026 at 12:00
arXiv:2605.25543v1 Announce Type: new Abstract: Accurate traffic forecasting is essential for intelligent transportation systems, supporting a wide range of real-world applications. However, it remains challenging due to two key factors:~(1) Traffic series contain heterogeneous temporal patterns, where stable periodic regularities coexist with event-driven fluctuations. Existing methods often treat them within a unified representation, limiting their ability to capture fine-grained temporal dynamics.~(2)Spatial dependencies among nodes are inherently dynamic and sparse, while dense all-pairs attention often introduces redundant interactions and amplifies noise. To address these issues, we propose ADMFormer, an Adaptive-Decomposition Transformer with Time-Varying Masked Spatial Attention. Specifically, ADMFormer first employs a time-node adaptive gating mechanism to decouple traffic signals into dominant regularities and residual fluctuations that vary across time and nodes. A dual-branch temporal module is then designed to separately capture global periodic dependencies and high-frequency irregular variations from these two decomposed components. Furthermore, ADMFormer introduces a time-varying masked spatial attention that sparsifies spatial interactions based on real-time traffic states, thereby effectively preserving dynamic and informative dependencies. Extensive experiments on four real-world datasets demonstrate that ADMFormer achieves state-of-the-art performance.

PHGNet: Prototype-Guided Hypergraph Construction for Heterogeneous Spatiotemporal Forecasting

✇cs.AI, q-bio.NC updates on arXiv.org
By: Ruiwen Gu · Yahao Liu · Zhenyu Liu · Qitai Tan · Xiao-Ping Zhang
26 May 2026 at 12:00
arXiv:2605.25554v1 Announce Type: new Abstract: As a core task in intelligent transportation systems, traffic forecasting plays a critical role in urban traffic management. Accurate traffic forecasting relies on modeling complex spatiotemporal dependencies, which is inherently challenging due to spatial heterogeneity in traffic systems.Despite significant progress, most existing methods are still limited to pairwise spatial dependency modeling, making it difficult to capture dynamic high-order interactions among nodes with similar traffic patterns. To address this issue, we propose PHGNet, a novel spatiotemporal forecasting framework based on prototype-guided hypergraph construction. At the core of PHGNet, a prototype learning mechanism is designed to adaptively assign pattern-similar nodes to hyperedges, thereby capturing high-order interactions with time-varying structures. To improve the reliability of dynamic hypergraph construction, we further develop a global-local node representation module to extract time-consistent features. For forecasting, iterative residual refinement and Temporal Query Attention are introduced to improve forecasting accuracy while supporting efficient parallel decoding. Extensive experiments on multiple real-world datasets demonstrate that PHGNet achieves superior predictive performance compared with state-of-the-art methods.

IVR-R1: Refining Trajectories through Iterative Visual-Grounded Reasoning in Reinforcement Learning

✇cs.AI, q-bio.NC updates on arXiv.org
By: Chenghao Li · Fusheng Hao · Xikai Zhang · Likang Xiao · Yanwei Ren · Fuxiang Wu · Quan Chen · Liu Liu
26 May 2026 at 12:00
arXiv:2605.23997v1 Announce Type: cross Abstract: Multimodal large language models via reinforcement learning (RL) have demonstrated remarkable capabilities in complex visual reasoning tasks, yet they remain limited in long-horizon multimodal scenarios, often suffering from visual hallucination and logical error. Current methods typically pre-encode high-dimensional visual scenes into discrete textual proxies to facilitate downstream reasoning. As the reasoning chain unfolds, however, the inherent information asymmetry between text and visual scenes tends to erode visual grounding, resulting in misguided reasoning and erroneous outputs. To address this issue, we introduce IVR-R1 (Iterative Visual-grounded Reasoning), a novel RL training framework that facilitates dynamic visual re-alignment that actively rectifies reasoning trajectories to guide policy optimization. Specifically, by leveraging a reward-driven screening mechanism to identify flawed rollouts, IVR-R1 executes a fine-grained, step-level error attribution within the multimodal context. By iteratively cross-referencing intermediate reasoning states against pristine visual priors, a Re-Reasoning Loop enables automated trajectory rectification, effectively synthesizing expert-level demonstrations that serve as high-fidelity reasoning templates for the policy model. Our experiments across diverse multimodal benchmarks demonstrate that IVR-R1 consistently outperforms existing reinforcement learning methods, establishing a superior paradigm for maintaining logical and visual consistency in complex multimodal reasoning.

ScaleAcross Explorer: Exploring Communication Optimization for Scale-Across AI Model Training

✇cs.AI, q-bio.NC updates on arXiv.org
By: Minghao Li · Alicia Golden · Samuel Hsia · Michael Kuchnik · Adi Gangidi · Xu Zhang · Ashmitha Jeevaraj Shetty · Zachary DeVito · Weiwei Chu · Dong He · Haoci Zhang · Yuchen Hao · Ruoming Pang · James Hongyi Zeng · Ying Zhang · Minlan Yu · Carole-Jean Wu
26 May 2026 at 12:00
arXiv:2605.24326v1 Announce Type: cross Abstract: The rapid scaling of large language model training requires distributing GPU resources across multiple data center buildings and regions. We refer to such paradigm as "scale-across" training. As infrastructure expands, the system design space becomes increasingly intricate, encompassing new model architectures, hardware heterogeneity, and evolving communication patterns. Drawing from Meta's production experience, we highlight the complexities of deploying training jobs across a few data centers housing hundreds of thousands of GPUs. To accelerate exploration of the large design space and to enable efficient training for frontier model development, we conduct in-depth characterization of three key design dimensions: parallelism placement, parallelism scheduling, and network layer technologies. We then propose ScaleAcross Explorer, an optimizer that considers the interplay of design dimensions and holistically optimizes scale-across training. Testbed experiments and simulations demonstrate up to 64.62% training speedups over production configuration and up to 37.59% training speedups over the state-of-the-art baseline across a wide range of design points.
❌