❌

Normal view

Received β€” 27 May 2026 ⏭

SimuWoB: Simulating Real-World Mobile Apps for Fast and Faithful GUI Agent Benchmarking

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Guohong Liu Β· Jialei Ye Β· Pengzhi Gao Β· Wei Liu Β· Jian Luan Β· Yunxin Liu Β· Yuanchun Li
26 May 2026 at 12:00
arXiv:2605.25160v1 Announce Type: new Abstract: Mobile GUI agents powered by large language models have progressed rapidly, creating urgent needs for realistic and comprehensive evaluation. Existing benchmarks prioritize reproducibility but are often limited to open-source apps or file-operation tasks for the difficulty of constructing rewards on real applications, leaving a gap between benchmark settings and real-world usage. Moreover, most benchmarks focus on basic grounding and navigation, with limited coverage of complex, long-horizon interactions. To address these limitations, we introduce SimuWoB, a fully synthetic benchmark for mobile GUI agents with 120 challenging tasks spanning diverse types and difficulty levels. We build a robust virtual environment generation framework that synthesizes high-fidelity tasks and environments, and automatically provides valid rewards for each task. Each environment is deployed as a backend-free webpage accessible via URL, enabling efficient and reproducible evaluation. We conduct comprehensive experiments on several state-of-the-art mobile GUI agents. The average success rate is only 27.92%, dropping to 17.82% on long-horizon tasks, which reveals substantial weaknesses in current agents under complex scenarios. Evaluation result comparison with real-world sample tasks demonstrate that agent assessments based on our synthetic environment generalize well. We further provide diagnostic insights across key capability dimensions and discuss implications for future mobile GUI agent development.
Received β€” 4 April 2026 ⏭

A distinct plasma lipidomic signature and multi-omics network in depression of polycystic ovary syndrome

βœ‡(Multiomics OR Omics) AND (Lung OR gastric OR Hepatocellular)
By: Furong Yan Β· Ming Sui Β· Hongzhi Gao Β· Yifei Liu Β· Liying Yu
2 April 2026 at 18:00

J Pharm Biomed Anal. 2026 Mar 29;276:117486. doi: 10.1016/j.jpba.2026.117486. Online ahead of print.

ABSTRACT

Patients with polycystic ovary syndrome (PCOS) are at an elevated risk of depression, yet the underlying mechanisms remain elusive. Emerging evidence implicates the gut-brain axis and systemic lipid homeostasis alterations as potential key contributors. We profiled untargeted plasma lipidomes of PCOS patients with and without comorbid depression (PCOS-DP) and integrated these data with our prior gut microbial and host transcriptomic datasets to construct multi-omics interaction networks. The causal role of the candidate gut microbial was preliminary explored in a germ-free PCOS mouse model using fecal microbiota transplantation, followed by behavioral phenotyping and ELISA-based protein quantification. We identified a distinct plasma lipidomic signature differentiating PCOS-DP from PCOS alone, characterized primarily by the downregulation of 26 lipid species. Most of these altered lipids were triacylglycerols (TAGs) enriched with FA18:1 and FA18:2, whose levels correlated with coagulation dysfunction. Multi-omics network analysis revealed significant interconnections between depression-associated gut microbiota (including Bacteroides eggerthii), specific altered lipids such as TAG (60:12/FA22:6), and host genes involved in inflammation (e.g., IL22, NLRP7), metabolism, and neural processes. Animal validation demonstrated that B. eggerthii colonization in PCOS mice specifically exacerbated anhedonia and hyperlocomotion, alongside modulating plasma IL-22 expression, suggesting its context-dependent neurobehavioral effect role. This study delineates a TAG-downregulated lipid signature with diagnostic potential and reveals a novel "gut microbiota-lipid-host gene" interaction network underpinning PCOS-DP, with B. eggerthii as a key microbial modulator of neurobehavioral phenotypes in the context of PCOS. These findings provide new pathophysiological insights and highlights potential diagnostic biomarkers for PCOS-DP.

PMID:41924769 | DOI:10.1016/j.jpba.2026.117486

Received β€” 2 April 2026 ⏭

GUIDE: Resolving Domain Bias in GUI Agents through Real-Time Web Video Retrieval and Plug-and-Play Annotation

βœ‡cs.AI, q-bio.NC updates on arXiv.org
By: Rui Xie Β· Zhi Gao Β· Chenrui Shi Β· Zirui Shang Β· Lu Chen Β· Qing Li
1 April 2026 at 12:00
arXiv:2603.26266v2 Announce Type: replace Abstract: Large vision-language models have endowed GUI agents with strong general capabilities for interface understanding and interaction. However, due to insufficient exposure to domain-specific software operation data during training, these agents exhibit significant domain bias - they lack familiarity with the specific operation workflows (planning) and UI element layouts (grounding) of particular applications, limiting their real-world task performance. In this paper, we present GUIDE (GUI Unbiasing via Instructional-Video Driven Expertise), a training-free, plug-and-play framework that resolves GUI agent domain bias by autonomously acquiring domain-specific expertise from web tutorial videos through a retrieval-augmented automated annotation pipeline. GUIDE introduces two key innovations. First, a subtitle-driven Video-RAG pipeline unlocks video semantics through subtitle analysis, performing progressive three-stage retrieval - domain classification, topic extraction, and relevance matching - to identify task-relevant tutorial videos. Second, a fully automated annotation pipeline built on an inverse dynamics paradigm feeds consecutive keyframes enhanced with UI element detection into VLMs, inferring the required planning and grounding knowledge that are injected into the agent's corresponding modules to address both manifestations of domain bias. Extensive experiments on OSWorld demonstrate GUIDE's generality as a plug-and-play component for both multi-agent systems and single-model agents. It consistently yields over 5% improvements and reduces execution steps - without modifying any model parameters or architecture - validating GUIDE as an architecture-agnostic enhancement to bridge GUI agent domain bias.
There are no more articles
❌