OptoAgent is a multi‑agent framework called SightSentinel that transforms a classroom wall display into a continuous vision screening tool. It embeds short, calibrated optotype probes into ordinary educational content, and eight specialized agents analyze each encounter to recover viewing distance, recognition accuracy, gaze stability, and other metrics. A quality agent filters out poor data, a longitudinal agent builds a child‑specific baseline, and an orchestrator produces a Vision Concern Score that is routed through a safety gate to avoid diagnosis or prescription. "whyItMatters":"The system aims to detect vision problems earlier and more reliably than periodic school screenings by leveraging everyday classroom interactions and automated, low‑cost data collection."
arXiv:2609.36838v1 Announce Type: cross Abstract: Visual agents solve problems by interleaving reasoning with image operations, and on-policy distillation (OPD) provides guidance from a strong teache...
VeriCUA is a platform-agnostic framework that treats verification of computer-using agents as dynamic evidence gathering and assessment. It constructs a task rubric from available context, then selects GUI or code checks to resolve unmet requirements and combines their findings under explicit evidence rules. Experiments demonstrate that VeriCUA achieves higher agreement with reference labels than baseline methods, especially on hard offline cases and in rejecting false completion claims online.
The paper introduces Mind2Web-Injection, a benchmark of 9,954 instruction‑screenshot pairs designed to evaluate whether vision‑language models (VLMs) use visual evidence when judging conflicts between on‑screen text and user instructions. Across six VLMs, models with similar overall precision can differ dramatically in Evidence‑Aligned Detection (EAD), revealing that verdict‑only metrics miss critical grounding failures. The authors also propose two training‑free interventions—ReadGate and CmdCompare—to improve evidence grounding and instruction‑command consistency, and argue for separate reporting of verdict correctness, evidence localization, and counterfactual responsiveness.
arXiv:2608. 10827v1 Announce Type: cross Abstract: Medical visual agents can use tools to inspect images and retrieve external knowledge, but indiscriminate tool use may introduce noisy or misleading evidence.
arXiv:2607. 24770v1 Announce Type: new Abstract: Procedural tasks such as furniture assembly and home repair impose substantial cognitive demands because users must interpret instructions, track task progress, reason about spatial state, and recover from errors while performing physical actions.
VeriCUA is a platform-agnostic framework that verifies computer-using agents by treating verification as dynamic evidence gathering and assessment. It constructs a task rubric from available context, then selects GUI or code checks to resolve unmet requirements, combining findings under explicit evidence rules. Experiments demonstrate higher agreement with reference labels than baselines, especially on hard offline cases and in rejecting false completion claims online.
arXiv:2609.36201v1 Announce Type: cross Abstract: Computer-use agents (CUAs), while capable of completing computer tasks in everyday and professional workflows, can cause unintended harm even under b...
arXiv:2607. 04334v1 Announce Type: new Abstract: Multimodal GUI agents read an interface through two redundant channels: the rendered pixels of a screenshot and a serialized structure such as a DOM or accessibility tree.
arXiv:2605.31351v2 Announce Type: replace-cross Abstract: AI-based Visually Impaired Assistance (VIA) remains challenging, largely due to the high cost of human evaluation. The VLM-as-a-Judge paradig...
RSIgen is a task‑centric reciprocal self‑improvement framework for agentic visual generation that coordinates learning allocation, supervision refresh, and recursive curriculum construction. It employs Dynamic Responsibility Routing (DRR) to match interventions on evidence, inputs, and actions, selecting learning recipients and building compatible training targets. After generator updates, RSIgen re‑evaluates action utility with fixed branch inputs to refresh agent supervision, then constructs the next curriculum, yielding measurable gains on benchmarks and real‑user evaluations.
arXiv:2608. 05573v1 Announce Type: new Abstract: LLM agents increasingly execute long-horizon tasks through tool use and environment interaction, shifting evaluation from final-response scoring to verification of complete executions.
arXiv:2609.22582v1 Announce Type: new Abstract: End-to-end and vision-language-action (VLA) driving policies are compared by leaderboard rank, but a rank reports an outcome, not the behaviour behind...