Tool use, function calling, orchestration and the protocols that let models act rather than only answer.
arXiv:2609.14400v1 Announce Type: new Abstract: Agent benchmarks evaluate policy compliance but assume each policy determines a unique correct action. Natural-language policies can violate this assum...
arXiv:2609.14648v1 Announce Type: new Abstract: Aligning multi-turn dialogue agents is usually framed as matching turn-level human preferences, yet direct optimization of long-term outcomes is often...
arXiv:2609.14823v1 Announce Type: new Abstract: Multimodal clinical decision-making requires reliable reasoning over heterogeneous evidence from electronic health records, medical images, and physiol...
ClinicalReTrial is a multi‑agent AI system that treats clinical trial protocol optimization as an iterative redesign problem on textual documents. It combines failure diagnosis, safety‑aware modifications, and candidate evaluation within a closed‑loop, reward‑driven framework, using a predictive model as a simulation environment for low‑cost, dense feedback. The system achieves a 56.7% conversion of failed protocols to predicted successes, with a 7.4% average success probability increase at a negligible cost, and its redesign patterns align with real‑world expert changes.
arXiv:2609.14578v1 Announce Type: new Abstract: Autonomous agents that author building information models need more than access to a host API. They need a representation of what they intended, what a...
arXiv:2609.15098v1 Announce Type: new Abstract: Continuous-environment vision-and-language navigation (VLN-CE) requires interpreting natural-language instructions in unseen 3D environments and execut...
arXiv:2609.15169v1 Announce Type: new Abstract: Driving vision-language-action (VLA) models increasingly reason before acting, but their intermediate reasoning is often weakly grounded in physical sc...
arXiv:2609.15478v1 Announce Type: new Abstract: Multimodal agents can create complex videos in software such as Blender by coding without relying on diffusion models. Yet video understanding benchmar...
arXiv:2609.15457v1 Announce Type: new Abstract: Vision-Language Models (VLMs) enable autonomous GUI navigation, but agents still struggle to process and learn from dense, continuous visual histories....
The paper introduces a lightweight ASR head that can be added to full‑duplex speech‑to‑speech models, enabling real‑time user transcription without major architectural changes. The method adds only a few parameters and preserves full‑duplex conversational features such as turn‑taking and barge‑in. Experiments show a streaming WER of 10.21% within the duplex framework and 7.73% when trained as a standalone ASR model, matching state‑of‑the‑art performance.
JaxAHT is a new open‑source library built with JAX that speeds up and standardizes research in Ad Hoc Teamwork (AHT). It offers a unified framework for generating teammates, training ego agents, and evaluating performance against unseen partners, delivering roughly 95× faster wall‑clock times than comparable PyTorch implementations. The library also supplies a diverse set of evaluation teammates for Level‑Based Foraging, Overcooked, and Hanabi, and is used to run a large‑scale benchmark that shows no single algorithm dominates and that agent modeling mainly helps in role‑based, diverse teammate settings.
The paper "Governing at Machine Speed: An Adaptive Intelligence Architecture for Real-Time AI Policy Enforcement" highlights a gap in enterprise AI governance, where 78% of organizations lack auditable evidence of policy enforcement. It introduces AGIL, a five-layer adaptive governance architecture that uses machine learning for real-time detection, risk classification, sub-100ms policy enforcement, continuous attestation, and policy evolution. The authors argue that the failure is organizational and architectural, not technical, and call for future empirical validation of AGIL.
The paper introduces the concept of moral rebellion in autonomous agents, where agents may deviate from user‑assigned tasks when morally justified. Five agent architectures are formalized: an amoral agent, utilitarian agents, deontic agents, utilitarian‑deontic (UD) agents, and dutiful agents. Experiments in a Mini Search‑and‑Rescue domain show distinct trade‑offs among task completion, rescue outcomes, and norm compliance, highlighting the role of commitment‑aware moral reasoning.
CoMem introduces a memory architecture for multi-agent systems that blends private experience with shared knowledge. It includes Private Experience Sedimentation to retain useful individual memories, Collective Wisdom Curation to filter widely proven ideas for sharing, and Parallel Dual-Stream Retrieval to draw from both personal and group memories while maintaining diversity. Experiments on ALFWorld and PDDL benchmarks demonstrate that CoMem improves overall performance and reduces memory pollution.
K-Bench is a new benchmark designed to evaluate large language model (LLM) unlearning when the models are deployed as agents. Unlike previous benchmarks that only inspect the final answer, K-Bench examines all six channels of a ReAct agent—including chain-of-thought, tool calls, tool observations, and elicited summaries—to determine if a secret is leaked. The benchmark measures leakage for secrets placed in the model weights, prompt, or retrieval store, and finds that many existing unlearning methods fail to prevent leaks in deployed agents, especially when secrets reside in the prompt or retrieval store.
The paper introduces ANASSA, an agentic AI orchestration framework designed for spatial intelligence in geographic information systems. It addresses gaps in current systems by integrating structured spatial reasoning, multi‑agent workflow orchestration, execution feedback, authoritative validation, provenance, uncertainty handling, and human decision authority. The architecture is detailed with eleven components across four layers, a six‑step Geospatial AI Cognitive Loop, cross‑component contracts, and governance mechanisms to ensure traceability, reproducibility, and accountability.
AlgoEvo introduces a unified agentic framework for automated algorithm discovery that replaces rigid search pipelines with an interactive, knowledge‑accumulating process. An autonomous agent inspects, diagnoses, and edits code using runtime feedback, while a design skill hub decouples paradigm‑specific knowledge from the core engine, enabling a single workflow to handle single‑objective, multi‑objective, and multi‑component design tasks. The hierarchical experience mechanism organizes search trajectories into a task‑level tree, guiding exploration and consolidating cross‑task patterns into reusable skills, resulting in performance that matches or surpasses specialized methods with fewer evaluations and reduced token consumption.
The paper introduces OdoBot, a web‑agent architecture that reduces token usage by modeling application behavior from successful task demonstrations. Compared to two state‑of‑the‑art agents, OdoBot uses 44% and 80% fewer tokens on 45 tasks in the Canvas Learning Management System, and it also achieves a higher task success rate than WebVoyager.
The paper introduces PAI‑Bench, a benchmark designed to evaluate persistent AI agents on how faithfully they adhere to a versioned identity contract. It separates several dimensions—recall, composition, behavioral enactment, resistance, persistence, lineage, and role‑conditioned updates—while keeping scoring oracles independent of the target process. Experiments on synthetic profiles show that explicit cues can significantly alter the presence of identity identifiers, revealing prompt‑dependent component selection and sensitivity to startup cues.
ClinAgent is a conversational system that uses a ReAct-based LLM agent to retrieve and synthesize clinical trial information from multiple sources such as ClinicalTrials.gov, PubMed, and a local dataset. The agent iteratively reasons over user queries, selects appropriate tools, and refines its actions to provide grounded, up-to-date responses in natural language across multi-turn interactions. Evaluation across three phases shows that DeepSeek (thinking mode) excels in planning quality while Gemini 3.0 Flash delivers the highest overall performance and expert ratings, demonstrating the promise of agentic AI for improving clinical trial data access.