Tag
TeachMateGPT presents a multi-agent framework that improves retrieval-augmented generation for creating pedagogical assessments from science textbooks, achieving higher faithfulness and answer relevancy compared to baseline systems.
This paper introduces AEROBAT, the first multi-agent system to automate behavioral scientific research on AI agents, generating hypotheses, designing and executing controlled experiments, and writing reports. The authors demonstrate its efficacy across 12 target behaviors, finding statistical evidence for 26 hypotheses.
This paper presents IntelliAudit, a retrieval-grounded multi-agent system that uses large language models to evaluate IT audit controls against evidence corpora, generating cited recommendations and remediation guidance. The authors instantiate it on ISO/IEC 27001 and find it useful for audit preparation while emphasizing the need for human oversight.
QFoldAgent is a closed-loop multi-agent framework for quantum-classical protein structure prediction that iteratively optimizes Hamiltonian penalties using VQE and feedback, achieving improved RMSD and structural validity on 5-residue fragments.
This paper proposes a multi-agent embodied conversational system that generates level-appropriate dialogue for English learners using a generate-evaluate-regenerate loop with LLMs and a CEFR classifier. A pilot study with Japanese university students showed improved level appropriateness but no statistically significant reduction in foreign language anxiety.
This paper describes a hybrid multi-agent LLM system for conversational depression screening submitted to the eRisk 2026 challenge, using either a paid GPT-5-nano or open-source Gemma 27B model with algorithmic guidance (dialogue tree, reliability-weighted aggregation, cluster-based imputation) to achieve competitive BDI-II assessment at lower cost.
Pythia is a multi-agent system that autonomously writes and optimizes extraction prompts for clinical concepts without manual prompt engineering or fine-tuning, using a locally hosted open-weights model. It achieves mean sensitivity of 0.76 and specificity of 0.95 on clinical symptom detection, outperforming lexicon-based methods on specificity.
This paper presents a question-type-specific LLM framework for the BioASQ 14b Task B biomedical QA challenge, using different inference strategies (snippet shuffling, chain-of-thought, multi-agent collaboration) for yes/no, factoid, and list questions, achieving competitive results including first place in the factoid subtask of Batch 4.
Prompt-to-Paper is a multi-stage multi-agent AI framework for automated bioinformatics manuscript generation that uses deterministic retrieval-augmented generation, an autonomous coding agent for real experiments, and an eight-dimensional quality scorer to produce submission-formatted PDFs at low cost, with validated quality improvements.
Agent4cs is a multi-agent framework that summarizes large, hierarchical codebases in a bottom-up fashion using specialized agents for summarization, keyword extraction, and quality assurance, achieving up to 8% improvement in semantic consistency and 38% gain in keyword coverage over baselines.
Proposes DDIAgents, a mechanism-conditioned multi-agent framework for drug-drug interaction prediction that dynamically routes relevant biomedical knowledge to specialized expert agents and aggregates their analyses, outperforming existing feature-based, graph-based, and LLM-based methods.
MetaResearcher proposes a framework for training deep research agents using self-reflective reinforcement learning in adversarial virtual environments, addressing limitations of static environments and fact-retrieval-only tasks.
Introduces MODE-RAG, a multi-agent system using Variational Free Energy and Monte Carlo Tree Search to dynamically gate interventions for mitigating hallucinations in Multimodal Retrieval-Augmented Generation systems, along with the ModeVent evaluation dataset.
The author describes building a Minsky brain: a runtime of 40+ LLM agents wired in a connectome and staged phylogenetically to simulate a brain. They ask for advice on where to post this project on Reddit.
This paper proposes Embodied-BenchClaw, an autonomous multi-agent system that automatically constructs embodied spatial intelligence benchmarks from user intent through a five-stage pipeline with process quality control and an extensible Skill Library.
CAF-Gen is a multi-agent LLM-driven framework that enriches shallow argument structures into formal Carneades Argumentation Framework models using an iterative Creator-Reviewer pipeline, achieving improved structural alignment and quality.
This paper presents CatDT, a self-evolving multi-agent digital twin that autonomously predicts heterogeneous catalyst properties from bulk crystal and reaction description, achieving experimental accuracy across seven benchmarks and discovering non-precious catalyst candidates for propane dehydrogenation.
A developer used a multi-agent AI system (Claude, Gemini, Quen) to build and deploy a full cybersecurity product, ScannerSend, including miner detection, threat intel API, and payment processing, with agents managing production infrastructure.
This paper presents Traj-Evolve, a self-evolving multi-agent system that uses an experience pool and multi-agent reinforcement learning to model patient trajectories from longitudinal EHRs for lung cancer early detection, outperforming strong baselines.
AutoSci is a memory-centric agentic system designed to automate the full scientific research lifecycle, from literature understanding to rebuttal, using LLM-based agents with persistent memory and self-evolution capabilities.