{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:PLD2WB2XOV3YJ7SOWGQEYS6OJS","short_pith_number":"pith:PLD2WB2X","schema_version":"1.0","canonical_sha256":"7ac7ab0757757784fe4eb1a04c4bce4c9feab39ea7ccb1d22245e9b2adf2a13c","source":{"kind":"arxiv","id":"2504.03601","version":4},"attestation_state":"computed","paper":{"title":"APIGen-MT: Agentic Pipeline for Multi-Turn Data Generation via Simulated Agent-Human Interplay","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Akshara Prabhakar, Caiming Xiong, Haolin Chen, Huan Wang, Jianguo Zhang, Juan Carlos Niebles, Ming Zhu, Shelby Heinecke, Shiyu Wang, Silvio Savarese, Thai Hoang, Tulika Awalgaonkar, Weiran Yao, Zhiwei Liu, Zuxin Liu","submitted_at":"2025-04-04T17:13:57Z","abstract_excerpt":"Training effective AI agents for multi-turn interactions requires high-quality data that captures realistic human-agent dynamics, yet such data is scarce and expensive to collect manually. We introduce APIGen-MT, a two-phase framework that generates verifiable and diverse multi-turn agent data. In the first phase, our agentic pipeline produces detailed task blueprints with ground-truth actions, leveraging a committee of LLM reviewers and iterative feedback loops. These blueprints are then transformed into complete interaction trajectories through simulated human-agent interplay. We train a fam"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2504.03601","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2025-04-04T17:13:57Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"bb2948dd846b51689eaf16d47c18bb755a7d1e8d4f7694a817ff49944630a218","abstract_canon_sha256":"8d2291f2196e497ad462d65c742e714d237d0f5b9d4a28c6ab68bd8ef9c1c5b7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:39:52.216051Z","signature_b64":"t95QR7JEq5MAB89CMkN6ZwEyF1AE0tibq2/oUNm2ne1uKJR0kpsLJgEfu01MxnEuyr3KhUSkhWkZGPy6as1sDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7ac7ab0757757784fe4eb1a04c4bce4c9feab39ea7ccb1d22245e9b2adf2a13c","last_reissued_at":"2026-07-05T11:39:52.215518Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:39:52.215518Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"APIGen-MT: Agentic Pipeline for Multi-Turn Data Generation via Simulated Agent-Human Interplay","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CL","authors_text":"Akshara Prabhakar, Caiming Xiong, Haolin Chen, Huan Wang, Jianguo Zhang, Juan Carlos Niebles, Ming Zhu, Shelby Heinecke, Shiyu Wang, Silvio Savarese, Thai Hoang, Tulika Awalgaonkar, Weiran Yao, Zhiwei Liu, Zuxin Liu","submitted_at":"2025-04-04T17:13:57Z","abstract_excerpt":"Training effective AI agents for multi-turn interactions requires high-quality data that captures realistic human-agent dynamics, yet such data is scarce and expensive to collect manually. We introduce APIGen-MT, a two-phase framework that generates verifiable and diverse multi-turn agent data. In the first phase, our agentic pipeline produces detailed task blueprints with ground-truth actions, leveraging a committee of LLM reviewers and iterative feedback loops. These blueprints are then transformed into complete interaction trajectories through simulated human-agent interplay. We train a fam"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.03601","kind":"arxiv","version":4},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2504.03601/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2504.03601","created_at":"2026-07-05T11:39:52.215595+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.03601v4","created_at":"2026-07-05T11:39:52.215595+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.03601","created_at":"2026-07-05T11:39:52.215595+00:00"},{"alias_kind":"pith_short_12","alias_value":"PLD2WB2XOV3Y","created_at":"2026-07-05T11:39:52.215595+00:00"},{"alias_kind":"pith_short_16","alias_value":"PLD2WB2XOV3YJ7SO","created_at":"2026-07-05T11:39:52.215595+00:00"},{"alias_kind":"pith_short_8","alias_value":"PLD2WB2X","created_at":"2026-07-05T11:39:52.215595+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":21,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.12908","citing_title":"SENTINEL: Failure-Driven Reinforcement Learning for Training Tool-Using Language Model Agents","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2606.11520","citing_title":"ISE: An Execution-Grounded Recipe for Multi-Turn OS-Agent Trajectories","ref_index":54,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03892","citing_title":"Synthesize and Reward -- Reinforcement Learning for Multi-Step Tool Use in Live Environments","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2606.02001","citing_title":"Scaling Agentic Capabilities via Grounded Interaction Synthesis","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10832","citing_title":"Towards On-Policy Data Evolution for Visual-Native Multimodal Deep Search Agents","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2606.16364","citing_title":"Looking Is Not Picking: An Attention-Segment Account of Tool-Selection Failures in LLM Agents","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00135","citing_title":"On Effectiveness and Efficiency of Agentic Tool-calling and RL Training","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12191","citing_title":"Agentic Environment Engineering for Large Language Models: A Survey of Environment Modeling, Synthesis, Evaluation, and Application","ref_index":268,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20876","citing_title":"Terminal-World: Scaling Terminal-Agent Environments via Agent Skills","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2605.18805","citing_title":"RecoAtlas: From Semantic Plausibility to Set-Level Utility in LLM Recommendation Agents","ref_index":44,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17558","citing_title":"Firefly: Illuminating Large-Scale Verified Tool-Call Data Generation from Real APIs","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2511.21686","citing_title":"Matrix: Peer-to-Peer Multi-Agent Synthetic Data Generation Framework","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2508.07407","citing_title":"A Comprehensive Survey of Self-Evolving AI Agents: A New Paradigm Bridging Foundation Models and Lifelong Agentic Systems","ref_index":69,"is_internal_anchor":false},{"citing_arxiv_id":"2603.24709","citing_title":"Training LLMs for Multi-Step Tool Orchestration with Constrained Data Synthesis and Graduated Rewards","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12521","citing_title":"ToolWeave: Structured Synthesis of Complex Multi-Turn Tool-Calling Dialogues","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2605.03476","citing_title":"CuraView: A Multi-Agent Framework for Medical Hallucination Detection with GraphRAG-Enhanced Knowledge Verification","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2506.07982","citing_title":"$\\tau^2$-Bench: Evaluating Conversational Agents in a Dual-Control Environment","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10832","citing_title":"Towards On-Policy Data Evolution for Visual-Native Multimodal Deep Search Agents","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2604.09813","citing_title":"Controllable and Verifiable Tool-Use Data Synthesis for Agentic Reinforcement Learning","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18292","citing_title":"Agent-World: Scaling Real-World Environment Synthesis for Evolving General Agent Intelligence","ref_index":75,"is_internal_anchor":false},{"citing_arxiv_id":"2605.02572","citing_title":"On Training Large Language Models for Long-Horizon Tasks: An Empirical Study of Horizon Length","ref_index":48,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PLD2WB2XOV3YJ7SOWGQEYS6OJS","json":"https://pith.science/pith/PLD2WB2XOV3YJ7SOWGQEYS6OJS.json","graph_json":"https://pith.science/api/pith-number/PLD2WB2XOV3YJ7SOWGQEYS6OJS/graph.json","events_json":"https://pith.science/api/pith-number/PLD2WB2XOV3YJ7SOWGQEYS6OJS/events.json","paper":"https://pith.science/paper/PLD2WB2X"},"agent_actions":{"view_html":"https://pith.science/pith/PLD2WB2XOV3YJ7SOWGQEYS6OJS","download_json":"https://pith.science/pith/PLD2WB2XOV3YJ7SOWGQEYS6OJS.json","view_paper":"https://pith.science/paper/PLD2WB2X","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.03601&json=true","fetch_graph":"https://pith.science/api/pith-number/PLD2WB2XOV3YJ7SOWGQEYS6OJS/graph.json","fetch_events":"https://pith.science/api/pith-number/PLD2WB2XOV3YJ7SOWGQEYS6OJS/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PLD2WB2XOV3YJ7SOWGQEYS6OJS/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PLD2WB2XOV3YJ7SOWGQEYS6OJS/action/storage_attestation","attest_author":"https://pith.science/pith/PLD2WB2XOV3YJ7SOWGQEYS6OJS/action/author_attestation","sign_citation":"https://pith.science/pith/PLD2WB2XOV3YJ7SOWGQEYS6OJS/action/citation_signature","submit_replication":"https://pith.science/pith/PLD2WB2XOV3YJ7SOWGQEYS6OJS/action/replication_record"}},"created_at":"2026-07-05T11:39:52.215595+00:00","updated_at":"2026-07-05T11:39:52.215595+00:00"}