{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:IM4AWCYTOHPCCPII6P53FR5KIE","short_pith_number":"pith:IM4AWCYT","schema_version":"1.0","canonical_sha256":"43380b0b1371de213d08f3fbb2c7aa413ce62797134c751d7552cd868801e24d","source":{"kind":"arxiv","id":"2506.18096","version":2},"attestation_state":"computed","paper":{"title":"Deep Research Agents: A Systematic Examination And Roadmap","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Haozheng Zhang, Huichi Zhou, Jianye Hao, Jun Wang, Kang Li, Kun Shao, Lifeng Shang, Linyi Yang, Meng Fang, Songcen Xu, Xiaoguang Li, Yihang Chen, Yuxuan Huang","submitted_at":"2025-06-22T16:52:48Z","abstract_excerpt":"The rapid progress of Large Language Models (LLMs) has given rise to a new category of autonomous AI systems, referred to as Deep Research (DR) agents. These agents are designed to tackle complex, multi-turn informational research tasks by leveraging a combination of dynamic reasoning, adaptive long-horizon planning, multi-hop information retrieval, iterative tool use, and the generation of structured analytical reports. In this paper, we conduct a detailed analysis of the foundational technologies and architectural components that constitute Deep Research agents. We begin by reviewing informa"},"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":"2506.18096","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-06-22T16:52:48Z","cross_cats_sorted":[],"title_canon_sha256":"dda22068adfd2dc202c139cbc6348117d69f0df4412e300f8117eb8e7bc21745","abstract_canon_sha256":"27a4ad1795cf4f8f9888c5b8df39b1023fd1f5c3fa3f6e85cb1bd6e77dcc5efc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:04:16.059843Z","signature_b64":"8jHFDLEtp8dJbJ6yLUqX0IKn/nDJ6OWNK4pl0Z3f1t+OKqtIx7gxtIGa1t4iXB5t57Nfco72sDqeWO7jaAIGBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"43380b0b1371de213d08f3fbb2c7aa413ce62797134c751d7552cd868801e24d","last_reissued_at":"2026-07-05T12:04:16.059310Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:04:16.059310Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Deep Research Agents: A Systematic Examination And Roadmap","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Haozheng Zhang, Huichi Zhou, Jianye Hao, Jun Wang, Kang Li, Kun Shao, Lifeng Shang, Linyi Yang, Meng Fang, Songcen Xu, Xiaoguang Li, Yihang Chen, Yuxuan Huang","submitted_at":"2025-06-22T16:52:48Z","abstract_excerpt":"The rapid progress of Large Language Models (LLMs) has given rise to a new category of autonomous AI systems, referred to as Deep Research (DR) agents. These agents are designed to tackle complex, multi-turn informational research tasks by leveraging a combination of dynamic reasoning, adaptive long-horizon planning, multi-hop information retrieval, iterative tool use, and the generation of structured analytical reports. In this paper, we conduct a detailed analysis of the foundational technologies and architectural components that constitute Deep Research agents. We begin by reviewing informa"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.18096","kind":"arxiv","version":2},"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/2506.18096/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":"2506.18096","created_at":"2026-07-05T12:04:16.059383+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.18096v2","created_at":"2026-07-05T12:04:16.059383+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.18096","created_at":"2026-07-05T12:04:16.059383+00:00"},{"alias_kind":"pith_short_12","alias_value":"IM4AWCYTOHPC","created_at":"2026-07-05T12:04:16.059383+00:00"},{"alias_kind":"pith_short_16","alias_value":"IM4AWCYTOHPCCPII","created_at":"2026-07-05T12:04:16.059383+00:00"},{"alias_kind":"pith_short_8","alias_value":"IM4AWCYT","created_at":"2026-07-05T12:04:16.059383+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":29,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2607.06223","citing_title":"Information Gain-based Rollout Policy Optimization: An Adaptive Tree-Structured Rollout Approach for Multi-Turn LLM Agents","ref_index":7,"is_internal_anchor":true},{"citing_arxiv_id":"2607.06374","citing_title":"VaseMuseum: Digital Intelligent Museum for Ancient Greek Pottery","ref_index":21,"is_internal_anchor":true},{"citing_arxiv_id":"2606.24155","citing_title":"MedBench v5: A Dynamic, Process-Oriented, and Hallucination-Aware Benchmark for Clinical Multimodal Models","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2606.24155","citing_title":"MedBench v5: A Dynamic, Process-Oriented, and Hallucination-Aware Benchmark for Clinical Multimodal Models","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18508","citing_title":"MCompassRAG: Topic Metadata as a Semantic Compass for Paragraph-Level Retrieval","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2607.01523","citing_title":"Multi-Head Recurrent Memory Agents","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2606.12882","citing_title":"HarnessBridge: Learnable Bidirectional Controller for LLM Agent Harness","ref_index":1,"is_internal_anchor":false},{"citing_arxiv_id":"2606.06741","citing_title":"OpenSkill: Open-World Self-Evolution for LLM Agents","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2606.04308","citing_title":"Creative Reading: Scaffolding Reading for Transformation","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26200","citing_title":"Workflow Closure Is Not Scientific Closure in Auto-Research Systems","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2605.26835","citing_title":"Helicase: Uncertainty-Guided Supply Chain Knowledge Graph Construction with Autonomous Multi-Agent LLMs","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2601.11650","citing_title":"Large Language Model Agent for User-friendly Chemical Process Simulations","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07021","citing_title":"Behavior Cue Reasoning: Monitorable Reasoning Improves Efficiency and Safety through Oversight","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2605.00505","citing_title":"LLM-Oriented Information Retrieval: A Denoising-First Perspective","ref_index":71,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16346","citing_title":"PropGuard: Safeguarding LLM-MAS via Propagation-Aware Exploration and Remediation","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2510.01553","citing_title":"IoDResearch: Deep Research on Private Heterogeneous Data via the Internet of Data","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2510.00861","citing_title":"Erase to Improve: Erasable Reinforcement Learning for Search-Augmented LLMs","ref_index":22,"is_internal_anchor":false},{"citing_arxiv_id":"2603.27771","citing_title":"Emergent Social Intelligence Risks in Generative Multi-Agent Systems","ref_index":63,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13034","citing_title":"ViDR: Grounding Multimodal Deep Research Reports in Source Visual Evidence","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.13534","citing_title":"Scaling Retrieval-Augmented Reasoning with Parallel Search and Explicit Merging","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2604.04949","citing_title":"Learning to Retrieve from Agent Trajectories","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2604.02988","citing_title":"Self-Optimizing Multi-Agent Systems for Deep Research","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2604.27725","citing_title":"AgentEconomist: An End-to-end Agentic System Translating Economic Intuitions into Executable Computational Experiments","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2604.27221","citing_title":"Web2BigTable: A Bi-Level Multi-Agent LLM System for Internet-Scale Information Search and Extraction","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09915","citing_title":"Position: Academic Conferences are Potentially Facing Denominator Gaming Caused by Fully Automated Scientific Agents","ref_index":8,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IM4AWCYTOHPCCPII6P53FR5KIE","json":"https://pith.science/pith/IM4AWCYTOHPCCPII6P53FR5KIE.json","graph_json":"https://pith.science/api/pith-number/IM4AWCYTOHPCCPII6P53FR5KIE/graph.json","events_json":"https://pith.science/api/pith-number/IM4AWCYTOHPCCPII6P53FR5KIE/events.json","paper":"https://pith.science/paper/IM4AWCYT"},"agent_actions":{"view_html":"https://pith.science/pith/IM4AWCYTOHPCCPII6P53FR5KIE","download_json":"https://pith.science/pith/IM4AWCYTOHPCCPII6P53FR5KIE.json","view_paper":"https://pith.science/paper/IM4AWCYT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.18096&json=true","fetch_graph":"https://pith.science/api/pith-number/IM4AWCYTOHPCCPII6P53FR5KIE/graph.json","fetch_events":"https://pith.science/api/pith-number/IM4AWCYTOHPCCPII6P53FR5KIE/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IM4AWCYTOHPCCPII6P53FR5KIE/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IM4AWCYTOHPCCPII6P53FR5KIE/action/storage_attestation","attest_author":"https://pith.science/pith/IM4AWCYTOHPCCPII6P53FR5KIE/action/author_attestation","sign_citation":"https://pith.science/pith/IM4AWCYTOHPCCPII6P53FR5KIE/action/citation_signature","submit_replication":"https://pith.science/pith/IM4AWCYTOHPCCPII6P53FR5KIE/action/replication_record"}},"created_at":"2026-07-05T12:04:16.059383+00:00","updated_at":"2026-07-05T12:04:16.059383+00:00"}