{"paper":{"title":"Agent4POI: Agentic Context-Conditioned Affordance Reasoning for Multimodal Point-of-Interest Recommendation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"No pre-computed encoder can satisfy context-sensitive POI ranking under bilinear scoring, so Agent4POI generates dynamic affordance representations at recommendation time instead.","cross_cats":["cs.AI","cs.MA"],"primary_cat":"cs.IR","authors_text":"Jinze Wang, Lu Zhang, Tiehua Zhang, Xingjun Ma, Yangchen Zeng, Yongchao Liu, Yuze Liu, Zhu Sun","submitted_at":"2026-04-03T01:53:05Z","abstract_excerpt":"We introduce Agent4POI, the first POI recommendation framework that generates context-conditioned multimodal representations at recommendation time, rather than relying on static POI embeddings pre-computed independently of context. Existing multimodal systems encode each POI once as a static embedding, a design that precludes reasoning about why the same cafe affords solo work on Monday but group celebration on Friday evening. We formally prove that no pre-computed encoder can satisfy context-sensitive ranking under standard bilinear scoring, motivating inference-time item-side representation"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"No pre-computed encoder can satisfy context-sensitive ranking under standard bilinear scoring; Agent4POI achieves a 23.2% relative gain over the strongest baseline and degrades by only 7.5% under context-shift versus 16-17% for baselines.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The four-phase LLM agent can reliably generate accurate uncertainty-aware affordance representations through five-step cross-modal chain-of-thought reasoning over image, review, and metadata evidence.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"Agent4POI generates context-conditioned multimodal affordance representations via a four-phase LLM agent, achieving 23.2% relative gains over baselines on POI benchmarks with reduced degradation under context shifts.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"No pre-computed encoder can satisfy context-sensitive POI ranking under bilinear scoring, so Agent4POI generates dynamic affordance representations at recommendation time instead.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"1351f276266967069e44ba4c5ed6de76e0c7ee0e80803ff0330f6ece0bcf2f26"},"source":{"id":"2605.15203","kind":"arxiv","version":1},"verdict":{"id":"72ae2663-3c91-45aa-ac8d-b4c8ec91df70","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T17:31:31.145060Z","strongest_claim":"No pre-computed encoder can satisfy context-sensitive ranking under standard bilinear scoring; Agent4POI achieves a 23.2% relative gain over the strongest baseline and degrades by only 7.5% under context-shift versus 16-17% for baselines.","one_line_summary":"Agent4POI generates context-conditioned multimodal affordance representations via a four-phase LLM agent, achieving 23.2% relative gains over baselines on POI benchmarks with reduced degradation under context shifts.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The four-phase LLM agent can reliably generate accurate uncertainty-aware affordance representations through five-step cross-modal chain-of-thought reasoning over image, review, and metadata evidence.","pith_extraction_headline":"No pre-computed encoder can satisfy context-sensitive POI ranking under bilinear scoring, so Agent4POI generates dynamic affordance representations at recommendation time instead."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.15203/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":40,"sample":[{"doi":"","year":2020,"title":"Paola Ardón, Èric Pairet, Katrin S Lohan, Subramanian Ramamoorthy, and Ronald Petrick. 2020. Affordances in robotic tasks–a survey.arXiv preprint arXiv:2004.07400(2020)","work_id":"a4734864-eb5d-40df-b0c6-aca2dd7fcc61","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Keqin Bao, Jizhi Zhang, Wenjie Wang, Yang Zhang, Zhengyi Yang, Yanchen Luo, Chong Chen, Fuli Feng, and Qi Tian. 2025. A bi-step grounding paradigm for large language models in recommendation systems.A","work_id":"a1706006-4d47-493d-afb2-519efccba14c","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Keqin Bao, Jizhi Zhang, Yang Zhang, Wenjie Wang, Fuli Feng, and Xiangnan He. 2023. TALLRec: An effective and efficient tuning framework to align large language model with recommendation. InProceedings","work_id":"ad4097fc-3910-4514-88a9-97dd9234e633","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"Ramesh Baral, XiaoLong Zhu, SS Iyengar, and Tao Li. 2018. REEL: Review aware explanation of location recommendation. InProceedings of the 26th Conference on User Modeling, Adaptation and Personalizati","work_id":"64676e6b-f7c9-407a-a8fc-90cc5765ca33","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":1991,"title":"William W Gaver. 1991. Technology affordances. InProceedings of the SIGCHI Conference on Human Factors in Computing Systems. 79–84","work_id":"5ec16b1c-d99c-4a07-b374-5200ad5b4030","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":40,"snapshot_sha256":"9b24f6d28d5c5c6d8194e9bb493c766e9fac4922ae3450465e692bc89da86c0d","internal_anchors":3},"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"}