{"paper":{"title":"TeachAnything: A Multimodal Crowdsourcing Platform for Training Embodied AI Agents in Symmetrical Reality","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"TeachAnything platform collects multimodal demonstrations via crowdsourcing and physics simulation to train embodied agents for Symmetrical Reality.","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Rongkai Liu, Yue Li, Zhenliang Zhang, Zidong Liu","submitted_at":"2026-05-14T08:32:46Z","abstract_excerpt":"Symmetrical Reality (SR) is emerging as a future trend for human-agent coexistence, placing higher demands on agents to acquire human-like intelligence. It calls for richer and more diverse human guidance. We introduce a three-stage demonstration paradigm integrating multimodal demonstration signals. Building on this paradigm, we developed TeachAnything, a cloud-based, crowdsourcing-oriented demonstration platform with physics simulation capable of collecting diverse demonstration data across varied scenes, tasks, and embodiments. By unifying virtual and physical interactions through both meth"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Building on this paradigm, we developed TeachAnything, a cloud-based, crowdsourcing-oriented demonstration platform with physics simulation capable of collecting diverse demonstration data across varied scenes, tasks, and embodiments.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That integrating multimodal demonstration signals via a three-stage paradigm and crowdsourcing platform will provide the richer human guidance required for embodied agents to acquire human-like intelligence in Symmetrical Reality.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"TeachAnything is a new crowdsourcing platform with physics simulation for collecting multimodal demonstration data to train embodied AI agents in Symmetrical Reality.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"TeachAnything platform collects multimodal demonstrations via crowdsourcing and physics simulation to train embodied agents for Symmetrical Reality.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"d8d9e8220d06ef9a9d902ac53ec909de4d477da5724be8be0d451780d8dad8af"},"source":{"id":"2605.14556","kind":"arxiv","version":1},"verdict":{"id":"7a45ebf4-af3f-4f4c-ae7b-7d522318e0ac","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T01:36:36.839678Z","strongest_claim":"Building on this paradigm, we developed TeachAnything, a cloud-based, crowdsourcing-oriented demonstration platform with physics simulation capable of collecting diverse demonstration data across varied scenes, tasks, and embodiments.","one_line_summary":"TeachAnything is a new crowdsourcing platform with physics simulation for collecting multimodal demonstration data to train embodied AI agents in Symmetrical Reality.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That integrating multimodal demonstration signals via a three-stage paradigm and crowdsourcing platform will provide the richer human guidance required for embodied agents to acquire human-like intelligence in Symmetrical Reality.","pith_extraction_headline":"TeachAnything platform collects multimodal demonstrations via crowdsourcing and physics simulation to train embodied agents for Symmetrical Reality."},"references":{"count":5,"sample":[{"doi":"","year":2024,"title":"On the emergence of symmetrical reality,","work_id":"a8d505d8-659c-4924-9be5-b1746437401a","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Embodied AI: A Survey on the Evolution from Perceptive to Behavioral Intelligence,","work_id":"525ef5ae-6cde-4895-aa31-a0168c4504e9","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"Roboturk: A crowdsourcing platform for robotic skill learning through imitation,","work_id":"cd48c7c4-90f5-48d5-b7e9-04d6e034f87c","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Rt-2: Vision-language-action models transfer web knowledge to robotic control,","work_id":"5c440657-aea5-48e9-8bdf-97d7c7c0c267","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Reconstructing hands in 3d with transformers,","work_id":"74903762-8251-41b2-83a0-d6ee264a6833","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":5,"snapshot_sha256":"1ea43136dd6a1b9903b3e092187baa0f3c37b46bfe91dc058dff136dde39d950","internal_anchors":0},"formal_canon":{"evidence_count":1,"snapshot_sha256":"30025efc9e1cc0a222c9d7177d4b4784bd3d461863a42d30e5b09079deac631b"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}