{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:2IFRYFVV5LGFHEW7SOLZEBZ7IB","short_pith_number":"pith:2IFRYFVV","schema_version":"1.0","canonical_sha256":"d20b1c16b5eacc5392df939792073f406609c58caea4f90ec68384b95a2c913d","source":{"kind":"arxiv","id":"2605.14417","version":1},"attestation_state":"computed","paper":{"title":"Before the Body Moves: Learning Anticipatory Joint Intent for Language-Conditioned Humanoid Control","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"DAJI learns dynamics-aligned joint intents so language commands produce humanoid actions that are both immediate and anticipatory of future contacts and balance shifts.","cross_cats":["cs.CV"],"primary_cat":"cs.RO","authors_text":"Haozhe Jia, Honglei Jin, Jianfei Song, Kuimou Yu, Lei Wang, Shaofeng Liang, Shuxu Jin, Wenshuo Chen, Youcheng Fan, Yuan Zhang, Yutao Yue, Zinuo Zhang","submitted_at":"2026-05-14T06:05:24Z","abstract_excerpt":"Natural language is an intuitive interface for humanoid robots, yet streaming whole-body control requires control representations that are executable now and anticipatory of future physical transitions. Existing language-conditioned humanoid systems typically generate kinematic references that a low-level tracker must repair reactively, or use latent/action policies whose outputs do not explicitly encode upcoming contact changes, support transfers, and balance preparation. We propose \\textbf{DAJI} (\\emph{Dynamics-Aligned Joint Intent}), a hierarchical framework that learns an anticipatory join"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2605.14417","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-05-14T06:05:24Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"aef4c6b08e86424c0911a15e05a0879f6faefb8e7ad3ce9f1d6ba3400abf6c73","abstract_canon_sha256":"6288f1b1ae9c3eb8a08b65b33187f89b3afbae98dfac0c7893929907c4154a9e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:39:07.291487Z","signature_b64":"xi8EHI75llZZj7RZVn8cOr8s3KyX6j6xX7OhnCWYAfOw1E91AZaB9COlD+/ZDZn9C5LNuCC/gVaAS14r0kFoAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d20b1c16b5eacc5392df939792073f406609c58caea4f90ec68384b95a2c913d","last_reissued_at":"2026-05-17T23:39:07.290905Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:39:07.290905Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Before the Body Moves: Learning Anticipatory Joint Intent for Language-Conditioned Humanoid Control","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"DAJI learns dynamics-aligned joint intents so language commands produce humanoid actions that are both immediate and anticipatory of future contacts and balance shifts.","cross_cats":["cs.CV"],"primary_cat":"cs.RO","authors_text":"Haozhe Jia, Honglei Jin, Jianfei Song, Kuimou Yu, Lei Wang, Shaofeng Liang, Shuxu Jin, Wenshuo Chen, Youcheng Fan, Yuan Zhang, Yutao Yue, Zinuo Zhang","submitted_at":"2026-05-14T06:05:24Z","abstract_excerpt":"Natural language is an intuitive interface for humanoid robots, yet streaming whole-body control requires control representations that are executable now and anticipatory of future physical transitions. Existing language-conditioned humanoid systems typically generate kinematic references that a low-level tracker must repair reactively, or use latent/action policies whose outputs do not explicitly encode upcoming contact changes, support transfers, and balance preparation. We propose \\textbf{DAJI} (\\emph{Dynamics-Aligned Joint Intent}), a hierarchical framework that learns an anticipatory join"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"DAJI achieves strong results in anticipatory latent learning, single-instruction generation, and streaming instruction following, reaching 94.42% rollout success on HumanML3D-style generation and 0.152 subsequence FID on BABEL.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That distilling a future-aware teacher policy into a deployable diffusion action policy via student-driven rollouts preserves the anticipatory properties without substantial degradation in closed-loop performance.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"DAJI learns future-aware joint intents from language to enable proactive humanoid control, reporting 94.42% rollout success on HumanML3D-style tasks and 0.152 subsequence FID on BABEL.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"DAJI learns dynamics-aligned joint intents so language commands produce humanoid actions that are both immediate and anticipatory of future contacts and balance shifts.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"4f373da91245bcdc67226c56822d6ca3e94b2bdc8b8c1810cba8c7613af494bf"},"source":{"id":"2605.14417","kind":"arxiv","version":1},"verdict":{"id":"2179876d-1c20-40a3-a8c7-6f0105617676","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-15T02:11:06.582991Z","strongest_claim":"DAJI achieves strong results in anticipatory latent learning, single-instruction generation, and streaming instruction following, reaching 94.42% rollout success on HumanML3D-style generation and 0.152 subsequence FID on BABEL.","one_line_summary":"DAJI learns future-aware joint intents from language to enable proactive humanoid control, reporting 94.42% rollout success on HumanML3D-style tasks and 0.152 subsequence FID on BABEL.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That distilling a future-aware teacher policy into a deployable diffusion action policy via student-driven rollouts preserves the anticipatory properties without substantial degradation in closed-loop performance.","pith_extraction_headline":"DAJI learns dynamics-aligned joint intents so language commands produce humanoid actions that are both immediate and anticipatory of future contacts and balance shifts."},"references":{"count":48,"sample":[{"doi":"","year":1992,"title":"J.Massion,“Movement,postureandequilibrium:Interaction and coordination”,Progress in neurobiology, vol. 38, no. 1, pp.35–56,1992","work_id":"6cb63c6e-69c1-4aaf-b94b-a8f2b21703eb","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2008,"title":"S.BouissetandM.-C.Do,“Posture,dynamicstability,andvol- untarymovement”,NeurophysiologieClinique/ClinicalNeuro- physiology,vol.38,no.6,pp.345–362,2008","work_id":"0094cddd-6532-48b3-9f3a-5d337c17915e","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1109/iros.2012.6386109","year":2012,"title":"MuJoCo: A physics engine for model-based control","work_id":"5f5e6f59-3edd-4ba3-9cc2-7e7483d4d151","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2016,"title":"The kit motion- languagedataset","work_id":"8c1bfffa-65b6-4898-b933-cf4a05104c0b","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2017,"title":"Proximal Policy Optimization Algorithms","work_id":"240c67fe-d14d-4520-91c1-38a4e272ca19","ref_index":5,"cited_arxiv_id":"1707.06347","is_internal_anchor":true}],"resolved_work":48,"snapshot_sha256":"ad895c95f94df84021534a2baf35eed00455eba9ee26fcdbd0f0a8fc72c8d311","internal_anchors":8},"formal_canon":{"evidence_count":2,"snapshot_sha256":"37413b106950f3efe441ef4d8aa49686a1c6e3b8f70552feede16d596d2724b4"},"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":"2605.14417","created_at":"2026-05-17T23:39:07.291010+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.14417v1","created_at":"2026-05-17T23:39:07.291010+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.14417","created_at":"2026-05-17T23:39:07.291010+00:00"},{"alias_kind":"pith_short_12","alias_value":"2IFRYFVV5LGF","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"2IFRYFVV5LGFHEW7","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"2IFRYFVV","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2IFRYFVV5LGFHEW7SOLZEBZ7IB","json":"https://pith.science/pith/2IFRYFVV5LGFHEW7SOLZEBZ7IB.json","graph_json":"https://pith.science/api/pith-number/2IFRYFVV5LGFHEW7SOLZEBZ7IB/graph.json","events_json":"https://pith.science/api/pith-number/2IFRYFVV5LGFHEW7SOLZEBZ7IB/events.json","paper":"https://pith.science/paper/2IFRYFVV"},"agent_actions":{"view_html":"https://pith.science/pith/2IFRYFVV5LGFHEW7SOLZEBZ7IB","download_json":"https://pith.science/pith/2IFRYFVV5LGFHEW7SOLZEBZ7IB.json","view_paper":"https://pith.science/paper/2IFRYFVV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.14417&json=true","fetch_graph":"https://pith.science/api/pith-number/2IFRYFVV5LGFHEW7SOLZEBZ7IB/graph.json","fetch_events":"https://pith.science/api/pith-number/2IFRYFVV5LGFHEW7SOLZEBZ7IB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2IFRYFVV5LGFHEW7SOLZEBZ7IB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2IFRYFVV5LGFHEW7SOLZEBZ7IB/action/storage_attestation","attest_author":"https://pith.science/pith/2IFRYFVV5LGFHEW7SOLZEBZ7IB/action/author_attestation","sign_citation":"https://pith.science/pith/2IFRYFVV5LGFHEW7SOLZEBZ7IB/action/citation_signature","submit_replication":"https://pith.science/pith/2IFRYFVV5LGFHEW7SOLZEBZ7IB/action/replication_record"}},"created_at":"2026-05-17T23:39:07.291010+00:00","updated_at":"2026-05-17T23:39:07.291010+00:00"}