{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:2RAJHYNJGMJLLWDLPYHWXXYTQM","short_pith_number":"pith:2RAJHYNJ","schema_version":"1.0","canonical_sha256":"d44093e1a93312b5d86b7e0f6bdf13833971165d726c28c3f9038defdce06da8","source":{"kind":"arxiv","id":"2407.05996","version":1},"attestation_state":"computed","paper":{"title":"Multimodal Diffusion Transformer: Learning Versatile Behavior from Multimodal Goals","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.RO","authors_text":"Fabian Wenzel, Moritz Reuss, \\\"Omer Erdin\\c{c} Ya\\u{g}murlu, Rudolf Lioutikov","submitted_at":"2024-07-08T14:46:44Z","abstract_excerpt":"This work introduces the Multimodal Diffusion Transformer (MDT), a novel diffusion policy framework, that excels at learning versatile behavior from multimodal goal specifications with few language annotations. MDT leverages a diffusion-based multimodal transformer backbone and two self-supervised auxiliary objectives to master long-horizon manipulation tasks based on multimodal goals. The vast majority of imitation learning methods only learn from individual goal modalities, e.g. either language or goal images. However, existing large-scale imitation learning datasets are only partially label"},"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":"2407.05996","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.RO","submitted_at":"2024-07-08T14:46:44Z","cross_cats_sorted":[],"title_canon_sha256":"d2908a0668cec9fedf70e55a262f2f9018a67ab7f7967186817189da450c77de","abstract_canon_sha256":"faf9d766a4eb8cd2f8616f540bc2611f3997da82940c9eeb521854a7d9376112"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:41:22.115869Z","signature_b64":"5yI6CPS8LvtJoQXfdkDFk0c/TmuhNK4KkmWGTe3vOxqZ54BqJKICMUM99xm6s2E4wPiEKjD/BAYzPiBc9DavDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d44093e1a93312b5d86b7e0f6bdf13833971165d726c28c3f9038defdce06da8","last_reissued_at":"2026-07-05T08:41:22.115324Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:41:22.115324Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Multimodal Diffusion Transformer: Learning Versatile Behavior from Multimodal Goals","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.RO","authors_text":"Fabian Wenzel, Moritz Reuss, \\\"Omer Erdin\\c{c} Ya\\u{g}murlu, Rudolf Lioutikov","submitted_at":"2024-07-08T14:46:44Z","abstract_excerpt":"This work introduces the Multimodal Diffusion Transformer (MDT), a novel diffusion policy framework, that excels at learning versatile behavior from multimodal goal specifications with few language annotations. MDT leverages a diffusion-based multimodal transformer backbone and two self-supervised auxiliary objectives to master long-horizon manipulation tasks based on multimodal goals. The vast majority of imitation learning methods only learn from individual goal modalities, e.g. either language or goal images. However, existing large-scale imitation learning datasets are only partially label"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.05996","kind":"arxiv","version":1},"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/2407.05996/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":"2407.05996","created_at":"2026-07-05T08:41:22.115378+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.05996v1","created_at":"2026-07-05T08:41:22.115378+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.05996","created_at":"2026-07-05T08:41:22.115378+00:00"},{"alias_kind":"pith_short_12","alias_value":"2RAJHYNJGMJL","created_at":"2026-07-05T08:41:22.115378+00:00"},{"alias_kind":"pith_short_16","alias_value":"2RAJHYNJGMJLLWDL","created_at":"2026-07-05T08:41:22.115378+00:00"},{"alias_kind":"pith_short_8","alias_value":"2RAJHYNJ","created_at":"2026-07-05T08:41:22.115378+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":16,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26588","citing_title":"Inference-Time Robot Behavior Steering through Physically-Aware Reconfiguration of Task-Structure","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2606.21501","citing_title":"UniviewVLA: A Unified Multiview Vision-Language-Action Model with World Modeling","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2606.04233","citing_title":"What Are We Actually Benchmarking in Robot Manipulation?","ref_index":59,"is_internal_anchor":false},{"citing_arxiv_id":"2405.14093","citing_title":"A Survey on Vision-Language-Action Models for Embodied AI","ref_index":114,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22816","citing_title":"AwareVLN: Reasoning with Self-awareness for Vision-Language Navigation","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2505.17016","citing_title":"Interactive Post-Training for Vision-Language-Action Models","ref_index":27,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17923","citing_title":"AdaptiveLoad: Towards Efficient Video Diffusion Transformer Training","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2510.12710","citing_title":"Reflection-Based Task Adaptation for Self-Improving VLA","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2507.15493","citing_title":"GR-3 Technical Report","ref_index":64,"is_internal_anchor":false},{"citing_arxiv_id":"2601.07060","citing_title":"PALM: Progress-Aware Policy Learning via Affordance Reasoning for Long-Horizon Robotic Manipulation","ref_index":103,"is_internal_anchor":false},{"citing_arxiv_id":"2508.19236","citing_title":"MemoryVLA: Perceptual-Cognitive Memory in Vision-Language-Action Models for Robotic Manipulation","ref_index":20,"is_internal_anchor":false},{"citing_arxiv_id":"2603.25903","citing_title":"Emergent Neural Automaton Policies: Learning Symbolic Structure from Visuomotor Trajectories","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2412.14803","citing_title":"Video Prediction Policy: A Generalist Robot Policy with Predictive Visual Representations","ref_index":121,"is_internal_anchor":false},{"citing_arxiv_id":"2604.24622","citing_title":"CF-VLA: Efficient Coarse-to-Fine Action Generation for Vision-Language-Action Policies","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2502.19645","citing_title":"Fine-Tuning Vision-Language-Action Models: Optimizing Speed and Success","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05672","citing_title":"A1: A Fully Transparent Open-Source, Adaptive and Efficient Truncated Vision-Language-Action Model","ref_index":31,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2RAJHYNJGMJLLWDLPYHWXXYTQM","json":"https://pith.science/pith/2RAJHYNJGMJLLWDLPYHWXXYTQM.json","graph_json":"https://pith.science/api/pith-number/2RAJHYNJGMJLLWDLPYHWXXYTQM/graph.json","events_json":"https://pith.science/api/pith-number/2RAJHYNJGMJLLWDLPYHWXXYTQM/events.json","paper":"https://pith.science/paper/2RAJHYNJ"},"agent_actions":{"view_html":"https://pith.science/pith/2RAJHYNJGMJLLWDLPYHWXXYTQM","download_json":"https://pith.science/pith/2RAJHYNJGMJLLWDLPYHWXXYTQM.json","view_paper":"https://pith.science/paper/2RAJHYNJ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.05996&json=true","fetch_graph":"https://pith.science/api/pith-number/2RAJHYNJGMJLLWDLPYHWXXYTQM/graph.json","fetch_events":"https://pith.science/api/pith-number/2RAJHYNJGMJLLWDLPYHWXXYTQM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2RAJHYNJGMJLLWDLPYHWXXYTQM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2RAJHYNJGMJLLWDLPYHWXXYTQM/action/storage_attestation","attest_author":"https://pith.science/pith/2RAJHYNJGMJLLWDLPYHWXXYTQM/action/author_attestation","sign_citation":"https://pith.science/pith/2RAJHYNJGMJLLWDLPYHWXXYTQM/action/citation_signature","submit_replication":"https://pith.science/pith/2RAJHYNJGMJLLWDLPYHWXXYTQM/action/replication_record"}},"created_at":"2026-07-05T08:41:22.115378+00:00","updated_at":"2026-07-05T08:41:22.115378+00:00"}