{"id":"449019c6-e4f8-43f0-996d-ec4d1c077b08","slug":"dextacwam","name":"DexTacWAM","description":"DexTacWAM (Dexterous Tactile World-Action Model) is a Sep 2026 academic visuo-tactile WAM (Yuan, Wang, Shao, Lu, Darrell, Lourentzou, Zhan; UIUC + UC Berkeley + Northwestern; arXiv 2609.24976) that jointly models visual scene dynamics and distributed multi-finger contact dynamics. A finger- and pose-aware tactile compressor maps five fingertip streams into per-hand latents injected into a video diffusion world model; an action expert conditions on the joint visuo-tactile state. Author-reported mean 70.6 across six real-robot tasks versus 38.0 for the strongest baseline; compressor retains 89.4% pre-fusion contact recall with 2.26x train / 1.29x infer speedups. Research demo on a 22-DoF bimanual Sharpa tactile platform, not a commercial product.","brainType":"research-model","isOpen":false,"maturityStage":"research","architecture":"Per-finger tactile maps via frozen pretrained vision VAE + grayscale-to-RGB adapter; finger- and pose-aware 5:1 tactile compressor into per-hand latents; video diffusion DiT world model jointly denoises visual + tactile views; per-modality K/V RMS-normalized action expert (flow matching) for dexterous actions","reviewStatus":"reviewed","sources":[{"url":"https://arxiv.org/abs/2609.24976","title":"DexTacWAM: A Visuo-Tactile World-Action Model for Dexterous Manipulation (arXiv:2609.24976)","sourceName":"arXiv","publishedAt":"2026-09-21"},{"url":"https://arxiv.org/pdf/2609.24976","title":"DexTacWAM PDF (arXiv:2609.24976)","sourceName":"arXiv","publishedAt":"2026-09-21"},{"url":"https://dextacwam.github.io/","title":"DexTacWAM project site","sourceName":"DexTacWAM project","publishedAt":"2026-09-21"},{"url":"https://x.com/robotsdigest/status/2105059594288681337","title":"Secondary summary: DexTacWAM 70.6% avg across six real-robot tasks (robotsdigest)","sourceName":"X (@robotsdigest)","publishedAt":"2026-09-29"}],"keyFacts":[{"label":"System (project-stated)","value":"Project-stated (project-stated; also project site): DexTacWAM is a visuo-tactile world-action model (Yuan / Wang / Shao / Lu / Darrell / Lourentzou / Zhan; UIUC + UC Berkeley + Northwestern; arXiv 2609.24976, Sep 21 2026) that jointly predicts visual scenes and multi-finger contact dynamics. Encodes each fingertip independently, aggregates via a finger- and pose-aware tactile compressor, and injects tactile latents into a video diffusion world model for joint visuo-tactile world modeling. Academic research system, not a commercial product"},{"label":"Six-task real-robot results (project-stated)","value":"Project-stated (project-stated): on six contact-rich dexterous manipulation tasks on a 22-DoF bimanual platform (20 real-robot trials per method per task), DexTacWAM averages 70.6 versus 38.0 for the strongest baseline (RDP). Cap: author-reported lab results; not independent third-party verification"},{"label":"Tactile compressor retention / speedups (project-stated)","value":"Project-stated (project-stated): finger- and pose-aware tactile compressor retains 89.4% of pre-fusion contact recall while enabling 2.26x faster training and 1.29x faster inference. Ablation: removing tactile world modeling drops four-task mean from 74.7 to 26.6 with the same tactile features and action expert. Cap: paper-reported metrics"},{"label":"Tactile world-model ablation","value":"The project page says removing tactile world modeling drops the four-task mean from 74.7 to 26.6 while keeping the same tactile features and action expert (DexTacWAM)."},{"label":"Sharpa 22-DoF platform","value":"The paper says the six-task evaluation used a bimanual platform with multi-finger 22-DoF Sharpa tactile sensors (arXiv 2609.24976)."},{"label":"UIUC, Berkeley, and Northwestern","value":"The project page lists authors at the University of Illinois Urbana-Champaign, the University of California, Berkeley, and Northwestern University (DexTacWAM)."}],"aliases":["DexTacWAM","Dex Tac WAM","Dexterous Tactile World-Action Model","Visuo-Tactile World-Action Model"],"collisionRisk":"low","reviewNote":null,"builtOnBrainId":null,"createdAt":"2026-09-30T01:59:58.484Z","updatedAt":"2026-10-09T22:00:42.086Z","jsonLd":{"@context":"https://schema.org","@type":"SoftwareApplication","@id":"https://registry.deploy.report/brains/dextacwam","url":"https://registry.deploy.report/brains/dextacwam","name":"DexTacWAM","alternateName":["DexTacWAM","Dex Tac WAM","Dexterous Tactile World-Action Model","Visuo-Tactile World-Action Model"],"description":"DexTacWAM (Dexterous Tactile World-Action Model) is a Sep 2026 academic visuo-tactile WAM (Yuan, Wang, Shao, Lu, Darrell, Lourentzou, Zhan; UIUC + UC Berkeley + Northwestern; arXiv 2609.24976) that jointly models visual scene dynamics and distributed multi-finger contact dynamics. A finger- and pose-aware tactile compressor maps five fingertip streams into per-hand latents injected into a video diffusion world model; an action expert conditions on the joint visuo-tactile state. Author-reported mean 70.6 across six real-robot tasks versus 38.0 for the strongest baseline; compressor retains 89.4% pre-fusion contact recall with 2.26x train / 1.29x infer speedups. Research demo on a 22-DoF bimanual Sharpa tactile platform, not a commercial product.","identifier":"449019c6-e4f8-43f0-996d-ec4d1c077b08","applicationCategory":"research-model","publisher":{"@id":"https://deploy.report/#organization"}},"framework_metadata":{"framework_schema_version":"0.1.0","verification_status":"verified","maturity_stage":"research","lifecycle_state":null,"architectural_position":{"cohort":null,"sub_cohorts":[]},"within_cohort_verified_vs_claimed_pair":null,"cap_flags":[],"verification_depth":{"sources_count":4,"primary_source_types":["preprint"]}}}