{"record": {"aliases": ["Tabero: Learning Gentle Manipulation with Closed-Loop Force Feedback from Vision, Touch, and Language"], "artifacts": [{"id": "arxiv:2605.27886", "kind": "paper", "url": "https://arxiv.org/abs/2605.27886"}, {"id": "gh:nathanwu7/tabero", "kind": "repo", "url": "https://github.com/NathanWu7/Tabero"}], "categories": ["Robot manipulation", "Robotics", "Robotics & Embodied AI"], "description": {"en": "Tabero evaluates language-conditioned gentle manipulation using vision, touch, task success, and physical-interaction quality."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2605.27886", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2605.27886", "title": "Tabero"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2605.27886", "modality": null, "name": "Tabero", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"admission_policy_version": "2026-09-09.1", "confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-09-25T04:40:22Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_tabero-learning-gentle-manipulation-with-c_55e0b024", "origin_source": "arxiv", "recognition_confidence": "0.55", "record_sha256": "317fc65f2659bb115ec53cfa91b1920aca75940f86ca7dda1224905835d0635a", "relation": "evaluates_only", "review_model": "codex-cli-default", "review_state": "ai-reviewed", "reviewed_at": "2026-09-24T19:39:58.745325Z", "source_url": "https://arxiv.org/abs/2605.27886"}, "publisher": null, "released": "2026-05-27", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2605.27886", "source_url": "https://arxiv.org/abs/2605.27886"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2605-27886", "source": "claire_radar", "source_benchmark_id": "2605.27886", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": ["Tabero: Learning Gentle Manipulation with Closed-Loop Force Feedback from Vision, Touch, and Language"], "annotation": "Unknown", "applicationDomains": ["Robotics & Embodied AI"], "area": "Robotics & Embodied AI", "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "Multimodal robot learning and tactile feedback are likely to draw near-term research interest.", "score": 62}, "capabilities": ["Robot manipulation"], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-24T19:39:58.745325Z", "inputHash": "7fe7d6b4563ba4b75f2684565026830aafbb03f0229656a3a879068e130c743b", "model": "codex-cli-default"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "Tabero: Learning Gentle Manipulation with Closed-Loop Force Feedback from Vision, Touch, and Language", "canonicalNameSource": "paper_title", "decisionReason": "The source formally names the benchmark, states its task and multidimensional evaluation protocol, and supplies a currently public code repository.", "evaluationEvidence": {"reuse": "Our code is publicly available at https://github.com/NathanWu7/Tabero.", "scoring": "measures task success alongside physical interaction quality", "task": "gentle, language-conditioned robotic manipulation that demands fine-grained contact force perception"}, "model": "codex-cli-default", "reviewedAt": "2026-09-24T19:39:58.745325Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Tabero evaluates language-conditioned gentle manipulation using vision, touch, task success, and physical-interaction quality.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["benchmark term in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL", "named entity is a model, method, framework, or dataset used in evaluation"], "snippet": "To address these challenges, we introduce Tabero, a benchmark and model suite for gentle, language-conditioned robotic manipulation that demands fine-grained contact force perception."}, "familyId": "bmf_85665ae7bf0a", "firstSeenAt": "2026-08-19", "heat": null, "id": "bm_tabero-learning-gentle-manipulation-with-c_55e0b024", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": ["Robotics"], "links": {"code": "https://github.com/NathanWu7/Tabero", "data": null, "paper": "https://arxiv.org/abs/2605.27886", "pdf": "https://arxiv.org/pdf/2605.27886", "project": null, "report": "https://arxiv.org/abs/2605.27886"}, "metrics": [], "motivation": "Tactile sensing is essential for robots to achieve human-like gentle manipulation.", "name": "Tabero", "oneLine": "Tabero evaluates language-conditioned gentle manipulation using vision, touch, task success, and physical-interaction quality.", "paperTitle": "Tabero: Learning Gentle Manipulation with Closed-Loop Force Feedback from Vision, Touch, and Language", "primaryDomain": "Robotics & Embodied AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2605.27886", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-19T11:05:13.395391Z"}, "publications": [], "readiness": "Runnable", "recognitionConfidence": 0.55, "relation": "evaluates_only", "releasedAt": "2026-05-27", "reviewLifecycle": {"state": "pending"}, "source": {"authors": ["Qiwei Wu", "Rui Zhang", "Xin Xiang", "Tao Li", "Weihua Zhang", "Junjie Lai", "Renjing Xu"], "categories": ["cs.RO"], "id": "2605.27886", "title": "Tabero: Learning Gentle Manipulation with Closed-Loop Force Feedback from Vision, Touch, and Language", "type": "arxiv", "url": "https://arxiv.org/abs/2605.27886"}, "sourceUpdatedAt": "2026-05-27T03:08:21Z", "topics": ["Robotics"], "venueAttempts": [], "whyItMatters": "It adds force-sensitive evaluation to robotic manipulation rather than measuring task completion alone."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
