{"data":{"event":{"id":"98dcba48-2044-4eda-bd2f-cef678c07f36","slug":"what-you-can-t-see-is-still-what-you-learn-a-preregistered-sixty-society-ce7eba46ee","title":"What You Can't See Is Still What You Learn: A Preregistered Sixty-Society Confirmation That Evidence Masking Drives Compositional Generalization","short_summary":"arXiv:2609.17637v1 Announce Type: new \nAbstract: Restricting what a module can read may improve what a system learns to compute. We test this in a preregistered confirmation with sixty four-cell systems sharing a frozen language-model backbone and communicating through learned continuous packets. Five conditions vary evidence masking, ownership markers, and replacement of foreign evidence with neutral filler, across six initialization clusters, each with two data orders, on one fresh task world. With markers available in both regimes, masking improved accuracy on held-out two- and three-operat","full_description":"arXiv:2609.17637v1 Announce Type: new \nAbstract: Restricting what a module can read may improve what a system learns to compute. We test this in a preregistered confirmation with sixty four-cell systems sharing a frozen language-model backbone and communicating through learned continuous packets. Five conditions vary evidence masking, ownership markers, and replacement of foreign evidence with neutral filler, across six initialization clusters, each with two data orders, on one fresh task world. With markers available in both regimes, masking improved accuracy on held-out two- and three-operation compositions by median paired differences of 0.846 and 0.859; all twelve pairs cleared the required margins, and the full preregistered behavioral criterion passed. The unmarked replication also passed. No globally visible system passed the marker-following check, so the effect of usable role information remains unresolved. The filler condition yielded seven full generalizers, but its decomposition criteria were inconclusive. Packet interventions in all eighteen audited masked systems followed the predicted intermediate-value changes on eligible cases; these finite, success-conditioned audits do not establish mediation. The results confirm a large advantage of the tested masking regime, while leaving its finer attribution and generality open. Protocols, results, and checkpoints are public.\n\nUpdate (2026-09-17): SNOMED CT Concept Recommendation from Masked Clinical Context. arXiv:2609.17855v1 Announce Type: new \nAbstract: Standardizing clinical language to SNOMED CT supports interoperability, analytics, and reusable phenotyping, but concept recommendation remains difficult when relevant concepts are rare or absent from training data. We present a masked-concept recommendation benchmark using the SNOMED CT Entity Linking Challenge v1.2.1 data derived from MIMIC-IV-Note. The dataset contains 75,491 annotations across 272 discharge summaries, with 204 notes used for training and 68 for historical testing. For each unique note-concept pair, the target mention is masked from a local clinical context and the system ranks SNOMED CT concepts observed during training. We compare a popularity baseline, sparse TF-IDF concept prototypes, dense latent semantic analysis embeddings, sparse-dense fusion, retrieved-note evidence, and a retrieval-augmented hybrid. Sparse TF-IDF performs best, achieving Recall@1 of 14.81%, Recall@10 of 33.43%, MRR of 0.2114, and nDCG@10 of 0.2297. Retrieval augmentation does not improve this baseline, with Recall@10 of 31.99% and MRR of 0.1937. Perform\n\nUpdate (2026-09-22): Hapi: A Multivariable Land-Surface Transformer for Medium-Range Hydrological Forecasting at Continental Scale. arXiv:2609.22702v1 Announce Type: new \nAbstract: Accurate flood forecasts several days in advance are essential for flood control, water-resource management, and emergency response. Producing them at high resolution over a continental domain calls for local hydrological detail together with spatial context extending from river basins to synoptic weather systems. We developed Hapi, a U-Net Swin Transformer that uses fine three-dimensional patches and hierarchical shifted-window attention to forecast discharge, surface runoff, snow water equivalent, and soil wetness across the contiguous United States. The model produces 24--72-hour forecasts at $0.05^{\\circ}$ resolution, with learned Laplacian task weights adjusting each variable's contribution to training. On 2024 test data using reconstructed weather and land-surface inputs from ERA5-Land, Hapi outperformed an operational physics-based model and a state-of-the-art AI model in flood detection. Independent validation against 3,881 U.S. Geological Survey gauges and a Hurricane Helene case study support\n\nUpdate (2026-09-23): X-Planner: Event-Structured Task Planning for Embodied Intelligence. arXiv:2609.25187v1 Announce Type: new \nAbstract: Task planning bridges high-level instructions and executable behavior in long-horizon manipulation, yet modern Vision-Language-Action (VLA) systems often leave this intermediate structure implicit. Existing chain-of-thought (CoT) planners also tend to rely on coarse task-level annotations or serialize long reasoning traces token by token. We present X-Planner, a planning front-end that addresses both the supervision and representation of embodied reasoning. Our planning data combine Ego, UMI, and teleoperation under a hierarchy granularity with source-dependent annotation depth. Takeover-time annotations and human-designed failures supervise ongoing error recognition. On the model side, a shared VLM backbone exposes two event-structured plan forms: a discrete interface that emits interpretable event states and a latent interface that relays continuous CoT states across staggered Transformer depths through Staircase Decoding. A frozen latent-to-text reconstruction objective provides a semantic anchor for the latent representation. Offline two-s","ledger_type":"benefit","primary_domain_id":"8f1af1b9-7302-4b51-aae9-fdfac04d158a","event_status":"provisional","event_date":"2026-09-17T00:00:00.000Z","discovery_date":"2026-09-17T00:00:00.000Z","first_published_date":"2026-09-17T00:00:00.000Z","last_reviewed_date":"2026-09-23T00:00:00.000Z","geographic_scope":"International","affected_population":null,"base_impact_tier":1,"base_score":"1.00","attribution_multiplier":"0.1000","evidence_multiplier":"0.4000","realization_multiplier":"0.2000","durability_multiplier":"0.5000","current_event_score":"0.004000","confidence_level":"low","score_explanation":"Ingest update: independent source raised evidence; added a follow-up source. Score remains provisional until review. Follow-up ingest: independent source raised evidence; added a follow-up source. Follow-up ingest: independent source raised evidence; added a follow-up source.","methodology_version_id":"8745541b-0f8e-4d5a-9b47-5291d3813a8d","original_methodology_version_id":"8745541b-0f8e-4d5a-9b47-5291d3813a8d","published_at":"2026-09-17T04:00:51.100Z","created_at":"2026-09-17T04:00:51.100Z","updated_at":"2026-09-23T04:01:04.383Z","flags":[],"domain_name":"Biology","domain_slug":"biology","methodology_version":"0.2"},"contributions":[{"id":"2e337af9-5d20-430b-aa63-08e8a5ca3366","event_id":"98dcba48-2044-4eda-bd2f-cef678c07f36","model_id":"535014f7-c941-4d39-aa9b-8f9c1d3d07b7","role_description":"Unspecified system mentioned or implied by a news item. Remap to a named release when identified.","attribution_multiplier":"0.1000","credit_share":"1.0000","contribution_score":"0.004000","attribution_rationale":"News ingest does not infer a named model from the publisher alone. Attribution stays unspecified until a release is identified.","attribution_confidence":"medium","first_used_date":"2026-09-17T00:00:00.000Z","model_version_if_known":null,"review_status":"approved","created_at":"2026-09-17T04:00:51.113Z","model_slug":"unspecified-ai-system","model_name":"Unspecified AI system","identity_class":"unknown","is_internal":false,"family_name":"Unspecified","family_slug":"unknown-unspecified","organization_name":"Unknown","organization_slug":"unknown"}],"sources":[{"id":"c0cc4616-b2a2-4029-aaff-2f4ea92836fc","event_id":"98dcba48-2044-4eda-bd2f-cef678c07f36","url":"https://arxiv.org/abs/2609.17637","canonical_url":"https://arxiv.org/abs/2609.17637","source_type":"preprint","publisher":"arXiv cs.AI","author":null,"publication_date":"2026-09-17T00:00:00.000Z","retrieved_at":"2026-09-17T04:00:51.119Z","title":"What You Can't See Is Still What You Learn: A Preregistered Sixty-Society Confirmation That Evidence Masking Drives Compositional Generalization","excerpt":"arXiv:2609.17637v1 Announce Type: new \nAbstract: Restricting what a module can read may improve what a system learns to compute. We test this in a preregistered confirmation with sixty four-cell systems sharing a frozen language-model backbone and communicating through learned continuous packets. Five conditions vary evidence masking, ownership markers, and replacement of foreign evidence with neutral filler, across six initialization clusters, each with two data orders, on one fresh task world.","content_hash":null,"source_reliability_class":"medium","is_primary_source":true,"is_independent":true,"is_peer_reviewed":false,"archived_url":null,"created_at":"2026-09-17T04:00:51.119Z"},{"id":"d2d56a66-e6c9-4042-b107-f4d0c6db3cf0","event_id":"98dcba48-2044-4eda-bd2f-cef678c07f36","url":"https://arxiv.org/abs/2609.17855","canonical_url":"https://arxiv.org/abs/2609.17855","source_type":"preprint","publisher":"arXiv cs.AI","author":null,"publication_date":"2026-09-17T00:00:00.000Z","retrieved_at":"2026-09-17T04:00:52.346Z","title":"SNOMED CT Concept Recommendation from Masked Clinical Context","excerpt":"arXiv:2609.17855v1 Announce Type: new \nAbstract: Standardizing clinical language to SNOMED CT supports interoperability, analytics, and reusable phenotyping, but concept recommendation remains difficult when relevant concepts are rare or absent from training data. We present a masked-concept recommendation benchmark using the SNOMED CT Entity Linking Challenge v1.2.1 data derived from MIMIC-IV-Note. The dataset contains 75,491 annotations across 272 discharge summaries, with 204 notes used for t","content_hash":null,"source_reliability_class":"medium","is_primary_source":false,"is_independent":false,"is_peer_reviewed":false,"archived_url":null,"created_at":"2026-09-17T04:00:52.346Z"},{"id":"eecc1108-bd2f-4f3e-ad12-df218734c7b3","event_id":"98dcba48-2044-4eda-bd2f-cef678c07f36","url":"https://arxiv.org/abs/2609.22702","canonical_url":"https://arxiv.org/abs/2609.22702","source_type":"preprint","publisher":"arXiv cs.AI","author":null,"publication_date":"2026-09-22T00:00:00.000Z","retrieved_at":"2026-09-23T03:01:09.280Z","title":"Hapi: A Multivariable Land-Surface Transformer for Medium-Range Hydrological Forecasting at Continental Scale","excerpt":"arXiv:2609.22702v1 Announce Type: new \nAbstract: Accurate flood forecasts several days in advance are essential for flood control, water-resource management, and emergency response. Producing them at high resolution over a continental domain calls for local hydrological detail together with spatial context extending from river basins to synoptic weather systems. We developed Hapi, a U-Net Swin Transformer that uses fine three-dimensional patches and hierarchical shifted-window attention to forec","content_hash":null,"source_reliability_class":"medium","is_primary_source":false,"is_independent":false,"is_peer_reviewed":false,"archived_url":null,"created_at":"2026-09-23T03:01:09.280Z"},{"id":"8902ef9c-880d-4c20-ae69-8dea22c42cf2","event_id":"98dcba48-2044-4eda-bd2f-cef678c07f36","url":"https://arxiv.org/abs/2609.25187","canonical_url":"https://arxiv.org/abs/2609.25187","source_type":"preprint","publisher":"arXiv cs.AI","author":null,"publication_date":"2026-09-23T00:00:00.000Z","retrieved_at":"2026-09-23T04:01:04.366Z","title":"X-Planner: Event-Structured Task Planning for Embodied Intelligence","excerpt":"arXiv:2609.25187v1 Announce Type: new \nAbstract: Task planning bridges high-level instructions and executable behavior in long-horizon manipulation, yet modern Vision-Language-Action (VLA) systems often leave this intermediate structure implicit. Existing chain-of-thought (CoT) planners also tend to rely on coarse task-level annotations or serialize long reasoning traces token by token. We present X-Planner, a planning front-end that addresses both the supervision and representation of embodied ","content_hash":null,"source_reliability_class":"medium","is_primary_source":false,"is_independent":false,"is_peer_reviewed":false,"archived_url":null,"created_at":"2026-09-23T04:01:04.366Z"}],"claims":[{"id":"7f664a9d-375b-4866-ad9b-71621460f98e","event_id":"98dcba48-2044-4eda-bd2f-cef678c07f36","claim_text":"What You Can't See Is Still What You Learn: A Preregistered Sixty-Society Confirmation That Evidence Masking Drives Compositional Generalization","claim_type":"outcome","claim_status":"supported","confidence_score":null,"created_at":"2026-09-17T04:00:51.125Z","updated_at":"2026-09-17T04:00:51.125Z"}],"revisions":[{"id":"1343f0c5-0d5f-4585-bed3-0b3887df08f8","event_id":"98dcba48-2044-4eda-bd2f-cef678c07f36","model_id":null,"previous_score":"0.000000","new_score":"0.001000","previous_factors":{},"new_factors":{"tier":1,"evidence":0.1,"base_score":1,"durability":0.5,"attribution":0.1,"realization":0.2},"change_reason":"Auto-published from news ingest.","trigger_type":"news_ingest","trigger_source_ids":null,"reviewer_id":null,"review_status":"published","created_at":"2026-09-17T04:00:51.131Z"},{"id":"f4e73cab-0fa4-4e35-a693-c98c40f73a52","event_id":"98dcba48-2044-4eda-bd2f-cef678c07f36","model_id":null,"previous_score":"0.001000","new_score":"0.002000","previous_factors":{"tier":1,"evidence":0.1,"base_score":1,"durability":0.5,"attribution":0.1,"realization":0.2},"new_factors":{"tier":1,"models":["535014f7-c941-4d39-aa9b-8f9c1d3d07b7"],"evidence":0.2,"base_score":1,"durability":0.5,"attribution":0.1,"realization":0.2},"change_reason":"Ingest update: independent source raised evidence; added a follow-up source.","trigger_type":"ingest_update","trigger_source_ids":["d2d56a66-e6c9-4042-b107-f4d0c6db3cf0"],"reviewer_id":null,"review_status":"published","created_at":"2026-09-17T04:00:52.367Z"},{"id":"b1febc44-2940-4411-98c8-721a435d510f","event_id":"98dcba48-2044-4eda-bd2f-cef678c07f36","model_id":null,"previous_score":"0.002000","new_score":"0.003000","previous_factors":{"tier":1,"evidence":0.2,"base_score":1,"durability":0.5,"attribution":0.1,"realization":0.2},"new_factors":{"tier":1,"models":["535014f7-c941-4d39-aa9b-8f9c1d3d07b7"],"evidence":0.3,"base_score":1,"durability":0.5,"attribution":0.1,"realization":0.2},"change_reason":"Ingest update: independent source raised evidence; added a follow-up source.","trigger_type":"ingest_update","trigger_source_ids":["eecc1108-bd2f-4f3e-ad12-df218734c7b3"],"reviewer_id":null,"review_status":"published","created_at":"2026-09-23T03:01:09.305Z"},{"id":"da8c6564-1482-4d94-aa77-1506e9184a6b","event_id":"98dcba48-2044-4eda-bd2f-cef678c07f36","model_id":null,"previous_score":"0.003000","new_score":"0.004000","previous_factors":{"tier":1,"evidence":0.3,"base_score":1,"durability":0.5,"attribution":0.1,"realization":0.2},"new_factors":{"tier":1,"models":["535014f7-c941-4d39-aa9b-8f9c1d3d07b7"],"evidence":0.4,"base_score":1,"durability":0.5,"attribution":0.1,"realization":0.2},"change_reason":"Ingest update: independent source raised evidence; added a follow-up source.","trigger_type":"ingest_update","trigger_source_ids":["8902ef9c-880d-4c20-ae69-8dea22c42cf2"],"reviewer_id":null,"review_status":"published","created_at":"2026-09-23T04:01:04.390Z"}],"secondary":[]},"methodology_version":"0.1"}