{"paper":{"title":"SiFo: Wireless Foundation Model for Low-Overhead Site-Specific CSI Feedback","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"SiFo pretrains a CSI feedback model across sites and adapts it to new deployments by matching users to calibration samples via RSRP measurements without parameter updates.","cross_cats":[],"primary_cat":"eess.SP","authors_text":"Cheng-Jie Zhao, Yuanwei Liu, Zhaolin Wang, Zongyao Zhao","submitted_at":"2026-05-15T16:23:10Z","abstract_excerpt":"SiFo, a wireless foundation model-based framework, is proposed for low-overhead site-specific channel state information (CSI) feedback. In 3GPP NR, Type-II feedback provides an expressive codebook-based CSI representation, but it requires substantial reference-signal overhead, UE-side search, and feedback. Learning-based site-specific feedback can reduce these online costs while retaining high-quality subspace representation by exploiting deployment-dependent propagation structure. However, existing site-specific designs typically train a dedicated neural network for each new site, which limit"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"SiFo achieves higher CSI-capture efficiency than separately trained site-specific learning baselines under the same target-site labeled budget, approaches the high-overhead 3GPP NR Type-II feedback reference using only RSRP measurements collected during online SSB probing, and converts the high CSI-capture efficiency and low overhead into effective spectral efficiency improvement under limited target-site data.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that RSRP fingerprints collected during standard SSB probing are sufficiently discriminative to match a served user to nearby calibration samples whose stored full-CSI subspace labels provide accurate site-specific guidance (abstract, paragraph describing online operation and calibration memory).","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"SiFo pretrains a CSI feedback model on source sites and uses RSRP-based user matching to calibration memory for site-specific subspace guidance at target sites without parameter updates.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"SiFo pretrains a CSI feedback model across sites and adapts it to new deployments by matching users to calibration samples via RSRP measurements without parameter updates.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"fae2b74b86482319c517e51d63a59fe1d40df4e96b515df87dd27367738aa7b9"},"source":{"id":"2605.16141","kind":"arxiv","version":1},"verdict":{"id":"4198e19a-62af-41a8-b7ce-f9a499483c18","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T21:42:36.007445Z","strongest_claim":"SiFo achieves higher CSI-capture efficiency than separately trained site-specific learning baselines under the same target-site labeled budget, approaches the high-overhead 3GPP NR Type-II feedback reference using only RSRP measurements collected during online SSB probing, and converts the high CSI-capture efficiency and low overhead into effective spectral efficiency improvement under limited target-site data.","one_line_summary":"SiFo pretrains a CSI feedback model on source sites and uses RSRP-based user matching to calibration memory for site-specific subspace guidance at target sites without parameter updates.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that RSRP fingerprints collected during standard SSB probing are sufficiently discriminative to match a served user to nearby calibration samples whose stored full-CSI subspace labels provide accurate site-specific guidance (abstract, paragraph describing online operation and calibration memory).","pith_extraction_headline":"SiFo pretrains a CSI feedback model across sites and adapts it to new deployments by matching users to calibration samples via RSRP measurements without parameter updates."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.16141/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T22:01:23.235148Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T21:51:59.522609Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"cited_work_retraction","ran_at":"2026-05-19T18:22:03.735106Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T17:33:31.048783Z","status":"skipped","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T16:41:55.452388Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"835d96ab9862d07407773d51f97940d56fb6e14cefec404c2bb569a55b51ee53"},"references":{"count":33,"sample":[{"doi":"","year":2021,"title":"On the Opportunities and Risks of Foundation Models","work_id":"a18039e9-928d-47c9-a836-32656a71bf71","ref_index":1,"cited_arxiv_id":"2108.07258","is_internal_anchor":true},{"doi":"","year":2024,"title":"WirelessLLM: Empowering large language models towards wireless intelligence,","work_id":"4ea1574e-c6c4-47d4-9d98-19a5d9e0d029","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"WiFo: Wireless foundation model for channel prediction,","work_id":"5e65c9c5-f9e4-4d3f-9b36-8ab180468235","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Large wireless model (LWM): A foundation model for wireless channels","work_id":"d43b57ab-b7d7-4512-96ef-8bcd1e300897","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"WirelessGPT: A generative pre-trained multi-task learning framework for wireless communication and sensing,","work_id":"090779c6-12ad-4482-bd5e-c07b953646dd","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":33,"snapshot_sha256":"01028f9b3b8abe45340462e3481c60781a64be7745fd39be82fc88f05e686336","internal_anchors":1},"formal_canon":{"evidence_count":2,"snapshot_sha256":"4223fba30c20e64bd03da5b03cd1e5cd20e9fb7ab7462803eea608be51ab7cea"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}