{"paper":{"title":"MIPIC: Matryoshka Representation Learning via Self-Distilled Intra-Relational and Progressive Information Chaining","license":"http://creativecommons.org/licenses/by/4.0/","headline":"MIPIC produces Matryoshka embeddings that remain competitive at any truncation size by aligning intra-layer relations through self-distillation and chaining semantics from deep to shallow layers.","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Hai An Vu, Linh Ngo Van, Minh-Phuc Truong, Phung Gia Huy, Thang Duc Tran, Thanh Hong Nguyen, Trung Le","submitted_at":"2026-04-27T12:07:40Z","abstract_excerpt":"Representation learning is fundamental to NLP, but building embeddings that work well at different computational budgets is challenging. Matryoshka Representation Learning (MRL) offers a flexible inference paradigm through nested embeddings; however, learning such structures requires explicit coordination of how information is arranged across embedding dimensionality and model depth. In this work, we propose MIPIC (Matryoshka Representation Learning via Self-Distilled Intra-Relational Alignment and Progressive Information Chaining), a unified training framework designed to produce structurally"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"MIPIC yields Matryoshka representations that are highly competitive across all capacities, with significant performance advantages observed under extreme low-dimensional.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That aligning intra-relational structures via top-k CKA self-distillation and incrementally chaining semantics from deeper to earlier layers will produce structurally coherent and semantically compact representations without introducing distortions or requiring heavy task-specific tuning.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"MIPIC trains nested Matryoshka representations via self-distilled intra-relational alignment with top-k CKA and progressive information chaining across depths, yielding competitive performance especially at extreme low dimensions.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"MIPIC produces Matryoshka embeddings that remain competitive at any truncation size by aligning intra-layer relations through self-distillation and chaining semantics from deep to shallow layers.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"19efbc5bb6a4d17cd2131b64050eb4a67fcac1757add79f647fa2597a598f131"},"source":{"id":"2604.24374","kind":"arxiv","version":2},"verdict":{"id":"4216dd7e-6530-44dd-8fc6-4265f83ee3b5","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-08T03:31:22.754461Z","strongest_claim":"MIPIC yields Matryoshka representations that are highly competitive across all capacities, with significant performance advantages observed under extreme low-dimensional.","one_line_summary":"MIPIC trains nested Matryoshka representations via self-distilled intra-relational alignment with top-k CKA and progressive information chaining across depths, yielding competitive performance especially at extreme low dimensions.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That aligning intra-relational structures via top-k CKA self-distillation and incrementally chaining semantics from deeper to earlier layers will produce structurally coherent and semantically compact representations without introducing distortions or requiring heavy task-specific tuning.","pith_extraction_headline":"MIPIC produces Matryoshka embeddings that remain competitive at any truncation size by aligning intra-layer relations through self-distillation and chaining semantics from deep to shallow layers."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.24374/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-21T06:41:31.633335Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T22:06:38.125027Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"34def373d91086ebef96a60c284c2ad6014e75262f99cd7a3d8030168451e316"},"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"}