{"paper":{"title":"MIST: Reliable Streaming Decision Trees for Online Class-Incremental Learning via McDiarmid Bound","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Streaming decision trees achieve reliable splits for online class-incremental learning by using a K-independent McDiarmid bound on Gini impurity together with Bayesian inheritance and quantile sketches.","cross_cats":["math.ST","stat.TH"],"primary_cat":"cs.LG","authors_text":"Chi-Nguyen Tran, Dao Sy Duy Minh, Huynh Trung Kiet, Long Tran-Thanh, Nguyen Lam Phu Quy, Phu-Hoa Pham","submitted_at":"2026-05-12T06:45:00Z","abstract_excerpt":"Streaming decision trees are natural candidates for open-world continual learning, as they perform local updates, enjoy bounded memory, and static decision boundaries. Despite these, they still fail in online class-incremental learning due to two coupled miscalibrations: (i) their split criterion grows unreliable as the class count K expands, and (ii) the absence of knowledge transfer at split time. Both failures share a common root: the range of Information Gain intrinsically scales with log2 K. Consequently, any Hoeffding-style confidence radius derived from it must inevitably grow with the "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"MIST resolves both failures through three integrated components: (i) a tight, K-independent McDiarmid confidence radius for Gini splitting that acts as a structural regulariser; (ii) a Bayesian inheritance protocol that projects parent statistics to child nodes via truncated-Gaussian moments, with variance reduction guarantees strongest precisely when splitting is most conservative; and (iii) per-leaf KLL quantile sketches that support both continuous threshold evaluation and geometry-adaptive leaf prediction from a single data structure.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that a McDiarmid bound applied to Gini impurity yields a practically tight and K-independent radius under the streaming, non-stationary data regime, and that the truncated-Gaussian moment projection accurately captures the statistical relationship between parent and child nodes without introducing bias that grows with class count.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"MIST fixes unreliable splits in streaming decision trees for class-incremental learning by using a K-independent McDiarmid bound on Gini impurity, Bayesian moment projection for knowledge transfer, and KLL quantile sketches for adaptive leaf predictions.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Streaming decision trees achieve reliable splits for online class-incremental learning by using a K-independent McDiarmid bound on Gini impurity together with Bayesian inheritance and quantile sketches.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"25e61609b000ebe80fd0728961c8c25189a7c36e848757b292c9c1cb260ef876"},"source":{"id":"2605.11617","kind":"arxiv","version":2},"verdict":{"id":"6ee00463-7afb-4605-bc58-eb64f46d28cf","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-13T01:15:42.621195Z","strongest_claim":"MIST resolves both failures through three integrated components: (i) a tight, K-independent McDiarmid confidence radius for Gini splitting that acts as a structural regulariser; (ii) a Bayesian inheritance protocol that projects parent statistics to child nodes via truncated-Gaussian moments, with variance reduction guarantees strongest precisely when splitting is most conservative; and (iii) per-leaf KLL quantile sketches that support both continuous threshold evaluation and geometry-adaptive leaf prediction from a single data structure.","one_line_summary":"MIST fixes unreliable splits in streaming decision trees for class-incremental learning by using a K-independent McDiarmid bound on Gini impurity, Bayesian moment projection for knowledge transfer, and KLL quantile sketches for adaptive leaf predictions.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that a McDiarmid bound applied to Gini impurity yields a practically tight and K-independent radius under the streaming, non-stationary data regime, and that the truncated-Gaussian moment projection accurately captures the statistical relationship between parent and child nodes without introducing bias that grows with class count.","pith_extraction_headline":"Streaming decision trees achieve reliable splits for online class-incremental learning by using a K-independent McDiarmid bound on Gini impurity together with Bayesian inheritance and quantile sketches."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.11617/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-19T11:40:50.257254Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T09:31:18.077895Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T08:17:50.935505Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"be9aa7012b0628e329a96be3ca1a35cacf03ab5f13d4e9981fb88330dc588b0e"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"4840c08c042eb84f2c72a8e5ee2ac2810e9da83fe101b8dd3c9bfad9fdc306df"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}