{"paper":{"title":"CHUCKLE -- When Humans Teach AI To Learn Emotions The Easy Way","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Ordering emotion samples by human annotator agreement boosts model performance and efficiency","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Ankush Pratap Singh, Houwei Cao, Yong Liu","submitted_at":"2025-10-10T13:38:06Z","abstract_excerpt":"Curriculum learning (CL) structures training from simple to complex samples, facilitating progressive learning. However, existing CL approaches for emotion recognition often rely on heuristic, data-driven, or model-based definitions of sample difficulty, neglecting the difficulty for human perception, a critical factor in subjective tasks like emotion recognition. We propose CHUCKLE (Crowdsourced Human Understanding Curriculum for Knowledge Led Emotion Recognition), a perception-driven CL framework that leverages annotator agreement and alignment in crowd-sourced datasets to define sample diff"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experimental results suggest that CHUCKLE enhances the performance of LSTMs and Transformers over non-curriculum baselines, while reducing the number of gradient updates, thereby enhancing both training efficiency and model robustness in both subject-dependent and subject-independent settings.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The central assumption that clips challenging for humans (measured by annotator disagreement and alignment) are similarly hard for neural networks; this premise is stated explicitly in the abstract and is required for the curriculum ordering to transfer from human perception to model training.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"CHUCKLE defines training sample difficulty for emotion recognition using crowdsourced annotator agreement and alignment, then applies this ordering to improve LSTM and Transformer performance while cutting gradient updates.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Ordering emotion samples by human annotator agreement boosts model performance and efficiency","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"e8f59e105ef2a1da732c654d7f73753d072596cb90aab019d87c18c26a5b6ade"},"source":{"id":"2510.09382","kind":"arxiv","version":3},"verdict":{"id":"2c40b0d1-0900-49fc-b1de-636e25b883a9","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-18T08:12:37.587689Z","strongest_claim":"Experimental results suggest that CHUCKLE enhances the performance of LSTMs and Transformers over non-curriculum baselines, while reducing the number of gradient updates, thereby enhancing both training efficiency and model robustness in both subject-dependent and subject-independent settings.","one_line_summary":"CHUCKLE defines training sample difficulty for emotion recognition using crowdsourced annotator agreement and alignment, then applies this ordering to improve LSTM and Transformer performance while cutting gradient updates.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The central assumption that clips challenging for humans (measured by annotator disagreement and alignment) are similarly hard for neural networks; this premise is stated explicitly in the abstract and is required for the curriculum ordering to transfer from human perception to model training.","pith_extraction_headline":"Ordering emotion samples by human annotator agreement boosts model performance and efficiency"},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2510.09382/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":2,"snapshot_sha256":"61384898f63b3df75c7d57a43201706acd1090c600cc3dc877eb76145b9c8d85"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}