{"paper":{"title":"DeepSlide: From Artifacts to Presentation Delivery","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"DeepSlide is a multi-agent system that plans time-budgeted narratives and generates synced slides and scripts to improve delivery while matching visual quality.","cross_cats":["cs.CL","cs.IR"],"primary_cat":"cs.AI","authors_text":"Haoseng Liu, Jiahang Li, Ming Yang, Weiguo Zheng, Yuzheng Cai, Zhiwei Zhang","submitted_at":"2026-04-01T13:38:36Z","abstract_excerpt":"Presentations are a primary medium for scholarly communication, yet most AI slide generators optimize the artifact (a visually plausible deck) while under-optimizing the delivery process (pacing, narrative, and presentation preparation). We present DeepSlide, a human-in-the-loop multi-agent system that supports preparing the full presentation process, from requirement elicitation and time-budgeted narrative planning, to evidence-grounded slide--script generation, attention augmentation, and rehearsal support. DeepSlide integrates (i) a controllable logical-chain planner with per-node time budg"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Across 20 domains and diverse audience profiles, DeepSlide matches strong baselines on artifact quality while consistently achieving larger gains on delivery metrics, improving narrative flow, pacing precision, and slide--script synergy with clearer attention guidance.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The dual-scoreboard benchmark cleanly separates static artifact quality from dynamic delivery excellence without overlap or bias in the evaluation metrics.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"DeepSlide introduces a multi-agent system for full presentation preparation that matches baselines on slide quality but improves narrative flow, pacing, and script synergy via a new dual-scoreboard benchmark.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"DeepSlide is a multi-agent system that plans time-budgeted narratives and generates synced slides and scripts to improve delivery while matching visual quality.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"6581e496fad8b08fcb1a5f7b4e1c920a9cc841812ae389aa80de66d6d7600413"},"source":{"id":"2605.15202","kind":"arxiv","version":1},"verdict":{"id":"50396fef-bd72-4cce-b45e-7e2899d8a37b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T17:58:19.195575Z","strongest_claim":"Across 20 domains and diverse audience profiles, DeepSlide matches strong baselines on artifact quality while consistently achieving larger gains on delivery metrics, improving narrative flow, pacing precision, and slide--script synergy with clearer attention guidance.","one_line_summary":"DeepSlide introduces a multi-agent system for full presentation preparation that matches baselines on slide quality but improves narrative flow, pacing, and script synergy via a new dual-scoreboard benchmark.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The dual-scoreboard benchmark cleanly separates static artifact quality from dynamic delivery excellence without overlap or bias in the evaluation metrics.","pith_extraction_headline":"DeepSlide is a multi-agent system that plans time-budgeted narratives and generates synced slides and scripts to improve delivery while matching visual quality."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.15202/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":94,"sample":[{"doi":"10.1093/oso/9780190936600.001.0001","year":2020,"title":"Public policy and superintelligent AI: A vector field approach","work_id":"d7919ddc-14ed-47de-b84c-2058b64e2dd8","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1063/1.1784305","year":2004,"title":"The craft of scientific presentations: Critical steps to succeed and critical errors to avoid.Physics Today, 57, 07 2004","work_id":"fa7b0a1e-d26e-477e-9f61-c78339fa1207","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Autopresent: Designing structured visuals from scratch","work_id":"fdaf43d6-f597-48cd-a2ec-dc1c99c5eeb9","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Presentations are not always linear! gnn meets llm for document-to-presentation transformation with attribution.arXiv preprint arXiv:2405.13095, 2024","work_id":"c6e8e42f-c826-433b-b5d0-ff94ac191b6a","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.18653/v1/2024.eacl-long.163","year":2024,"title":"Presentations by the humans and for the humans: Harnessing LLMs for generating persona-aware slides from documents","work_id":"fc77f98a-1eb6-473a-aebe-9bbb9e2ea9d2","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":94,"snapshot_sha256":"a0d1a5868ed3d8e4d9539d3c0c23daffb3511b51d6552799212d43ff1ea759ba","internal_anchors":1},"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"}