Metraipolis › archive › dc_d1871da893ae

Parser, Chunking, and Embedding Interactions in Retrieval-Augmented Generation over Indian Government Regulatory Documents — arXiv:2609.31660

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.31660
License (as archived)
CC0-1.0
Archived
2026-10-07T20:59:18.310Z
SHA-256 of the archived text
879eb636e4623f32a1979b99a95d2dc54b83e99f1c0ac2d67ce2afb57ffdfcc5
Stable URL of this page
https://metraipolis.com/documento/dc_d1871da893ae

How to cite

Parser, Chunking, and Embedding Interactions in Retrieval-Augmented Generation over Indian Government Regulatory Documents — arXiv:2609.31660 — arXiv (en), https://arxiv.org/abs/2609.31660 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d1871da893ae (Metraipolis archive dc_d1871da893ae, sha256 879eb636e4623f32a1979b99a95d2dc54b83e99f1c0ac2d67ce2afb57ffdfcc5)
BibTeX
@misc{dcd1871da893ae,
  title = {Parser, Chunking, and Embedding Interactions in Retrieval-Augmented Generation over Indian Government Regulatory Documents — arXiv:2609.31660},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.31660},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d1871da893ae; sha256 879eb636e4623f32a1979b99a95d2dc54b83e99f1c0ac2d67ce2afb57ffdfcc5},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.31660. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Retrieval-augmented generation (RAG) pipelines are typically assembled from independently-chosen components -- a document parser, a chunking strategy, and an embedding model -- yet these choices are rarely evaluated jointly, and evaluations that do combine them are usually run on a single document or corpus. We present a controlled factorial study of 3 parsers, 3 chunking strategies, and 5 dense embedding models, together with a sparse BM25 baseline, evaluated against 800 question instances, each with one or more required evidence strings, with evidence strings automatically validated against source text and a 10% random sample manually reviewed, across four structurally distinct Indian central-government regulatory documents. We fit linear mixed-effects models with document-query-level random intercepts to the resulting 72,000-row result set, run Holm-corrected paired comparisons between matched dense and sparse configurations, and report clustered bootstrap confidence intervals for all 54 unique retriever configurations. We find that no single retriever family dominates across documents; parser and chunker choice interact significantly; MPNet-base is a consistent underperformer with a severe failure mode on table-derived questions; and the corpus exhibits a near-saturated evidence-preservation ceiling above 98%, indicating that retrieval differences are driven primarily by ranking quality rather than information loss during ingestion. We additionally report embedding-dimension and chunk-size/overlap ablations and an efficiency/quality Pareto analysis. We release our full evaluation harness, corpus manifest, and 800-question benchmark.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in cs.AI).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Information Retrieval (cs.IR).
Pagina del lavoro: https://arxiv.org/abs/2609.31660 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T20:29:02.006Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).