proomt

Search

Search posts, papers, and topics

All posts

Hugging Face Daily PapersHao Yu, Kang Liu, Linnan Zhao1 min readpaperadvanced

WeVisDoc: From Coverage to Capability for Robust End-to-End Document Parsing

Read on arxiv.org

Summary

WeVisDoc introduces a two‑stage data‑centric pipeline for end‑to‑end document parsing. Stage I expands coverage using heterogeneous data and structure‑preserving degradations. Stage II probes the Stage I model with a held‑out set, clusters residual errors, and directs targeted data creation and token‑budget reallocation. The 4‑billion‑parameter model reaches 95.38 Overall on OmniDocBench v1.6 and…

  • Coverage‑only data expansion leaves systematic blind spots; probing with a fixed visual‑structural split surfaces them.
  • Targeted data synthesis guided by error clusters yields measurable gains, especially on degraded documents.
  • A modest token‑budget reallocation (focusing on under‑represented token types) improves large‑scale parsers without changing architecture.

Document parsing under real‑world conditions (varied layouts, scans, noise) is a bottleneck for downstream automation. WeVisDoc shows that a disciplined, data‑centric feedback loop can close the gap between benchmark performance and field robustness, offering a repeatable recipe for other vision‑ML…

7/10
machine learningdata engineeringml infra

Related reading

  2. Glyph: A Multi-Strategy Agentic System for Column Description and Sensitivity-Ontology Tagging of Enterprise Data Catalogs

    Glyph is a production system that uses coordinated LLM agents and a fine‑tuned MiniLM encoder to automatically generate column descriptions and assign ontology tags in enterprise data catalogs. It combines code‑grounded retrieval, regex, and contrastive vector search, achieving NDCG@10 0.92 and MAP@100 0.90, and provides auditable provenance for each tag.

    Apple Machine Learning Researchapple.com1 minpaper

  3. Faster Maps: Chasing Swiss Speed

    ParparVM’s HashMap suffered catastrophic miss latency due to linear probing on dense integer keys. By adopting CPython‑style perturbed probing (Swiss‑table style) and extending tagged immediate values to more primitives, miss latency dropped from 32 s to ~45 ms, allocation pressure fell dramatically, and overall performance stayed roughly flat despite a modest hit‑time slowdown.

    CodeName Onecodenameone.com8 min

  4. When scanners miss the attack: how Cloudflare Client-Side Security protects storefronts

    Cloudflare’s Page Shield uses a graph‑neural‑network (GNN) to model JavaScript as a syntax‑tree graph, followed by a lightweight LLM for second‑opinion triage and an ensemble of frontier models for deep analysis. This pipeline caught eight malicious payloads across four distinct affiliate‑theft and backdoor techniques that traditional scanners missed, demonstrating the need for runtime, behavior‑…

    Cloudflarecloudflare.com21 minHN2

  5. Labeled matches: why is this not in every regex engine?

    The author shows how to label tokens (dates, money, emails, etc.) with a single regex pass using extended operators (`&` for intersection, `~` for complement) in the resharp library. A tiny benchmark compares 10 patterns against spaCy’s NER, reporting ~1.9 GB/s (≈4500× faster) on 8 threads. The post lists the concrete patterns and argues that for deterministic, regular‑language entities regex can…

    Lobstersiev.ee7 minHN5lobste.rs27

  6. Fathom: Per-Query Read Depth for Sparse Decoding over Offloaded KV Caches

    Fathom introduces a per-query read depth mechanism for sparse decoding over offloaded KV caches, allowing each query to adaptively decide how many bits of each key channel to read. This method significantly speeds up decoding for large language models with long contexts by reducing host memory traffic, achieving 1.67x faster GPU decoding on Qwen3-8B at one million tokens.

    Hugging Face Daily Papersarxiv.org1 minpaper