CoolFace
Datasetpublic

JensCS/top-ml-conference-papers-2023

Top ML Conference Papers 2023 A dataset of 7,035 papers from ACL 2023, CVPR 2023, ICLR 2023, and NeurIPS 2023, processed with KNOWHERE. Series This collection of KNOWHERE-processed paper datasets is continuously expanding. New datasets and editions will be released — stay tuned! Top ML Conference Papers 2023 (this) https://huggingface.co/datasets/JensCS/top-ml-conference-papers-2023… See the full description on the dataset page: https://huggingface.co/datasets/JensCS/top-ml-conference-papers-2023.

sourceHugging Facecc-by-4.0updated 1mo agoView on Hugging Face
10likes158downloads
Dataset Card

Top ML Conference Papers 2023

A dataset of 7,035 papers from ACL 2023, CVPR 2023, ICLR 2023, and NeurIPS 2023, processed with **KNOWHERE**.

Series

This collection of KNOWHERE-processed paper datasets is continuously expanding. New datasets and editions will be released — stay tuned!

Top ML Conference Papers

Embodied AI Conference Papers

About KNOWHERE

<p align="center"> <img src="https://github.com/user-attachments/assets/896e64d2-e50e-4158-b71c-bc69e11c7c65" alt="KNOWHERE" width="800"/> </p>

<h2 align="center"><strong>Prepare unstructured data for AI Agents</strong></h2>

<p align="center"><strong>KNOWHERE is the memory layer between complex, dirty documents and AI agents.</strong></p>

<p align="center">It ingests unstructured documents and produces persistent, navigable memory: parsing, hierarchy extraction, multi-modal structuring, and graph construction in a single pipeline. Every chunk retains full semantic context, making the output a natural fit for Agentic RAG, vector-based RAG, or any LLM workflow.</p>

<p align="center"><strong>GitHub Repo:</strong> <a href="https://github.com/Ontos-AI/knowhere">https://github.com/Ontos-AI/knowhere</a></p>

ConferencePapersDomain
ACL 20231,044Natural Language Processing
CVPR 20232,248Computer Vision
ICLR 2023545Machine Learning
NeurIPS 20233,198Machine Learning

Construction

Built with KNOWHERE:

  1. 1.PDF Parsing — KNOWHERE parses each PDF into structured chunks (chunks.json): text blocks with section paths. The dataset retains each chunk's content and path.
  2. 2.Hierarchy Extraction — KNOWHERE outputs a doc_nav.json section tree per paper, with title, path, level, summary, chunk_count, and recursive children.
  3. 3.Agentic Extraction — An LLM pipeline classifies papers into 4 types and extracts fields from KNOWHERE's parsed chunks and section hierarchy.
  4. 4.Assembly — Extraction fields + KNOWHERE chunks + hierarchy merge into unified records.

Applications

  • —Literature analysis and survey support — structured extraction fields (method, results, baselines) enable systematic literature review, cross-conference trend tracking, and automated survey generation
  • —Benchmark for PDF parsing & scientific understanding — evaluate document parsing pipelines and information extraction models on real academic papers with ground-truth hierarchy and typed fields
  • —Foundation for AI for Science experiments — structured paper representations serve as input for downstream tasks such as paper generation, research idea proposal, and citation-aware knowledge construction

Field Schema

All 24 extraction fields appear in every record. Fields not applicable to a paper's type are empty strings (or empty lists).

Common (all papers)

FieldTypeDescription
paper_titlestringVerbatim title
authorslist[string]Author names
affiliationstringFirst author's institution
research_domainstringAI/CS subfield
task_typestringSpecific technical task
core_problemstringUnresolved limitation
key_innovationstringPrimary contribution
code_repositorystringCode URL (if stated)

model_architecture (1,004 papers)

FieldDescription
model_nameProposed model name
architecture_typeDesign paradigm (e.g., Transformer)
model_size_parametersParameter count
training_evaluation_datasetDatasets used
key_resultsQuantitative results

methodandpipeline (5,196 papers)

FieldDescription
method_nameProposed method name
method_summaryCore mechanism
baseline_modelslist[string] — baselines compared
key_resultsAdvantage over baselines

theoryandanalysis (798 papers)

FieldDescription
analysis_targetObject/phenomenon studied
theoretical_toolsProof techniques
key_findingsTheorems or conclusions
prior_work_comparisonComparison to prior results
limitationsAssumptions or scope

agent_system (37 papers)

FieldDescription
agent_framework_nameProposed framework name
environment_or_toolsEnvironments/tools used
planning_mechanismReasoning strategy
eval_benchmarkEvaluation benchmarks
key_resultsBenchmark performance

Structural & Metadata

FieldTypeDescription
paper_typestringClassification key
type_namestringHuman-readable type name
extraction_qualityfloatNon-null field fraction
total_charsintPaper character count
conferencestringSource conference
chunkslist[{content, path}]KNOWHERE text chunks
hierarchylist[{title, path, level, summary, chunk_count}]KNOWHERE section tree

Formats

  • —JSONL — Nested hierarchy tree with recursive children, human-readable.
  • —Parquet — Columnar format, Zstd compression; hierarchy flattened to depth-first list.

Pipeline

KNOWHERE → Parse → Structure → Build Memory → Agentic extraction → Dataset assembly.

License

CC BY 4.0. Original papers retain their respective copyrights.