You need to agree to share your Hugging Face information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

CoRL Papers β€” Conference on Robot Learning (PDF archive)

A full-text PDF archive of accepted papers from the Conference on Robot Learning (CoRL), organized by year, with per-paper metadata. 2020–2025 are mirrored from the official PMLR proceedings; 2026 is from OpenReview (no PMLR volume yet). For personal research and archival use.

Coverage

Year Edition Venue Source Papers PDFs
2026 10th Austin, USA OpenReview 687 686
2025 9th Seoul, Korea PMLR v305 263 263
2024 8th Munich, Germany PMLR v270 264 264
2023 7th Atlanta, USA PMLR v229 199 199
2022 6th Auckland, NZ PMLR v205 197 197
2021 5th London, UK PMLR v164 166 166
2020 4th Virtual PMLR v155 165 165
Total 1941 1940

CoRL 2026 notes

  • The conference is Nov 9–12, 2026. The paper list comes from the official accepted-papers page, with 21 orals, 72 spotlights and 3 poster sessions of 229 papers each. PDFs, authors and abstracts come from the camera-ready submissions on OpenReview.
  • 1 paper has no PDF: S3FVTypVnZ (UniLatent: Cross-Embodiment Transfer via Latent Observation Alignment) has no PDF on OpenReview or arXiv. It is marked status = FAILED in papers.csv.
  • Titles are the OpenReview camera-ready titles. Some differ from the program listing, which is kept in site_title.

Structure

<year>/
  papers.csv        # per-paper metadata (one row per accepted paper)
  <key>_<slug>.pdf  # main paper PDF, e.g. huang21a_learning-a-decision-module.pdf
  tags.json         # 2020–2025: per-paper topic tags
  analysis.json     # 2020–2025: aggregate topic statistics
  abstracts.json    # 2026: {key: {title, abstract, keywords}}
  • <key> is the PMLR paper id (e.g. huang21a) for 2020–2025 and the OpenReview forum id (e.g. 4RLjGPEfpH) for 2026.
  • Only the main paper PDF is archived. Supplementary -supp.pdf files are excluded.

papers.csv columns

Column Years Description
key all PMLR id / OpenReview forum id
title, authors all Paper title, comma-separated authors
filename all PDF file name in the year folder
pdf_url, abs_url all Source PDF and abstract/forum page
status, detail all Download status (FAILED = no PDF)
site_title 2026 Title as listed in the conference program
session, session_title 2026 Oral/Spotlight session (e.g. Oral O3), empty for poster-only papers
poster_session, poster_no 2026 Poster session (1–3) and board number

Topic files (2020–2025)

tags.json is a list of {key, title, primary_area, app_tags, method_tags, modality_tags}, produced by LLM multi-label tagging of the abstracts. analysis.json holds counts per tag (primary_area, app_tags, method_tags, modality_tags) plus embedding-based clusters (MiniLM + KMeans).

Usage

from huggingface_hub import snapshot_download
import pandas as pd

path = snapshot_download("Ngseo/corl-papers", repo_type="dataset",
                         allow_patterns=["2026/*.csv", "2026/*.json"])  # metadata only
df = pd.read_csv(f"{path}/2026/papers.csv")
orals = df[df.session.str.startswith("Oral", na=False)]

# PDFs for one year (large: 2026 is ~8 GB)
path = snapshot_download("Ngseo/corl-papers", repo_type="dataset",
                         allow_patterns=["2024/*"])

Source & license

PDFs come from the official PMLR proceedings (https://proceedings.mlr.press/, volumes v155/v164/v205/v229/v270/v305) and, for 2026, from OpenReview. Copyright remains with the respective authors. CoRL/PMLR papers are typically released by their authors under CC BY. This archive is a convenience mirror, so please cite the original papers and CoRL.

Downloads last month
3,969