Conference ExplorersCoRL 2026 · Paper ExplorerMethodJSON ↓

CONFERENCE ON ROBOT LEARNING 2026

From generalist policies
to reliable physical action.

A map of 687 papers. Compare the full programme with Oral and Spotlight highlights,
and explore topics, papers, authors and institutions.

CoRL IN CONTEXT · 2024 → 2026

The paper count is now 2.61× last year

+424 papers · +161.2% vs 2025

Reading the scale · Counts were nearly unchanged from 2024 to 2025 (264 → 263), then reached 687 in 2026. Compared with 2024, this is +423 papers · +160.2%. With more papers to explore, session, topic and author filters help you focus on your research interests.

Counting method · Sources · Interpretation

For 2024 and 2025, we counted paper entries in the official PMLR main-conference proceedings. For 2026, we counted unique papers in the official CoRL programme as of 15 October 2026. Workshop papers are excluded. Proceedings counts may differ from initial acceptance decisions; the 2026 figure is not a final proceedings count.

Growth = (687 ÷ previous-year public paper count − 1) × 100. The 2026 figures of 2,094 active submissions and a 32.8% acceptance rate come from the official announcement. Submission definitions have not been harmonized across years, so we do not compare submission growth or acceptance-rate trends here. The increase in paper count is documented; changes in research quality or the growth of individual topics require separate analysis.

Memory, context and transfer in generalist policies; world models coupled with control; and reliable data, evaluation, contact and whole-body execution are prominent themes. This interpretation draws on official sessions and paper content, not year-over-year topic growth.

ALL vs ORAL vs SPOTLIGHT

Compare topic distributions

The same title-keyword rules were applied to all 687 papers. Values are percentages within each group, with papers allowed to match multiple topics. Click a number to browse the matching papers.

Data and evaluation keywords: 10.6% overall → 28.6% in Oral. Memory and long context: 5.1% → 14.3% in Oral. Safety and robustness: 8.6% → 15.3% in Spotlight. Relevant work without these title keywords is not counted; read these figures alongside the official session themes.
Title-derived topicAll · 687Oral · 21Spotlight · 72Poster only · 594
Topic rules and interpretation limits

These are automatic, overlapping tags, not a semantic classification of every paper. Quantitative comparisons retain the original title-keyword rules for consistency. The Oral group is small: one of its 21 papers represents 4.8 percentage points. These figures do not explain selection decisions, research quality or annual topic growth.

OFFICIAL CORL THEMES

Explore by day and session

Official CoRL session themes are preserved verbatim. Times are local to Austin. Every Oral and Spotlight paper also has a poster presentation.

FOLLOW THE RESEARCHERS

Start with authors and institutions

Click a name to filter the papers. Affiliations are recorded at paper level, not mapped to individual authors. Counts reflect available public metadata and are not a ranking of total institutional output.

Authors

Institutions

SOURCES & LIMITATIONS

How to read this explorer

Programme titles, presentation types, days and session themes follow the official CoRL programme. When an external title differs, it is shown separately. Authors and abstracts for all 687 papers, plus 686 PDF links were refreshed from public OpenReview records on 18 October 2026. Affiliations and other links retain their separately cited sources. Each paper detail includes sources and affiliation evidence.

Institution aliases are normalized only from names found in paper affiliations or bylines. A paper with multiple institutions counts once for each, so totals may exceed the number of papers. Authors with identical names may be merged by name string. Bookmarks are stored only in this browser.

Download source-backed JSON · Download Excel