wtf is ai-bio-conference-papers?

This repository is a curated dataset, not a piece of software you run. It collects 3,722 research papers at the intersection of artificial intelligence and biology that were accepted at three major AI conferences, ICLR, ICML, and NeurIPS, spanning from 2010 through 2026. Each paper entry includes its title, authors, abstract, and a link to the PDF. The papers were found using a two stage keyword filter built to catch biology related work across a huge pool of AI conference papers, and the README states this filter was checked against a smaller, hand reviewed set of 315 papers and correctly found about 96 percent of them. Every paper is tagged with one or more of ten topic labels, such as protein, molecule, single cell, clinical, and neuro, based on which biology keywords appear in its abstract. The data is provided in a few different forms: a master JSON and CSV file containing every paper record, and a set of markdown files organized by conference and year that you can browse directly on GitHub, where each paper appears as a collapsible entry you click to expand and read the full abstract. A table in the README lets you jump straight to the file for any specific venue and year, along with the paper count for that year. The README also includes charts showing how the number of papers and specific topics, like protein related research, have grown over time, especially picking up noticeably after 2021. The intended use is for researchers, students, or anyone curious about trends in AI applied to biology to browse, search, or load this dataset into their own analysis scripts, since it can be loaded directly as JSON in a couple lines of Python. The data is released under a combination of the MIT license and CC-BY-4.0.