AI-powered literature discovery and review engine for medical/scientific papers
paperai is an AI-powered literature discovery and review engine for medical/scientific papers. paperai helps automate tedious literature reviews allowing researchers to focus on their core work. Queries are run to filter papers with specified criteria. Reports powered by extractive question-answering are run to identify answers to key questions within sets of medical/scientific papers.
paperai was used to analyze the COVID-19 Open Research Dataset (CORD-19), winning multiple awards in the CORD-19 Kaggle challenge.
paperai and/or NeuML has been recognized in the following articles:
The easiest way to install is via pip and PyPI
pip install paperai
You can also install paperai directly from GitHub. Using a Python Virtual Environment is recommended.
pip install git+https://github.com/neuml/paperai
Python 3.6+ is supported
See this link to help resolve environment-specific install issues.
A Dockerfile with commands to install paperai, all dependencies and scripts is available in this repository.
Clone this git repository and run the following to build and run the Docker image.
docker build -t paperai -f docker/Dockerfile . docker run --name paperai --rm -it paperai
This will bring up a paperai command shell. Standard Docker commands can be used to copy files over or commands can be run directly in the shell to retrieve input content. All scripts in the following examples are available in this environment.
paperetl's Dockerfile can be combined with this Dockerfile to have a single image that can index and query content. The files from the paperetl project scripts directory needs to be placed in paperai's scripts directory. The paperetl Dockerfile also needs to be copied over (it's referenced as paperetl.Dockerfile here).
docker build -t base -f docker/Dockerfile . docker build -t paperai --build-arg BASE_IMAGE=base -f docker/paperetl.Dockerfile . docker run --name paperai --rm -it paperai
The following notebooks and applications demonstrate the capabilities provided by paperai.
|CORD-19 Analysis with Sentence Embeddings||Builds paperai-based submissions for the CORD-19 Challenge|
|CORD-19 Report Builder||Template for building new reports|
|Search||Search a paperai index. Set query parameters, execute searches and display results.|
paperai indexes databases previously built with paperetl. paperai currently supports querying SQLite databases.
The following sections show how to build an index for a SQLite articles database.
This example assumes the database and model path is cord19/models. Substitute as appropriate.
Download CORD-19 fastText vectors
A full vector model build can optionally be run with the following command.
python -m paperai.vectors cord19/models
CORD-19 fastText vectors are also available on Kaggle.
Build embeddings index
python -m paperai.index cord19/models cord19/vectors/cord19-300d.magnitude
The paperai.index process takes two optional arguments, the model path and the vector file path. The default model location is ~/.cord19 if no parameters are passed in.
Reports support generating output in multiple formats. An example report call:
python -m paperai.report tasks/risk-factors.yml 50 md cord19/models
The following report formats are supported:
In the example above, a file named tasks/risk-factors.md will be created.
The fastest way to run queries is to start a paperai shell
A prompt will come up. Queries can be typed directly into the console.
The tech stack is built on Python and creates a sentence embeddings index with FastText + BM25. Background on this method can be found in this Medium article.
The model is a combination of a sentence embeddings index and a SQLite database with the articles. Each article is parsed into sentences and stored in SQLite along with the article metadata. FastText vectors are built over the full corpus. The sentence embeddings index only uses tagged articles, which helps produce the most relevant results.
Multiple entry points exist to interact with the model.