You can find here the back of omniscience.academy. We are trying to be a more transparent and collaborative Google scholar.
Research work is one of humanity's most valuable assets. This content must be referenced and easily accessible. Our algorithm is customizable. The calibration parameters are transparent and visible to all.
The algorithm is calibrated by default to perform literature reviews.
We rely on SemanticScholar and Microsoft Academic databases for our article database.
This is a Beta, please open an Issue for any issue!
When you search for a title, we look for articles that match the semantics of your query and also match your keywords. We search a thousand articles, corresponding to these criteria, then we organize them by citations, by novelty. We give bonuses to articles that have been cited by other quality articles.
The seminal tag means that the article has been at least 3 other important articles that appear in the final search results. The seminal tag allows you to see the seminal articles of a field of study at a glance. We have noticed that it often has false negatives, but rarely false positives.
The questions were generated by the language models in advance of your search. We start by summarizing the abstracts, then we extract the question that the abstract answers.
- On the Back repo:
source djenv/bin/activate
python manage.py runserver
- On the front repo
The front is also open source. You can find it here: here.
First adapt the ./.env.local file, change the first and unique line to:
NEXT_PUBLIC_API_URL=http://127.0.0.1:8000
(before it was: NEXT_PUBLIC_API_KEY=https://aws.omniscienceapidev.com)
npm install
npm run dev
And go to htttp://localhost:3000 on your browser.
If you want to change substantially the algorithm, you can fork us and deploy on heroku. Ask us by email <omniscience.project@g m a i l.com> and we will give you access to the database and we can coach you during one hour to do so.
If you want to build your own academic search engine, we give detailled instructions to do so in our blog. We give detailled instruction to reproduce this project from scratch in order to reproduce the database, the pipeline enabling to index the titles, the abstracts and to generate the questions.
We have constructed other cool tools.
For example, this tool enables to explore the taxonomy of field of studies in a more visual way than on the defunt microsoft academic.
more comming soon...
Progress is made by trial and failure; the failures are generally a hundred times more numerous than the successes ; yet they are usually left unchronicled. WILLIAM RAMSAY
A fact acquires its true and full value only through the idea which is developed from it. JUSTUS VON LIEBIG