Open NLP meetup: Landscape of Vector Databases & Overcoming DPR's Input Limit

It's time for our next event on January 20th, 7pm CEST with two talks on "The Landscape of Vector Databases in NLP" and "Overcoming DPR's Input Length Limitation with PARM"! The link to the livestream of the virtual event will be shared here just before the event.

Talk 1 (7:00pm):
Dmitry Kan, Principal AI Scientist at Silo AI and host of the Vector Podcast [1], will give an overview of the landscape of vector search databases and their role in NLP, along with the latest news and his view on the future of vector search. Further, he will share how he and his team participated in the Billion-Scale Approximate Nearest Neighbor Challenge and improved recall by 12% over a baseline FAISS.

Talk 2 (7:35pm):
Sophia Althammer, PhD student at TU Vienna, will present "PARM: A Paragraph Aggregation Retrieval Model" [2], which got accepted as a research paper at ECIR 2022 [3]. PARM overcomes the input length limitation of dense passage retrieval models. Thereby, it enables dense document-to-document retrieval with documents of arbitrary length. Experiments on two legal case retrieval collections demonstrate its increased retrieval effectiveness.

Virtual Networking (8:05pm):
We will make sure that there is enough time and virtual space to chat with other attendees after the two talks.

[1] https://www.youtube.com/channel/UCCIMPfR7TXyDvlDRXjVhP1g
[2] https://github.com/sophiaalthammer/parm
[3] https://arxiv.org/abs/2201.01614