Mission

We develop knowledge-based technologies that allow humans and computers to deal better with the artificial world that surrounds us.

News

LIA researchers win awards in IARPA competition

A team consisting of EPFL researchers Naman Goel, Diego Antognini and Boi Faltings, together with 2 EPFL students, won several awards in the Geopolitical Forecasting Challenge organized by the US IARPA.

https://actu.epfl.ch/news/phd-students-win-the-geopolitical-forecasting-chal/


LIA has three papers at the IAAI/AAAI 2019 conferences, the biggest AI event in history:

Stephane Martin, Boi Faltings and Vincent Schickel:
Context-Tree Recommendation vs Matrix-Factorization: Algorithm Selection and Live Users Evaluation

Naman Goel and Boi Faltings:
Deep Bayesian Trust : A Dominant and Fair Incentive Mechanism for Crowd

Simina Branzei and Aris Filos-Ratsikas:
Walrasian Dynamics in Multi-unit Markets

The book “Game theory for data science: eliciting truthful information” is now appearing at Morgan Claypool publishers.

 

Recent results

The techniques for session-based recommendation using Bayesian Variable-order Markov Models, developed at LIA, are integrated into the recommendation system of Prediggo SA and often more than double sales.

See forthcoming paper

 

LIA co-organized events