Course details


Advanced methods for multi-agent communication

S + P
SS 2023 Dr. Elia Bruni OFFLINE
B.Sc modules:
CS-BWP-AI - Artificial Intelligence
CS-BWP-CL - (Computational) Linguistics
M.Sc modules:
CS-MWP-AI - Artificial Intelligence
CS-MWP-CL - (Computational) Linguistics

CS-BW - Bachelor elective course
CS-MW - Master elective course

Everyday interactions require a common understanding of language, i.e. for people to communicate effectively, words (for example ‘cat’) should invoke similar beliefs over physical concepts (what cats look like, the sounds they make, how they behave, what their skin feels like etc.). However, how this ‘common understanding’ emerges is still unclear. One appealing hypothesis is that language is tied to how we interact with the environment. As a result, meaning emerges by ‘grounding’ language in modalities in our environment (images, sounds, actions, etc.). This course will review recent works in machine learning which bridges visual and natural language understanding through visually-grounded language learning tasks. In particular, we will look into multi-agent communication games. The course will be split into half frontal teaching, half hands-on practical. In the second part, students will form group and will implement a multi-agent communication game, supervised by the instructors. As the grounding problem requires an interdisciplinary attitude, this course aims to gather students with broad expertise in various fields -- machine learning, computer vision, natural language processing -- and who are excited about this space of grounding and interactions.