The annual workshop of PC3 MATCHING will take place
🗓️ Thursday, October 9 and Friday, October 10
📍Sorbonne Université, Campus Pierre et Marie Curie, 4 place Jussieu, Paris 5e

He will present the progress of the project and welcome two speakers, who are introduced below.

Program

Thursday, October 9

9h30 – 10h	Welcome coffee
10h – 10h30	Opening
10h30 – 12h00	Presentation of progress on funded projects, session 1
12h00 – 13h30	Lunch
13h30 – 14h30	Keynote: The sense of agency in digital collaboration settings Ouriel Grynszpan, professor at Université Paris-Saclay, LISN member
14h30 – 15h30	Presentation of progress on funded projects, session 2
15h30 – 16h00	Break
16h00 – 16h40	Presentation of progress on funded projects, session 3
17h00 – 18h30	Collaborative work
18h30 – 20h00	Cocktail

Friday, October 10

8h30 – 9h	Welcome coffee
9h – 10h30	Presentation of progress on funded projects, session 4
10h30 – 11h	Collaborative work feedback
11h – 11h30	Break
11h30 – 12h30	Keynote: Physics-based simulation of Multi-Agent interactions using deep imitation learning Frank Multon, Senior researcher Inria, IRISA member
12h30 – 12h45	Conclusion
12h45 – 14h	Lunch

speakers

Sense of agency in digital collaboration settings:
The sense of agency can be defined as the experience of controlling one’s actions and their effects in the environment. It is considered critical to social interactions and forms the basis of the feeling of responsibility and therefore liability. Experimental evidence shows that disambiguating who is responsible for an action is not straightforward when the action is performed jointly with another person. This presentation will describe several experiments showing how joint action modifies the experience of agency. Changes occur at the explicit level of awareness but also at the implicit level of sensorimotor processes. Moreover, there is emerging evidence that the sense of agency is altered when interacting with artificial agents compared to human-human interaction.

Physics-based simulation of Multi-Agent interactions using deep imitation learning:
Simulating realistic interaction and motions for physics-based characters is of great interest for interactive applications, and automatic secondary character animation in the movie and video game industry. We propose a novel Multi-Agent Generative Adversarial Imitation Learning based approach that generalizes the idea of motion imitation for one character to deal with both the interaction and the motions of the multiple physics-based characters. Based on unstructured motion capture datasets, the system trains control policies allowing each character to imitate the interactive skills associated with each actor, while preserving the intrinsic style. This approach has been tested on two different fighting styles, boxing and full-body martial art, to demonstrate the ability of the method to imitate different styles.