Training a gaming agent on brainwaves

Bartolomé, Francisco; Moreno, Juan; Navas, Natalia; Vitali, José; Ramele, Rodrigo; Santos, Juan Miguel

Artículo de Publicación Periódica:
Training a gaming agent on brainwaves

dc.contributor.author	Bartolomé, Francisco
dc.contributor.author	Moreno, Juan
dc.contributor.author	Navas, Natalia
dc.contributor.author	Vitali, José
dc.contributor.author	Ramele, Rodrigo
dc.contributor.author	Santos, Juan Miguel
dc.date.accessioned	2022-06-27T15:19:39Z
dc.date.available	2022-06-27T15:19:39Z
dc.date.issued	2020-12-07
dc.description.abstract	"Error-related potential (ErrP) are a particular type of Event-Related Potential (ERP) elicited by a person attending a recognizable error. These Electroencephalographic (EEG) signals can be used to train a gaming agent by a Reinforcement Learning (RL) algorithm to learn an optimal policy. The experimental process consists of an observational human critic (OHC) observing a simple game scenario while their brain signals are captured. The game consists of a grid, where a blue spot has to reach a desired target in the fewest amount of steps. Results show that there is an effective transfer of information and that the agent successfully learns to solve the game efficiently, from the initial 97 steps on average required to reach the target to the optimal number of 8 steps. Our results are expressed in threefold: (i) the mechanics of a simple grid-based game that can elicit the ErrP signal component, (ii) the verification that the reward function only penalizes wrong steps, which means that type II error (not properly identifying a wrong movement) does not affect significantly the agent learning process; (iii) collaborative rewards from multiple observational human critics can be used to train the algorithm effectively and can compensate low classification accuracies and a limited scope of transfer learning schemes."	en
dc.identifier.uri	http://ri.itba.edu.ar/handle/123456789/3918
dc.language.iso	en	en
dc.relation	nfo:eu-repo/semantics/altIdentifier/doi/10.1109/TG.2020.3042900
dc.relation	info:eu-repo/grantAgreement/ITBACyT/2020-15/AR. Ciudad Autónoma de Buenos Aires
dc.subject	CEREBRO	es
dc.subject	JUEGOS	es
dc.subject	ALGORITMOS	es
dc.subject	APRENDIJZAJE	es
dc.title	Training a gaming agent on brainwaves	en
dc.type	Artículos de Publicaciones Periódicas	es
dc.type	info:eu-repo/semantics/publishedVersion
dspace.entity.type	Artículo de Publicación Periódica
itba.description.filiation	Fil: Bartolomé, Francisco. Instituto Tecnológico de Buenos Aires; Argentina.
itba.description.filiation	Fil: Moreno, Juan. Instituto Tecnológico de Buenos Aires; Argentina.
itba.description.filiation	Fil: Navas, Natalia. Instituto Tecnológico de Buenos Aires; Argentina.
itba.description.filiation	Fil: Vitali, José. Instituto Tecnológico de Buenos Aires; Argentina.
itba.description.filiation	Fil: Ramele, Rodrigo. Instituto Tecnológico de Buenos Aires; Argentina.
itba.description.filiation	Fil: Santos, Juan Miguel. Instituto Tecnológico de Buenos Aires; Argentina.