Video games Industry generated 150$ billion (approx. two times Facebook revenue) and involved one-third of the world population, in 2019 only. It is not hard to imagine how this attracted cyber-criminals, e.g.: 77 million PlayStation Network accounts were compromised in 2011; in 2015 Steam reported more than 70 thousand victims of scam monthly; cyberbullism events are also frequently reported. Being able to recognize gamers leveraging their gaming data could help to mitigate these issues, e.g., harmful players that are banned could be found again in all the other profiles they own. On the other side, this capability could be a further tool in the hand of cyber-criminals. In this paper, we are the first to demonstrate that players can be recognized based on their play-style. In particular, we observe the play-style through gaming data and use a Deep Neural Network for recognition. Our solution addresses games in which players control a character, and generic features are used to make our system possibly applicable to other games as well. To demonstrate the feasibility of our proposal, we run a thorough set of experiments based on players of Dota 2, which counts more than 10 million monthly active users. Our results show the efficiency and feasibility of the proposal, achieving 96% accuracy with only two minutes of gaming data.

PvP: Profiling Versus Player! Exploiting Gaming Data for Player Recognition

Conti M.
Supervision
;
Tricomi P. P.
2020

Abstract

Video games Industry generated 150$ billion (approx. two times Facebook revenue) and involved one-third of the world population, in 2019 only. It is not hard to imagine how this attracted cyber-criminals, e.g.: 77 million PlayStation Network accounts were compromised in 2011; in 2015 Steam reported more than 70 thousand victims of scam monthly; cyberbullism events are also frequently reported. Being able to recognize gamers leveraging their gaming data could help to mitigate these issues, e.g., harmful players that are banned could be found again in all the other profiles they own. On the other side, this capability could be a further tool in the hand of cyber-criminals. In this paper, we are the first to demonstrate that players can be recognized based on their play-style. In particular, we observe the play-style through gaming data and use a Deep Neural Network for recognition. Our solution addresses games in which players control a character, and generic features are used to make our system possibly applicable to other games as well. To demonstrate the feasibility of our proposal, we run a thorough set of experiments based on players of Dota 2, which counts more than 10 million monthly active users. Our results show the efficiency and feasibility of the proposal, achieving 96% accuracy with only two minutes of gaming data.
2020
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
978-3-030-62973-1
978-3-030-62974-8
File in questo prodotto:
Non ci sono file associati a questo prodotto.
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11577/3366875
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 4
  • ???jsp.display-item.citation.isi??? ND
  • OpenAlex ND
social impact