International Conference on Applied Mathematics, Statistics, and Computing (ICAMSAC 2023), Udayana University, 21st ? 22nd November 2023
Sprache des Tagungstitel:
Englisch
Original Kurzfassung:
Whether you believe in Artificial General Intelligence (AGI) where machines
will stand in tandem with humans in their ability to reason, learn, solve
problems, and make decisions or not, there will be a day not far in the future
where machines need to cooperate, collaborate and even compete with
other machines and humans to achieve the goals assigned to them. This will
be the time when the interdependencies of activities of humans and
machines need to be managed efficiently for which frameworks/tools are
required. We call this framework "Cooperative Artificial Intelligence" or CAI
for short. CAI is the study of the process of machines and humans working
together toward a goal or goals using artificial intelligence approaches and
tools.
Advances in machine learning especially in the field of reinforcement
learning focus on a single AI entity that acts in an environment where only
one agent is doing the sensing, planning and decision making which fails to
capture many real world scenarios in which an agent needs to make
decisions in an environments that contains multiple actors/agents.
In recent years, researchers approached the problem of cooperative
artificial intelligence from two perspectives one is multi-agent
reinforcement learning in which a zero-sum games are assumed most of the
time and the second is the deployment of AI tools to facilitate human
cooperation such as language translation, HCI, social networks, etc., in which
the focus is on solving the problems of cooperation.
The aim of this talk is to establish a solid theoretical foundation for
Cooperative Artificial Intelligence and to elaborate on the machine learning
concepts, tools and practices to promote human-AI cooperation.
Sprache der Kurzfassung:
Englisch
Vortragstyp:
Hauptvortrag / Eingeladener Vortrag auf einer Tagung