Hermine J. Grosinger

Hermine J.(asmin) Grosinger (PhD)

Researcher in AICognitive Robotic Systems (CRS), AASSÖrebro University, Sweden https://www.oru.se/personal/hermine_grosinger <firstname> [dot] <surname> [at] oru [dot] se

Research Interests (short)

Knowledge Representation and Reasoning, Logic and formal methods in Artificial Intelligence
Proactive Agents and Human-AI systems
Epistemic reasoning
Causal reasoning

Education

PhD in Computer Science (Örebro University, Sweden)
MSc in Computer Science (Technical University Vienna, Austria)

Research Interests (long)

Knowledge Representation and Reasoning (KRR), Logic and formal methods in Artificial Intelligence (AI)

KRR in AI provides powerful methods for creating human-understandable representations of the world. Formal methods and sound reasoning lead to verifiable outcomes. The focus of my research is using logic and formal approaches and/or developing logic with the aim to further the autonomy of artificial agents.

Proactive Agents and Human-AI systems

Proactive behavior is autonomously initiating action taking into account future state development as well as anticipating potential consequences of the agent’s actions on (the mind) of other agents and the environment. My investigation is on how to make artificial agents proactive and which joint cognitive abilities are interacting in producing proactive behavior of agents, as well as finding out how proactivity can enhance human-AI system collaborations in real world settings.

Epistemic reasoning

The ability to reason on own and other agents' knowledge and beliefs enhances the agent's proactive abilities. For example, by understanding the other agent's false beliefs/expectations and preferences, the first agent can proactively make relevant announcements to correct the second agent and prevent them from troubles. For this kind of work I have been using DEL (Dynamic Epistemic Logic).

Causal reasoning

Reasoning on a causal model with exogenous and endogenous variables together with their dependencies (structural functions) can inform an agent's proactive behavior. It can be intertwined with epistemic reasoning so as to model an agent's uncertainty in causal reasoning.

Page updated

Google Sites

Report abuse