Gaia Belardinelli

Visiting Scholar

Departement of Economics

University of California, Davis

I am currently a visiting scholar at the University of California at Davis, working with Burkhard C. Schipper on formal models of implicit cognition.  Starting from the next academic year, I will be a postdoc at Stanford University working on the impact of implicit biases on reasoning.


I completed my PhD studies at the University of Copenhagen. My PhD project was about formal models of explicit and implicit cognitive phenomena from an epistemic and multi-agent perspective, using modal logic or other formal methods to represent them. My thesis focuses on three interrelated themes: attention, awareness, and implicit reasoning. 


In general, I am interested in exploring applications of logical models to cognitive science, AI, and economics (especially their intersection), and I like to challenge the strong idealizations that are built in our logic models. Finally, I am passionate about methodological questions such as "why are mathematical models so illuminating about reality, when they abstract away so many details?".

News:

The funding agency is the Independent Research Fund Denmark. You can find a link to the project here (although there is not much information about it yet). I am extremely grateful for this and looking forward to start!

Many thanks to Thomas Bolander, Nina Gierasimczuk, Burkhard Schipper, and Thomas Icard for supporting me in this, either as collaborators or host. 

The talk is entitled Attention in Epistemic Logic.

Alessandro Burigana will also be there! We will work on syntactic event models together. In particular, we will try to find a syntactic way to induce updates in Dynamic Epistemic Logic.

The seminar is held by Snow Xueyin Zhang. My presentation will be about epistemic logic models of unawareness.

The talk is entitled Attention in Dynamic Epistemic Logic.

One week summer school on philosophical themes in artificial intelligence.

Keynote speakers:  Johanna Thoma, Joseph Halpern, and Marija Slavkovik.

The paper is entitled Implicit Knowledge in Unawareness Structures. 

It introduces a notion of implicit knowledge in Heifetz, Meier, Schipper (HMS) models for unawareness (see their 2006, 2008 papers). This is a lattice structure for unawareness that originally features explicit knowledge only. As HMS models have been applied to decision and game theory, we see this as a first step to modeling the behavioral implications of implicit cognition. 

The paper also introduces a category of Fagin Halpern (FH) models for unawareness (Fagin and Halpern, 1987). FH models originally provide an objective perspective on the epistemic situation, modeling it as seen from the outside, from a fully aware point of view. We show that models in an FH category can also provide each agent's subjective perspective, where the modeled situation is described using their subjective language only.

HMS models with implicit knowledge are shown modally equivalent to the category of FH models. By this result and the soundness and completeness a logic for explicit and implicit knowledge and awareness with respect to FH models, we get soundness and completeness of HMS models with implicit knowledge with respect to that logic. 

Yehee!

The talk is entitled A dynamic epistemic logic of attention with attention change and limits on attentional capacities. 

See here for an abstract.

The talk is entitled Attention! Dynamic Epistemic Logic models of (in)attentive agents. 

See here for an abstract.

The paper is entitled Attention! Dynamic Epistemic Logic Models of (In)attentive Agents. It presents two logical models of partial learning of events due to inattention to some aspects of them. In one model, the agents remain indifferent with respect to information they did notThe event models are complex and grow exponentially large in the number of agents and announced literals, so we also present a more succinct way to describe them via a fully syntactic representation.

See here for an abstract.

The workshop is organized by Burkhard C. Schipper and myself.

One week summer school about the use of logic in artificial intelligence.

Courses Followed: Automated Theorem Provers, by Joseph Urban; History of AI, by Stephanie Dick; Logics for Categories, by Alessandra Palmigiano; Multi-Agent Systems, by Michael Wooldridge.

Thomas Bolander and I organized this three-day workshop to explore applications of logic and multi-agent systems to the social sciences. We had standard conference-talks for more focused in presentations, and group discussion to investigate three themes: logic as a foundation to the social sciences; logic as a methodology in the social sciences; logic and agent-based models. The participants are now writing a paper together to report our discussions. 

Thanks to CIBS and the Carlsberg Foundation for funding it!

Courses Followed: Knowledge and Gossip, by Hans van Ditmarsch & Malvin Gattinger, Logic & Probability, by Thomas Icard & Krzysztof Mierzewski, Explainability in Integrated Cognitive Systems Combining Logic-based Reasoning and Data-driven Learning, by Mohan Sridharan. 

One week attendance.

Working on our joint project on cognitively plausible models of attention.  We are now focusing on modeling inattentional blindness, the phenomenon according to which agents completely miss certain events, even if they happen at fixation. 

See here for an abstract.

See here for an abstract.

Working on modeling implicit knowledge in unawareness structures, a lattice model of unawareness that Burkhard and coauthors created during Burkhard's PhD. We are planning to the use the resulting model to represent tacit knowledge, implicit biases, and other behaviorally relevant aspects of implicit cognition - exciting!

With Sonja I worked on logics for social network. We initiated a project with Lei Li and Anthia Solaki on cognitive (attention-based) and algorithmic filters in social networks. The basic idea is that not all the announcements posted in social media are received by the users, firstly because the algorithm filters out what is deemed not interesting to the user, and secondly because the user will only attend some of the announcements she receives.

With Johan I worked on the use of syntax in modeling cognitive and social phenomena: What does it add to the picture? In particular with respect to economics models of the same phenomena. Moreover, we explored lattice based formalizations of attention dynamics.

We initiated a project on modeling cognitive aspects of attention, as top-down vs bottom-up attention, inattentional blindness phenomena and other surprising aspects of attention. We are very excited to see that there are many possible applications, spanning from epistemic planning to attention economics.

Courses Followed: Introduction to Deontic Logic and its Applications, by Réka Markovich & Leon van der Torre, Temporal Logics, by Valentin Goranko,  Workshop on Automated Synthesis, organized by Natasha Alechina and Brian Logan, Abstract Argumentation and Modal Logic, by Davide Grossi and Carlo Proietti.

Earlier events are listed in my CV. If you are interested in having a look, feel free to send me an email !