Andrea Graziano

About me

Hello! I'm a student enrolled in single courses (from Informatics to Physics to Neurophysiology) at a few universities in Italy. I obtained my Bachelor's at Bocconi University in July 2023. I'm currently a research intern at the Institute for Neural Computation, Ruhr-Universität Bochum.

During my degree, I co-founded BAINSA (Bocconi AI & Neuroscience Student Association), a student group conducting student-level research projects, organizing events, weekly meetings, and other activities. I am still involved in some of its activities, i.e., I organzied a panel on NeuroAI at WAICF (World AI Cannes Festival) 2024.

Research

My research interest lies at the intersection of Theoretical Neuroscience and Machine Learning, in exploring the neural principles of cognitive processing allowing for adaptive behavior and flexible learning. I aim to build biologically-plausible computational models as I believe this is crucial for enhancing the utility of future artificial intelligence as well as ensuring a responsible development of next-generation AI systems.

My current research at the Institute for Neural Computation, Ruhr-Universität Bochum focuses on Clone-structured-causal-graphs (CSCGs) [1],[2]. Under the supervision of Tristan Stöber, my goal is to provide a better understanding of how CSCGs support causality by validating the cloned Hidden Markov Model (HMM) used for building CSCGs against counterfactual causal criteria. This is useful to demonstrate that the model can serve as a computational implementation of the episodic memory theory developed by M. Werning, proposing that episodic memory plays a crucial role in causal reasoning through counterfactuals.

In my previous work, and as a bachelor's thesis, under the supervision of Alessandro Sanzeni, I explored how visual information is processed in primary visual cortex (V1) during body movement. Through electrophysiological data analysis, I discovered that V1 contains an invariant representation of visual stimuli that is independent of passive head movements, and obtained evidence suggesting the presence of an abstract, yet high- dimensional representation of information that supports generalization.

Thesis

• Understanding visual information processing during body movement.
  (The thesis is password-protected: You can request the password via email if interested).

More

Here you can find a few blog articles I've written in the past years:

• - “You don’t conquer anything except things in yourself”. Flow: A book review
• - What is Data Science?

I enjoy reading, doing sports, and listening to house music.