Alessandro Ronca

Alessandro Ronca

Research Associate

Research Associate, University College

Associate Member, Oriel College

E: first.last@cs.ox.ac.uk

Wolfson Building, Parks Road, Oxford OX1 3QD

Themes:

• Artificial Intelligence and Machine Learning
• Data, Knowledge and Action

Interests

Areas: Artificial Intelligence, Logic, Automata Theory, Algebraic Automata Theory, Complexity Theory, Reinforcement Learning, Learning Theory, Automata Learning, Knowledge Representation and Reasoning, Temporal Logics, Datalog, Query Languages.

My current research focuses on three aspects of Artificial Intelligence.

• Machine learning models to capture temporal patterns, including Recurrent Neural Networks and Transformers. I study them from formal point of view, assessing their ability to capture temporal patterns.
• Logics to capture temporal patterns. I am particularly interested in the Transformation Logics, a new family of temporal logics that I have recently established and that allows for creating hierarchies of increasing expressivity and complexity, with a great potential to match the expressivity-complexity trade-off required by specific applications.
• Reinforcement learning in domains where an agent must learn to capture patterns over the history of past events. This work has the potential to greatly extend the number of applications where reinforcement learning can be employed.

Selected Publications

View AllManage publications

• On the Expressivity of Recurrent Neural Cascades

  Nadezda A. Knorozova and Alessandro Ronca

  In Proceedings of the AAAI Conference on Artificial Intelligence (AAAI). 2024.

  Details about On the Expressivity of Recurrent Neural Cascades | BibTeX data for On the Expressivity of Recurrent Neural Cascades | Link to On the Expressivity of Recurrent Neural Cascades

• The Transformation Logics

  Alessandro Ronca

  In Proceedings of the Thirty−Third International Joint Conference on Artificial Intelligence (IJCAI). 2024.

  Details about The Transformation Logics | BibTeX data for The Transformation Logics | Link to The Transformation Logics

• Provably Efficient Offline Reinforcement Learning in Regular Decision Processes

  Roberto Cipollone‚ Anders Jonsson‚ Alessandro Ronca and Mohammad Sadegh Talebi

  In Advances in Neural Information Processing Systems 36 (NeurIPS). 2023.

  Details about Provably Efficient Offline Reinforcement Learning in Regular Decision Processes | BibTeX data for Provably Efficient Offline Reinforcement Learning in Regular Decision Processes | Link to Provably Efficient Offline Reinforcement Learning in Regular Decision Processes