معرفی
Antonio Norelli is a Research Fellow at Oxford University and a PhD student in Computer Science at Sapienza University of Rome, specializing in AI and deep learning within the GLADIA research group. His academic journey includes a BSc in Physics, an MSc in Computer Science, and the SSAS interdisciplinary honors program. He has held research internships at Spiketrap (San Francisco) and Amazon Lablet in Tübingen. His research focuses on bridging the gap between human and artificial intelligence, emphasizing symbolic reasoning, explainable AI, and multimodal learning. Notable contributions include the ASIF framework for unsupervised multimodal alignment and Olivaw, an Othello-playing AI that achieves expert-level performance without human knowledge.
His key research interests span artificial scientific discovery, neural network explainability, and the theoretical foundations of intelligence. Recent work explores relative representation learning and zero-shot latent space communication, with applications in cross-modal retrieval and efficient model adaptation. Norelli’s interdisciplinary approach combines insights from philosophy of science (Popperian critical rationalism) with machine learning to develop systems capable of symbolic reasoning and hypothesis generation.
Publications highlight advancements in: (1) Multimodal learning without training through coupled data alignment, (2) Game AI strategies for Othello using minimal resources, (3) Neuro-symbolic architectures for interpretable decision-making. His work has been presented at NeurIPS and IEEE Transactions on Games, among others. Current efforts aim to formalize intelligence as information-processing mechanisms grounded in symbol manipulation and semantic alignment.
