Fabio Petroni – Co-Founder & CTO @ Samaya AI (https://samaya.ai/)

Wednesday, November 30th, 2022 –  11.15 AM

Fabio is Co-Founder & CTO at Samaya AI, building an AI-powered knowledge-discovery platform. Before that he was a Researcher at FAIR and Thomson Reuters, focusing on representing, gathering, extracting, using, reasoning on and creating world knowledge using AI.

“Improving Wikipedia Verifiability with AI”

Verifiability is a core content policy of Wikipedia: claims that are likely to be challenged need to be backed by citations. There are millions of articles available online and thousands of new articles are released each month. For this reason, finding relevant sources is a difficult task: many claims do not have any references that support them. Furthermore, even existing citations might not support a given claim or become obsolete once the original source is updated or deleted. Hence, maintaining and improving the quality of Wikipedia references is an important challenge and there is a pressing need for better tools to assist humans in this effort. In the talk, I’ll present our recent research showing that the process of improving references can be tackled with the help of AI. Our results indicate that an AI-based system could be used, in tandem with humans, to improve the verifiability of Wikipedia. More generally, we hope that our work can be used to assist fact-checking efforts and increase the general trustworthiness of information online.

Social Profiles




Giuseppe Attardi – Full Professor at the University of Pisa – MISE AI Expert

Wednesday, November 30th, 2022 –  5.00 PM

Giuseppe Attardi is a professor of Computer Science at the University of Pisa. He previously worked at MIT’s AI Lab, Sony Paris Research Laboratory, ICSI Berkeley, and Yahoo Research Barcelona. He developed Omega, a description logic; CMM, the garbage collector used in Java; and DeSR, a dependency parser. He participated in developing Arianna, the first Italian search engine, and iStella. He is the founder or partner of several startups in Italy and Spain. He has been involved in implementing the fiber optic networks of the JRC, the University of Pisa, and GARR. He promoted a campaign to reduce Internet access costs to facilitate deployment. He was responsible for the implementation of GARR’s cloud platform. He contributed to drafting the Italian strategy on Artificial Intelligence and the birth of the first national doctorate in Artificial Intelligence.

Large Language Models are All You Need

LLMs are among the three scientific breakthroughs of Deep Learning applied to NLP in just ten years: word embeddings, attention, and prompt engineering. They have shown surprising effectiveness in all tasks and have also been adopted in other areas, including cross-cutting tasks such as generating images from text descriptions. However, LLMs have also raised perplexities, starting with doubts about the release of GPT-2 to Galactica’s recent retirement. Their capabilities, which seem to grow with their size, have become a subject of study. We will discuss whether discrimination will continue to grow among groups of researchers capable of building them and about possible alternatives.

Social Profiles