Nicola Cancedda

RESEARCH SCIENTIST AND MANAGER | LONDON, UNITED KINGDOM

Nicola is a research scientist and manager with Meta's Fundamental AI Research (FAIR) organisation. His current research focus is on better understanding how Large Language Models realize complex behaviors to make them safer and more efficient. Nicola is an alumnus of the University of Rome "La Sapienza", and has held applied and fundamental research and management positions at Meta, Xerox, and Microsoft, pushing the state of the art in tasks in Machine Learning, Machine Translation, and Natural Language Processing, and leading the transfer of research results to large-scale production environment.

Nicola's Publications

August 10, 2024

NLP

Spectral Filters, Dark Signals, and Attention Sinks

Nicola Cancedda

August 10, 2024

June 13, 2024

NLP

Know When To Stop: A Study of Semantic Drift in Text Generation

Ava Spataru, Eric Hambro, Lena Voita, Nicola Cancedda

June 13, 2024

February 22, 2024

NLP

Toolformer: Language Models Can Teach Themselves to Use Tools

Timo Schick, Jane Yu, Roberto Dessì, Roberta Raileanu, Maria Lomeli, Eric Hambro, Luke Zettlemoyer, Nicola Cancedda, Thomas Scialom

February 22, 2024

December 06, 2023

NLP

Polar Ducks and Where to Find Them: Enhancing Entity Linking with Duck Typing and Polar Box Embeddings

Mattia Atzeni, Mike Plekhanov, Frederic Dreyer, Nora Kassner, Simone Merello, Louis Martin, Nicola Cancedda

December 06, 2023

June 15, 2023

NLP

Multilingual End to End Entity Linking

Mike Plekhanov, Nora Kassner, Kashyap Popat, Louis Martin, Simone Merello, Borislav Kozlovskii, Frederic Dreyer, Nicola Cancedda

June 15, 2023

May 25, 2022

NLP

EDIN: An End-to-end Benchmark and Pipeline for Unknown Entity Discovery and Indexing

Nora Kassner, Fabio Petroni, Mike Plekhanov, Sebastian Riedel, Nicola Cancedda

May 25, 2022