The event, organised under the EU COST Action EPIGRAPHY, with contributions from INESC-ID researchers Nuno Roma and Pedro Tomás (INESC-ID / DEEC), will consist of a colloquium followed by a panel discussion. The first part will feature keynote speaker Thea Aarrestad (ETH Zurich), with the talk “When the Smallest AI Meets the Biggest Data: Nanosecond Discovery at the CERN LHC & the Future of Fast ML”. The panel discussion will focus on “The Future of Fast Machine Learning” and will include INESC-ID researcher Arlindo Oliveira (INESC-ID / IST) and Mário Figueiredo (IT / IST) as panellists.

Date & Time: 13 January, 16h30
Location: Anfiteatro Abreu Faro, Instituto Superior Técnico

Abstract:
In particle physics, the Standard Model—our most accurate and predictive theory—remains only an effective description that breaks down at high energies. The search for new physics beyond it is driven by the CERN Large Hadron Collider (LHC), which produces an extraordinary 10,000 exabytes of raw data each year from high-energy proton collisions.

As the LHC enters its High-Luminosity era (HL-LHC), these underground systems—located 100 meters underground—must manage data rates approaching 5% of global internet traffic while dealing with increasing event complexity. Achieving the required data fidelity for new physics searches demands AI algorithms that deliver extreme throughput and ultra-low-latency inference.

In this talk, I will show how tiny, highly optimized AI models are redefining what is possible in real-time data processing—and how they are transforming the search for Beyond Standard Model phenomena.

Panel Discussion: The Future of “Fast Machine Learning”

Follow-up debate focused on future perspectives for the development of dedicate processors for fast machine learning.

Panelists:

• Thea Aarrestad (ETH Zurich) – Specialist in fast ML for the CMS detector at CERN

• Arlindo Oliveira (IST / INESC-ID) – Expert in Algorithms, AI, and Computer Architecture

• Mário Figueiredo (IST / IT) – Expert in Machine Learning and Signal Processing

About the Speakers

Thea Aarrestad: Researcher at ETH Zurich and a leading figure in developing low-latency ML algorithms applied to the trigger systems of the CMS detector at CERN.

Arlindo Oliveira: Distinguished Professor at IST and Researcher at INESC-ID. An expert in Artificial Intelligence, Algorithms, and Computer Architecture, and author of several influential books, including The Digital Mind.

Mário Figueiredo: Distinguished Professor at IST and Researcher at the Instituto de Telecomunicações (IT). One of the most highly cited Portuguese researchers worldwide in the fields of Machine Learning and Signal Processing.