SCONE meeting 25-02-2022

Venue

This hybrid event will run on Zoom and in the School of Computing Science at the University of Glasgow

Sir Alwyn Williams Building, room 422

How to get here

Take the train to Glasgow Queen Street. On arrival, take the right-hand exit from the station (past the ticket office). On leaving the station you should see the entrance to Buchanan Street subway station immediately across the street. Take the subway to Hillhead. On leaving Hillhead station, turn right and walk along Byres Road to Great George Street. Turn right and walk up the hill past the Marks & Spencer, then turn right and walk along Lilybank Gardens. The Sir Alwyn Williams Building is at the far end of Lilybank Gardens, on the left.

Keynote

Speaker: Rui Li

Title: Reciprocity in Communications and ML Research: Is there free lunch for AI in comms? And can communications help the advancement of ML?

Abstract:
Applying powerful machine learning tools to solve communications and
networking problems has led to many success stories. In this talk, I will
discuss my previous work along this line, on the optimisation of utility for
millimetre wave backhaul networks using machine learning, the lessons
learned, and future opportunities. I will then talk about our effort towards
studying a family of machine learning algorithms, namely meta-learning,
using the classical channel coding tasks in communications. I will discuss
how mathematical tools known by information theorists can be used to
evaluate existing meta-learning algorithms, and thereby provide an elastic
benchmark platform for future study of meta-learning.

Bio:
Rui Li a machine learning researcher at Samsung AI Centre in Cambridge,
UK. She is interested in wireless communications, machine learning, and
the intersections of these. In the past, Rui has worked on resource allocation and utility optimisation in millimetre wave backhaul networks, and
acoustic signal processing problems. She is increasingly interested in
cross-domain adaptation and generalisation of ML models, as well as their
performances in real-world systems. Rui obtained her PhD from the School
of Informatics, University of Edinburgh in 2019, advised by Dr Paul Patras,
and she remains a close friend of the Mobile Intelligence Lab at Edinburgh.

Schedule

13:00-13:30 – Arrival, Welcome, Coffee

13:30-14:00 – Alec Diallo (U. Edinburgh) Adversarial ML

14:00-14:30 – Kaushal Kumar (Stirling U.) Improving the scalability of 6TiSCH networks using smart scheduling reservation solution

14:30-15:00 – Jose Cano and Ferheen Ayaz (U. Glasgow) Multimodal AI-based Security at the Edge

15:00-15:30 – Mihail Yanev (U. Glasgow) Internet video stability

15:30-16:00 – Coffee break

16:00-16:30 – Panel discussion on interdisciplinary research with Paul Patras, Lito Michala and Richard Mortier

16:30-17:00 – Rui Li (Samsung) Keynote talk