women@CL Talklet Event

Add to your list(s) Download to your calendar using vCal

• Sandra Servia, Catalina Cangea, Angeliki Koutsoukou-Argyraki
• Friday 25 May 2018, 13:00-14:00
• Computer Laboratory, William Gates Building, Room SS03.

If you have a question about this talk, please contact Ayat Fekry.

This talk has been canceled/deleted

Title: Mobile sensing at the service of mental wellbeing

Abstract

The pervasiveness of smartphones and their rich-set of built-in sensors, including accelerometer, GPS and microphone, have allowed the emergence of many platforms to passively monitor health and behaviour through experience sampling and sensing, at low cost and large scale. However, studies at the confluence of mental health and mobile sensing have been longitudinally limited, controlled, and confined to a small number of participants. In this talk, I will report on what we believe is the largest longitudinal, in-the-wild study of mood through smartphones, which includes data from ~18,000 participants for a period of three years. Using data collected with an Android app, which includes self-reported moods, system triggered experience sampling data and passive sensing data, we are able to identify routines and their relation with demographics, perceived health and psychological traits, as well as exploring the predictability of users’ mood from passive sensing data. Although this large scale data collection is very suitable for population studies, the collection and use of sensitive data comes with privacy issues and chances of data misuses. In this line, I will comment on the trade-offs between utility, battery consumption and latency of private-by-design mobile health apps that rely on on-device processing with limited cloud offloading.

———————————————————————————

Title: Cross-modal techniques for data integration

Abstract

Multimodal learning is a natural and necessary progression from the traditional methods that typically learn to represent a single modality. Applications exist in a variety of scenarios, a few notable examples being medicine, environmental risk and robotics. In this talk, I will present two classification tasks from the audiovisual and chemical domains, along with the deep learning architectures that I have designed to improve on existing methods.

—————————————————————————————-

Title: Proof Mining Mathematics, Formalizing Mathematics-the ALEXANDRIA project

Abstract

Proof mining is a research program in applied proof theory involving the extraction of quantitative, computable information from (even nonconstructive) mathematical proofs of statements of a certain logical form, via a pen-and-paper i.e. \textit{not} automated logical analysis. The program originated as ``unwinding of proofs’’ in the ideas of Georg Kreisel from the fifties, and has been developed by Ulrich Kohlenbach and his collaborators during the past two decades. A great deal of applications for proofs in different research directions in Mathematics has been achieved. ALEXANDRIA is a new ERC project at the University of Cambridge under the leadership of Lawrence Paulson aiming at the creation of a proof development environment for working mathematicians through a collaboration of mathematicians and computer scientists. This will be achieved by formalizing mathematical proofs with the proof assistant \textit{Isabelle}. The focus of the project is the management and use of large-scale mathematical knowledge, both as theorems and as algorithms. In addition to the obvious importance of proof verification for Mathematics and the usefulness of libraries of formalized proofs for (the future generations of) mathematicians, the formalization of mathematical proofs could possibly shed light on interesting proof theoretic questions. Moreover, enriching the libraries with formalized proof-mined proofs could open the way for the exciting prospect of automating proof mining itself.

This talk is part of the Women@CL Events series.

This talk is included in these lists:

This talk is not included in any other list

Note that ex-directory lists are not shown.