I am co-organising SOAR, a tutorial on SOlving the sensor-based Activity Recognition problem at UbiComp tomorrow! Come join us if you're attending UbiComp/ISWC 2023. #UbiComp2023 #ISWC2023 - more details at sites.google.com/view/soar-tuto…

I attend Data-centric ML @ #ICLR2024 to present a work (in progress) with Ryuhaerang Choi @CHI2025 Soumyajit Chatterjee Dimitris Spathis Sung-Ju Lee Fahim Kawsar Tell me what you do & I tell you about Device Software Research Bell Labs: projects, collaborations, interns, jobs,... bell-labs.com/research-innov…

🤔Why do Large Language Models like ChatGPT struggle with temporal data? Our new work published in JAMIA w/ Fahim Kawsar explores LLMs' difficulties when processing time series data, particularly from sources like mobile sensors or medical records.

(Fundraising for Shelter) I'm running a half marathon to support Shelter's mission in securing safe homes for everyone. Please help someone feel a little safer! I will personally add £1 for every £10 donated. justgiving.com/page/running-a…

Working on Federated Learning and Audio Processing? Our workshop on federated learning for audio understanding at IEEE ICASSP 2025 provides an excellent opportunity to showcase your work. Please consider submitting!! Link: icasspflute.github.io #Conferences #ICASSP2025

🚀 The submission deadline for the #HumanSys24 Workshop (in conjunction with ACM SenSys24 ACM SenSys ACM SenSys) has been extended to September 25th, 2024, 23:59 AOE! #SenSys24 #ACM #DeadlineExtension Details: humansys-workshop.github.io

🚨 Submission Deadline Approaching! 🚨 Don't miss your chance to submit to the FLute Workshop on Federated Learning for Audio Understanding at ICASSP 2025! 🗓️ Deadline: Nov 1, 2024 Submit here: cmt3.research.microsoft.com/ICASSP2025/Tra… #AI #FederatedLearning #Audio #ICASSP2025

Epicurus, 23 centuries ago, believed "All sensations are true, and errors come from our interpretations". Sensation has always been fascinating! If you have experience and passion for human sensing and interpreting sensory data, I'd love to hear from you. We have opportunities!

We are accepting submissions for FLute@ICASSP 2025. Please consider submitting and get feedback on your ideas surrounding Federated Learning for Audio. You can also submit 2-page extended abstracts presenting your idea in brief. Submission Link: lnkd.in/eDpgVsQp

We are delighted to share PaPaGei, an open foundation model for PPG trained on over 57,000 hours of publicly available data. The PPG signal, which measures blood volume changes, is essential for various health tasks.

Self-supervised learning on different PPG datasets is a challenge that we have been excited to work on. See our recent work with Arvind Pillai Dimitris Spathis & Fahim Kawsar

🦜PaPaGei: an open foundation models for the most widely used biosignal🦜 It is pre-trained on over 57,000 hours of PPG signals (like those from smartwatches & pulse oximeters) from publicly available sources. Paper: arxiv.org/abs/2410.20542 Models/code: github.com/nokia-bell-lab…

Collecting new datasets for ML is costly, but what if we could reuse existing ones? Our work, SoundCollage, discovers hidden classes in audio datasets—automatically creating new label spaces for fresh applications. 🧵

Research Internship — Nokia Bell Labs, Cambridge, UK! Three topics for summer 2025: (1) Wearable Intelligence for MedTech (2) Sensory Wearables for MedTech (3) Wearable Systems for MedTech Apply:👇 linkedin.com/posts/fahimkaw…

We are accepting applications for summer internship at Nokia Bell Labs Cambridge, UK. If you are enrolled in a PhD program and interested in wearable intelligence please drop me an email with your updated CV and research interest.

Great news from Vancouver, happy to share that 🦜PaPaGei received the 🏆 Best Paper Award at the Time Series in the Age of Large Models (TSALM) workshop in #NeurIPS2024! Huge congratulations to Arvind Pillai and the team. Stay tuned for the v2 of the paper with more results.