Nic Lane

I am an Associate Professor in the Computer Science department at the University of Oxford, where I am part of the Cyber-Physical Systems cluster of faculty.

My research interests revolve around the systems and modelling challenges that arise when computers collect and reason about people-centric sensor data. For the past three years, I have been primarily focused on: (1) developing deep learning models of human behavior and context (using audio, image and inertial data), and (2) enabling state-of-the-art signal processing and modeling algorithms to efficiently execute on embedded- and mobile-class hardware.

Previously, I held dual academic and industrial appointments as a Senior Lecturer (Associate Professor) in the Computer Science department at University College London (UCL), and a Principal Scientist at Nokia Bell Labs. At UCL I was part of the Digital Health Institute and UCL Interaction Center, while at the Bell Labs I led DeepX -- an embedded focused deep learning unit at the Cambridge location.

Before moving to England, I spent four years at Microsoft Research based in Beijing. There I was a Lead Researcher within the Mobile and Sensing Systems group (MASS). In March 2011, I received a Ph.D. from Dartmouth College.

Please visit my publications page or google scholar profile to learn more about my work.

I can be reached at: nicholas dot lane at cs dot ox dot ac dot uk

July '16	Our paper 'Sparsifying Deep Learning Layers for Constrained Resource Inference on Wearables' has been (conditionally) accepted to SenSys 2016.


June '16	Our paper 'LEO: Scheduling Sensor Inference Algorithms across Heterogeneous Mobile Processors and Network Resources' has been accepted to MobiCom 2016; a collaboration with the University of Cambridge and Samsung Research.


	Our paper 'Engagement-Aware Computing: Modelling User Engagement with Mobile Contexts' has been accepted at UbiComp 2016!


April '16	Our work on deep activity models for smartwatches wins best paper at WristSense 2016!


Mar '16	BodyScan is provisionally accepted at MobiSys! This is part of our new initiative into forms of radio-based sensing for wearables. A sister system (HeadScan) will appear at IPSN in April; pre-print of the HeadScan paper is available here.


	Honored to be serving as the PC Chair of HotMobile 2017. I will be working with Elizabeth Belding who will be the General Chair. Stay tuned for more details.


Feb '16	Pre-prints of two of our recent papers on enabling deep learning for mobile and wearables are now available. "DeepX: A Software Accelerator for Low-Power Deep Learning Inference on Mobile Devices" will appear at IPSN 2016; while "From Smart to Deep: Robust Activity Recognition on Smartwatches using Deep Learning" is to be presented at WristSense 2016.


Jan '16	Two papers provisionally accepted to IPSN '16. The first details unobtrusive monitoring of internal body states (such as, eating and drinking) using a wearable that exploits PHY-level radio transmissions; a first-of-its kind device developed with collaborators at MSU. The second presents DeepX -- a software-based accelerator for deep learning models that enables wearable-class hardware to cope with the deepest forms of this important direction in machine learning.


Oct '15	Initial results of our measurement study examining the system resource overhead of deep learning inference phases on wearables, phones and embedded devices will appear at the IoT-App workshop at SenSys.


Sept '15	DeepEar wins best paper at UbiComp '15! Congrats to all my co-authors.