mRisk: Continuous Risk Estimation for Smoking Lapse from Noisy Sensor Data with Incomplete and Positive-Only Labels.

Passive detection of risk factors (that may influence unhealthy or adverse behaviors) via wearable and mobile sensors has created new opportunities to improve the effectiveness of behavioral interventions. A key goal is to find opportune moments for intervention by passively detecting rising risk of an imminent adverse behavior. But, it has been difficult due to substantial noise in the data collected by sensors in the natural environment and a lack of reliable label assignment of low- and high-risk states to the continuous stream of sensor data. In this paper, we propose an event-based encoding of sensor data to reduce the effect of noises and then present an approach to efficiently model the historical influence of recent and past sensor-derived contexts on the likelihood of an adverse behavior. Next, to circumvent the lack of any confirmed negative labels (i.e., time periods with no high-risk moment), and only a few positive labels (i.e., detected adverse behavior), we propose a new loss function. We use 1,012 days of sensor and self-report data collected from 92 participants in a smoking cessation field study to train deep learning models to produce a continuous risk estimate for the likelihood of an impending smoking lapse. The risk dynamics produced by the model show that risk peaks an average of 44 minutes before a lapse. Simulations on field study data show that using our model can create intervention opportunities for 85% of lapses with 5.5 interventions per day.

WristPrint: Characterizing User Re-identification Risks from Wrist-worn Accelerometry Data.

Public release of wrist-worn motion sensor data is growing. They enable and accelerate research in developing new algorithms to passively track daily activities, resulting in improved health and wellness utilities of smartwatches and activity trackers. But, when combined with sensitive attribute inference attack and linkage attack via re-identification of the same user in multiple datasets, undisclosed sensitive attributes can be revealed to unintended organizations with potentially adverse consequences for unsuspecting data contributing users. To guide both users and data collecting researchers, we characterize the re-identification risks inherent in motion sensor data collected from wrist-worn devices in users' natural environment. For this purpose, we use an open-set formulation, train a deep learning architecture with a new loss function, and apply our model to a new data set consisting of 10 weeks of daily sensor wearing by 353 users. We find that re-identification risk increases with an increase in the activity intensity. On average, such risk is 96% for a user when sharing a full day of sensor data.

Towards detecting cocaine use using smartwatches in the NIDA clinical trials network: Design, rationale, and methodology.

Cocaine use in clinical trials is often measured via self-report, which can be inaccurate, or urine drug screens, which can be intrusive and burdensome. Devices that can automatically detect cocaine use and can be worn conveniently in daily life may provide several benefits. AutoSense is a wearable, physiological-monitoring suite that can detect cocaine use, but it may be limited as a method for monitoring cocaine use because it requires wearing a chestband with electrodes. This paper describes the design, rationale, and methodology of a project that seeks to build upon and extend previous work in the development of methods to detect cocaine use via wearable, unobtrusive mobile sensor technologies. To this end, a wrist-worn sensor suite (i.e., MotionSense HRV) will be developed and evaluated. Participants who use cocaine (N = 25) will be asked to wear MotionSense HRV and AutoSense for two weeks during waking hours. Drug use will be assessed via thrice-weekly urine drug screens and self-reports, and will be used to isolate periods of cocaine use that will be differentiated from other drug use. The present study will provide information on the feasibility and acceptability of using a wrist-worn device to detect cocaine use.

Low energy-density recording with a high-repetition-rate laser beam in gold-nanorod-embedded discs.

In this paper, we report on the low energy-density recording with a high-repetition-rate femtosecond pulsed beam in homogenous gold-nanorod-dispersed discs by using low numerical aperture (NA) micro-optics. By focusing a femtosecond pulsed beam at a repetition rate of 82 MHz using a low NA DVD optical head, the spatially-stretched energy density introduces a temperature rising of the polymer matrix. This temperature rising facilitates the surface melting of gold nanorods, which leads to over one-order-of-magnitude reduction in the energy-density threshold for recording, compared with that by focusing single pulses through a high NA objective. Applying this finding, we demonstrate the dual-layer recording in gold-nanorod-dispersed discs with an equivalent capacity of 69 GB. Our results demonstrate the potential of ultra-high density three-dimensional optical memory with a low-cost and DVD-compatible apparatus.