Amulet

Hester, Josiah; Peters, Travis; Yun, Tianlong; Peterson, Ronald; Skinner, Joseph; Golla, Bhargav; Storer, Kevin; Hearndon, Steven; Freeman, Kevin; Lord, Sarah; Halter, Ryan J.; Kotz, David; Sorber, Jacob

doi:10.1145/2994551.2994554

Cited by 46 publications

(3 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Resource constraints: Intermittent systems are severely resource-constrained. In this paper we study an intermittent system built using a TI MSP430 microcontroller (MCU), which is the most commonly used processor in existing intermittent systems [17,[37][38][39]71]. Such an MCU's frequency is typically 1-16MHz, leaving a substantial performance gap compared to, e.g., a full-fledged, 2GHz Xeon-based system.…”

Section: Intermittent Execution On Energy-mentioning

confidence: 99%

Intelligence Beyond the Edge

Gobieski

Lucia

Beckmann

2019

Proceedings of the Twenty-Fourth International Conference on Architectural Support for Programming Languages and Operating Syst

151

View full text Add to dashboard Cite

Energy-harvesting technology provides a promising platform for future IoT applications. However, since communication is very expensive in these devices, applications will require inference "beyond the edge" to avoid wasting precious energy on pointless communication. We show that application performance is highly sensitive to inference accuracy. Unfortunately, accurate inference requires large amounts of computation and memory, and energy-harvesting systems are severely resource-constrained. Moreover, energy-harvesting systems operate intermittently, suffering frequent power failures that corrupt results and impede forward progress. This paper overcomes these challenges to present the first full-scale demonstration of DNN inference on an energyharvesting system. We design and implement SONIC, an intermittence-aware software system with specialized support for DNN inference. SONIC introduces loop continuation, a new technique that dramatically reduces the cost of guaranteeing correct intermittent execution for loop-heavy code like DNN inference. To build a complete system, we further present GENESIS, a tool that automatically compresses networks to optimally balance inference accuracy and energy, and TAILS, which exploits SIMD hardware available in some microcontrollers to improve energy efficiency. Both SONIC & TAILS guarantee correct intermittent execution without any hand-tuning or performance loss across different power systems. Across three neural networks on a commercially available microcontroller, SONIC & TAILS reduce inference energy by 6.9× and 12.2×, respectively, over the state-of-the-art. CCS Concepts • Computer systems organization → Embedded software.

show abstract

Section: Intermittent Execution On Energy-mentioning

confidence: 99%

Intelligence Beyond the Edge

Gobieski

Lucia

Beckmann

2019

Proceedings of the Twenty-Fourth International Conference on Architectural Support for Programming Languages and Operating Syst

151

View full text Add to dashboard Cite

show abstract

“…Although these sensors show promise, they are not without challenges, as reliability can be an issue when ambient energy is not readily available. Wearable sensors are increasingly being developed to last several months [91,92], but most commercial sensors las t <12 hours [93] when attempting to collect continuous inertial measurement unit data. Low-maintenance sensor solutions must be designed, and careful consideration of battery lifetime must exist in every phase of system and study design.…”

Section: Technology-enabled Domains For Measuring Calorie and Macronumentioning

confidence: 99%

Counting Bites With Bits: Expert Workshop Addressing Calorie and Macronutrient Intake Monitoring

et al. 2019

Self Cite

View full text Add to dashboard Cite

BackgroundConventional diet assessment approaches such as the 24-hour self-reported recall are burdensome, suffer from recall bias, and are inaccurate in estimating energy intake. Wearable sensor technology, coupled with advanced algorithms, is increasingly showing promise in its ability to capture behaviors that provide useful information for estimating calorie and macronutrient intake.ObjectiveThis paper aimed to summarize current technological approaches to monitoring energy intake on the basis of expert opinion from a workshop panel and to make recommendations to advance technology and algorithms to improve estimation of energy expenditure.MethodsA 1-day invitational workshop sponsored by the National Science Foundation was held at Northwestern University. A total of 30 participants, including population health researchers, engineers, and intervention developers, from 6 universities and the National Institutes of Health participated in a panel discussing the state of evidence with regard to monitoring calorie intake and eating behaviors.ResultsCalorie monitoring using technological approaches can be characterized into 3 domains: (1) image-based sensing (eg, wearable and smartphone-based cameras combined with machine learning algorithms); (2) eating action unit (EAU) sensors (eg, to measure feeding gesture and chewing rate); and (3) biochemical measures (eg, serum and plasma metabolite concentrations). We discussed how each domain functions, provided examples of promising solutions, and highlighted potential challenges and opportunities in each domain. Image-based sensor research requires improved ground truth (context and known information about the foods), accurate food image segmentation and recognition algorithms, and reliable methods of estimating portion size. EAU-based domain research is limited by the understanding of when their systems (device and inference algorithm) succeed and fail, need for privacy-protecting methods of capturing ground truth, and uncertainty in food categorization. Although an exciting novel technology, the challenges of biochemical sensing range from a lack of adaptability to environmental effects (eg, temperature change) and mechanical impact, instability of wearable sensor performance over time, and single-use design.ConclusionsConventional approaches to calorie monitoring rely predominantly on self-reports. These approaches can gain contextual information from image-based and EAU-based domains that can map automatically captured food images to a food database and detect proxies that correlate with food volume and caloric intake. Although the continued development of advanced machine learning techniques will advance the accuracy of such wearables, biochemical sensing provides an electrochemical analysis of sweat using soft bioelectronics on human skin, enabling noninvasive measures of chemical compounds that provide insight into the digestive and endocrine systems. Future computing-based researchers should focus on reducing the burden of wearable sensors, aligning data across m...

show abstract

“…Ryan Halter (Engineering) described four projects underway at Dartmouth, beginning with Amulet [2,3], a collaborative effort to develop a wearable computing platform that can run multiple mHealth apps with strong security guarantees and weeks-long battery life. In a related project, Halter’s group is developing a wristband to measure electro-dermal activity which, in combination with the Amulet and a chest-strap heart-rate monitor, they use to measure stress in free-living conditions.…”

Section: Mobile Sensingmentioning

confidence: 99%

Workshop on Emerging Technology and Data Analytics for Behavioral Health

et al. 2018

Self Cite

View full text Add to dashboard Cite

Wearable and portable digital devices can support self-monitoring for patients with chronic medical conditions, individuals seeking to reduce stress, and people seeking to modify health-related behaviors such as substance use or overeating. The resulting data may be used directly by a consumer, or shared with a clinician for treatment, a caregiver for assistance, or a health coach for support. The data can also be used by researchers to develop and evaluate just-in-time interventions that leverage mobile technology to help individuals manage their symptoms and behavior in real time and as needed. Such wearable systems have huge potential for promoting delivery of anywhere-anytime health care, improving public health, and enhancing the quality of life for many people. The Center for Technology and Behavioral Health at Dartmouth College, a P30 “Center of Excellence” supported by the National Institute on Drug Abuse at the National Institutes of Health, conducted a workshop in February 2017 on innovations in emerging technology, user-centered design, and data analytics for behavioral health, with presentations by a diverse range of experts in the field. The workshop focused on wearable and mobile technologies being used in clinical and research contexts, with an emphasis on applications in mental health, addiction, and health behavior change. In this paper, we summarize the workshop panels on mobile sensing, user experience design, statistics and machine learning, and privacy and security, and conclude with suggested research directions for this important and emerging field of applying digital approaches to behavioral health. Workshop insights yielded four key directions for future research: (1) a need for behavioral health researchers to work iteratively with experts in emerging technology and data analytics, (2) a need for research into optimal user-interface design for behavioral health technologies, (3) a need for privacy-oriented design from the beginning of a novel technology, and (4) the need to develop new analytical methods that can scale to thousands of individuals and billions of data points.

show abstract

Amulet

Cited by 46 publications

References 24 publications

Intelligence Beyond the Edge

Intelligence Beyond the Edge

Counting Bites With Bits: Expert Workshop Addressing Calorie and Macronutrient Intake Monitoring

Workshop on Emerging Technology and Data Analytics for Behavioral Health

Contact Info

Product

Resources

About