Thesis: S.M., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2016. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 83-88). / A line is a form that is primitive yet versatile. With its transformation capability, the line can transform into curves, surfaces and solid shapes. As a familiar form factor we encounter in everyday life in the shape of strings, tapes or wires, the form of line provides various tangible interactions, including knotting, wrapping, tying, and connecting. Lines are also used to represent abstract information, from the original drawings using line from thousands of years ago, to now when they are used to compose digital geometrical models through wireframe or Bézier data. The customization capability is another characteristic of line; strings or tape can be cut and rearranged to create various shapes and configurations. In this thesis, we will introduce "LineFORM, a novel concept for shape changing interfaces which uses physical lines to bridge the dynamic digital environment and the tangible physical world. Utilizing the versatile characteristics of lines mentioned above, we explore and define the design space for interactions with actuated curve interfaces. We implement two major versions of prototypes based on serpentine robotics technology to demonstrate a wide range of applications. For the applications, both digital and physical environment-based applications are being proposed. The system design and the implementation of LineFORM are evaluated through preliminary technical evaluation, and the limitations and future works are discussed. Through this initial design and technical exploration of actuated curve interfaces for interaction design, we envision the future where the adaptability and dynamism of digital environment will be seamlessly woven into our daily life. / by Ken Nakagaki. / S.M.

Thesis: Ph. D., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2016. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 165-182). / This thesis carries out a series of studies and develops a methodology and tools to measure and analyze ambulatory physiological, behavioral and social data from wearable sensors and mobile phones with trait data such as personality, for learning about behaviors and traits that impact human health and wellbeing. This thesis also validates the methodology and tools on a selected subset of the questions that can be answered by the data collected. First, I conducted a study to characterize wrist electrodermal activity (EDA) patterns with concurrent polysomnography and conventional palm EDA measurement. I developed a tool to analyze the EDA data quantitatively and found that wrist EDA peaks occur during Non REM2 and 3 sleep. Then, with multi-modal wearable sensor data, I conducted several studies showing how multi-modal wearable sensors can improve characterization of sleep/wake states over motion-sensing alone, and predict sleep-related memory consolidation. We found that wrist-EDA helps discriminate when there is improved sleep-related memory consolidation. Next, with colleagues at MIT and Brigham and Women's hospital, I designed and carried out the first four semesters of the "SNAPSHOT study", which measured over 100,000 hours of multi-sensor and smartphone use data from 168 college students, recruited together with their social groups. Each student contributed intensive multi-modal ambulatory data (physiological, behavioral, environmental, and social) for 30 days. Each student also filled out standardized questionnaires on mental health, personality, stress, social interactions, sleep and GPA, and provided a measure of dim light melatonin, enabling circadian phase to be measured. To investigate the value of the data, I examined a subset of the large set of questions that these new data enable us to answer: I examined the associations between sleep regularity and sleep duration on academic performance, physical/mental health, perceived stress and wellbeing-related measures using coarsened exact matching to control covariates. Our data showed that sleep irregularity was statistically significantly more associated with bad health, reported in the morning, and with worse mental health than sleep duration. I also identified features useful for recognition of monthly reported perceived stress (high vs low): daily activities, personality, sleep, physiology, social interactions, phone usage, and mobility. / by Akane Sano. / Ph. D.

Thesis: S.M., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2017. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 93-97). / Years of heavy regulation and a long-standing focus on compliance have co-opted the ability of the healthcare industry to implement novel data sharing approaches. We now face a critical need for such innovation, as personalization and data science prompt patients to engage in the details of their healthcare and restore agency over their medical data. This thesis proposes MedRec: a novel, decentralized record management system to handle EHRs (Electronic Health Records), using blockchain technology. The system design gives patients a comprehensive, immutable log and access to their medical information across providers and treatment sites. Leveraging unique blockchain properties, MedRec manages authentication, data retrieval, update tracking for existing records, data entry (both for patients and providers) and data sharing. MedRec accomplishes record management without creating any centralized data repositories; a modular system design integrates with providers' existing, local data storage solutions, facilitating interoperable data exchange between data sources and the patients. We incentivize healthcare industry stakeholders (government-funded researchers, public health authorities, etc.) to participate in the network as blockchain "miners". This provides them with access to aggregate, anonymized data as mining rewards, in return for sustaining and securing the MedRec network via Proof of Work. We emphasize the flexibility and extensibility of our system components to other dimensions of the healthcare industry and to applications beyond healthcare as well. This thesis describes the MedRec technical design and early-stage prototype, our pilot with Beth Israel Deaconess Medical Center (BIDMC), and an analysis of MedRec's contribution in the context of national healthcare priorities. This work is supported by the MIT Media Lab Consortium. / by Ariel C. Ekblaw. / S.M.

Thesis: S.M., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2017. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 44-47). / Nutritional assessment is an important problem for every American. Studies suggest that as many as 90% of Americans fall short of Vitamins D and E as a result of their regular dietary habits, and up to 50% of Americans do not get enough Vitamin A and Calcium. On the other hand, 68.8% of American adults over 20 were considered overweight or obese (had BMI of over 25), with excessive consumption of added sugars, fats, and carbohydrates being a key factor. There are two potential challenges that, if solved, may help many ordinary Americans manage their diets healthily. The first is recording dietary intake so that we have sufficient information regarding an individual's dietary pattern, and the second is interpreting nutritional profiles from the foods people are eating. It's after these two steps that nutritional intake can be inferred and insights into dietary balance can be gained. This thesis focuses on the first challenge, enabling more convenient tracking of dietary patterns supported by automatic image recognition. Our goal was to provide an improved alternative to current mainstream methods of keeping dietary records such as written records for clinical studies, or text input based digital trackers such as MyFitnessPal. Both current methods are quite tiresome, and we saw opportunities in utilizing computer vision methods to automate the recognition of what a user is eating, therefore hoping to reduce the need for manual input and making the process easier. In practice, we implemented an image classifier based on the Inception architecture of GoogLeNet, and trained it on the Food- 101 dataset. The performance of the classifier on the validation set achieved around 87% for top 5 accuracy. We then deployed our image recognition apparatus in the form of a mobile application, to examine the actual performance of this apparatus in an in-field setting with actual consumer eating patterns. The overall in-field recognition performance was around 28% (top 5), however, since only 30% of our meals observed were actually of foods belonging to the 101 classes we had trained the classifier to recognize, the in-field recognition accuracies for when foods to record were of foods we had trained on was around 92%. Furthermore, in subjective user surveys, 67% of users preferred our computer vision based apparatus to existing text input based digital trackers like MyFitnessPal, with 22% being neutral. Therefore, we believe that this approach to diet tracking is a promising one to explore in the future, as the main cause of low in-field recognition performance seems to be mainly caused by lack of coverage of the training data, and if we can curate a training set that captures the visual food domain appropriately, this approach can yield high in-field results and provide a tangibly more convenient tool for users to log and track their diets. / by Guolong Wang. / S.M.

Thesis: S.M., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2017. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 73-76). / Educational games and digital learning environments provide opportunities to collect fine-grained data on how learners engage with these technologies. The number of technologies targeted at literacy learning for children is increasing. However, the majority of them are structured and reward-based. Therefore, the users' behavior and data collected from them have the same limits. In this thesis, however, we assess children's engagement with a constructionist literacy learning app. The open ended nature of play in such an environment gives us the opportunity to analyze children's play not only through what they made while playing but also how they did it. This thesis provides an analytics pipeline from data acquisition to modeling behavioral patterns. This systematic way of capturing significant events in children's play can be used to inform stakeholders such as parents, peers and teachers and engage them with the learning process. It also gives the learning environment more intelligence on when and what to provide scaffolding on. To collect data, we ran two pilot studies and gathered audio and video recordings of play sessions. In addition, all of the children's interactions within the app were automatically logged. The fine-grained longitudinal data collected during the pilot studies provides a rich yet raw corpus. To reveal the patterns hidden in the data, the analytics pipeline parses logs of low-level interactions into abstract representations for sequences of actions in a word construction process. Next, it visualizes the process for each play session and the entire play history. Using the visualizations, I identified and annotated repeated motifs for more intentional sequences of actions during play and used supervised learning models to capture those patterns. The results of this analytical pipeline are currently being used by literacy experts to provide feedback to parents and suggest activities based on the child's process. / by Mina Soltangheis. / S.M.

Thesis: S.M., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2016. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 68-77). / Eye exams via a slit lamp are critical in screening for conditions such as cataracts, corneal opacities and pterygia early on to avert vision loss. The slit lamp, however, is a purely qualitative optical device that is bulky, expensive, can cause eye discomfort due to light sensitivity. It also requires a trained physician to operate, making it unsuitable for large-scale screening in resource-constrained settings. In this thesis, we propose a spectrum of portable anterior segment imaging solutions that can be operated by minimally trained health workers. On one end, we present a smartphone attachment with minimal optics and no electronic components beyond what is present in the smartphone itself to examine and image the anterior segment of the eye. This cost-effective, easily scalable solution would help extend the reach of anterior segment examination to extremely resource constrained settings, such as mass-screening camps, mobile ophthalmology clinics, war zones etc. On the other end, we propose purely solid-state instrumentation that employs programmable illumination and light steering optics to simulate the motion of a slit on the eye, thereby exhibiting functionality similar to that of a slit lamp with no moving parts. Finally, we discuss potential deployment strategies for two implementations of this technology in the specific cases of two contrasting healthcare systems in India. / by Shantanu Sinha. / S.M.

Thesis: S.M., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2018. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 41-43). / It is inevitable that personal architectural environments in the future will contain dozens, if not hundreds, of connected devices that will require human control. Currently, most interactions involve standalone devices, each with a unique and often non-intuitive interface, which places an unacceptable cognitive load on the occupants. To address this challenge, I propose to develop a framework that enables the devices to understand and react to the user's intentions towards the environment using pattern recognition coupled with a single user interface that is consistent across output modes. The system, designed to work within complex transformable environments, will allow the user to control furniture configuration, lighting, transparency, audio, fragrance, health systems, etc. Using machine learning, the system will correlate the user's preferences to those of others who have exhibited similar patterns of behavior in order to predict appropriate settings for novel situations. Overall, the system is expected to reduce the amount of time and cognitive load required to configure an environment for optimal comfort and utility. The proposed framework will be tested and validated in a small, re-configurable workplace environment designed to accommodate private work, phone calls, conversations, napping, and meditating. Occupants will be able to tune many parameters of the environment in response to context-aware transformations triggered by their ongoing activities. The proposed system will have an architecture based on a database infrastructure, sensor data collection, and algorithms for activity and pattern recognition. / by Lucas Cassiano Pereira Silva. / S.M.

Thesis: Ph. D., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2017. / Cataloged from PDF version of thesis. Due to the condition of the original material with text runs off the edges of the pages, the reproduction may have unavoidable flaws. / Includes bibliographical references (pages 167-172). / Generative models are venerated as full probabilistic models that randomly generate observable data given a set of latent variables that cannot be directly observed. They can be used to simulate values for variables in the model, allowing analysis by synthesis or model criticism, towards an iterative cycle of model specification, estimation, and critique. However, many datasets represent a combination of several viewpoints - different ways of looking at the same data that leads to various generalizations. For example, a corpus that has data generated by multiple people may be mixtures of several perspectives and can be viewed with different opinions by others. It isn't always possible to represent the viewpoints by clean separation, in advance, of examples representing each perspective and train a separate model for each point of view. In this thesis, we introduce lensing, a mixed-initiative technique to (i) extract lenses or mappings between machine-learned representations and perspectives of human experts, and (2) generate lensed models that afford multiple perspectives of the same dataset. We explore lensing of latent variable model in their configuration, parameter and evidential spaces. We apply lensing to three health applications, namely imbuing the perspectives of experts into latent variable models that analyze adolescent distress and crisis counseling. / by Karthik Dinakar. / Ph. D.

Thesis: S.M., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2018. / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 116-118). / In architecture, the building skin is the primary interface for mediating the environment of the external with the internal. But today, this mediation is mechanical, deterministic, and static-often seeing the human as a generalizable and problematic input. With advances in material science however, there is great potential to disrupt these traditional manufactured environments of architecture and turn them into responsive mediated environments. What this thesis aims to explore is this idea of the receptive skin-a sensate and dynamic multimaterial interface for environmental mediation. This suggests that by departing from the view that buildings are static artifacts, we may instead begin to see buildings as organic, living entities. Through the development of a working prototype, this thesis explores how such an interface may manifest itself, through dynamic material composites, instead of mechanical and electronic means. The final prototype is a "proof of concept," a built example of this novel design methodology, which unites material performance with sensate technologies, as a way to enable new interactions between building and environment. / by Chrisoula Kapelonis. / S.M.

Thesis: S.M., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2016. / "September 2016." Cataloged from PDF version of thesis. / Includes bibliographical references. / (lm)possible Baby is a speculative design project which aims to stimulate discussions about the social, cultural and ethical implications of emerging biotechnologies that could enable same-sex couples to have their own, genetically related children. Delivering a baby from same-sex parents is starting to not look like a sci-fi dream anymore recent developments in genetics and stem cell research, such as the achievements of scientists from Cambridge University in England and Israel's Weizmann Institute of Science 1 have made this dream much closer to reality. Jacob Hanna, the specialist leading the project's Israeli arm, said it may be possible to use the technique to create a baby in just two years. "It has already caused interest from gay groups because of the possibility of making egg and sperm cells from parents of the same sex," he said."2 Is creating a baby from same-sex parents the ethical thing to do? Who has the right to decide this, and how? This project aims to design and inspire debate about the bioethics of producing babies from same-sex couples. In this project, the DNA data of a lesbian couple was analyzed using 23andMe to simulate and visualize their potential children, and then we created a set of fictional, "what if' future family photos using this information to produce a hardcover album which was presented to the couple as a gift. To achieve more public outreach, we worked with the Japanese national television service, NHK, to create a 30-minute documentary film following the whole process, which aired in October 2015. / by Ai Hasegawa. / S.M.