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 157-162). / With recent developments in visual interface technologies, which are now capable of delivering rich and highly immersive experiences, the need has arisen to develop equally capable haptic-feedback technologies that can complement and be easily integrated with their visual counterparts to enable fully immersive, multimodal, interactive experiences. Moreover, with touchscreen devices rapidly replacing physical controls, as in the case of automobiles and industrial equipment, there has been a loss of natural haptic cues and an increased need for operators to divert their attention from the task at hand into the touchscreen just to find the locations of different controls - which previously they could feel by relying on haptic cues without having to employ the visual sense. This is especially concerning for drivers, where even a brief loss of attention could lead to catastrophic consequences. This work aims to begin addressing these problems by presenting an approach - based on air-vortex-ring, free-space haptics - that enables existing visual interfaces to be augmented with haptic feedback capabilities. After presenting our goals and motivations for this work, the latest approaches to haptic feedback interaction, and a discussion of the advantages and limitations of each approach, we describe the toroidal air-vortex approach to free-space haptics in full detail. We then present a multimodal interactive interface system that we built based on that approach, named AirTap, which is the main focus of this thesis. AirTap is an open-source system that uses 16 air-vortex-ring generators to deliver targeted, localized, unambiguous, free-space haptic feedback onto a user's hand when interacting with 3D virtual objects on a stereoscopic display. We show how AirTap can serve as a multimodal interactive interface, as a research tool for studying air-vortex-ring based haptics, and as an open platform for creative expressions. Finally, we provide an extensive discussion of how our system can be extended beyond its present form and adopted for automotive and other applications. / by Ali Shtarbanov. / S.M.

Thesis: Ph. D., 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 261-273). / Understanding how light travels through macroscopic scenes can transform autonomous driving, medical imaging and consumer photography. Unfortunately, this understanding is difficult to achieve: trillions of light paths are measured by millions of pixels. The framework of computational light transport was introduced to model this complex interaction between light and matter in a tractable space. In this thesis, we study new methods to invoke space, time, and polarization into a computational light transport framework. First, we study how probing the time dimension enables cameras to separate bounces from multiple light paths. Our solutions are inspired by prior work on multipath in wireless and telecommunications. We then invoke both time and space to provide the first provable bound on resolution for seeing around corners or through scattering media. Finally, we jointly invoke space, time, and polarization to propose an ultra-high quality form of 3D imaging. This thesis contributes a few analytical theories, including: (1) provable bounds on multipath separation; (2) provable bounds on seeing around corners; and (3) proof of shape reconstruction from polarimetric measurements. The thesis also contributes new applications that span: (a) micron-scale 3D cameras; (b) real-time object tracking around corners; and (c) single-shot computational relighting of images. Future applications encompass equipping self-driving cars the ability to see through fog, or enabling doctors to see deeper inside the body using light. / by Achuta Kadambi. / Ph. D.

Thesis: Ph. D., 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 251-281). / The vision of virtual reality has always been to create worlds that look, sound, act, and feel real. However, researchers and developers have largely favored visual perception over other senses. This over-valuation of the visual may be traced back to a partial interpretation of the seminal work on visual perception by psychologist JJ Gibson. Oculocentrism in design overlooks the fact that Gibson's theory of perception encompasses the entire range of perceptual processes integrated with action, including kinesthesia and affordances of the environment. Starting with Gibson's ecological approach to the reality of experience, I develop a four-dimensional framework for creating immersive experiences that blend extrinsic elements, meaning elements related to the user's real world context, and intrinsic elements, i.e., those related to the device, application and content. I present a series of novel methods and techniques, demonstrated through implemented systems to show how transferring real world affordances to virtual experiences can enhance the sense of presence, while also arguing for a shift from oculocentrism to sensorimotor processes and to the experiential modalities of touch, proprioception, and kinesthesia. My work contrasts with the currently dominant design approach premised on the notion that the richness of sensory perception can be recreated with vision alone. The hybrid systems described in this thesis present techniques for integrating space, kinesthesia, touch and other sensations, social interaction, and the user's physiology into the virtual experience. / by Misha Sra. / 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 78-80). / Acknowledging and reflecting upon the cultural and scientific discourses that shape the notions the female body, this thesis presents a biopolitical feminist work in which a woman controls the motility of spermatozoa through the agency of her thoughts. The intellectual investigation of this project is threefold: i. to engineer a system that directs the movement of sperm via the signals of brain activity, 2. to communicate the project aesthetically and expressively through art and design, 3. to situate the project philosophically and pose critical cultural questions. Technologically, this project implements a brain-computer interface, where the electric signals generated by the brain are translated into a system engineered to control the movement of sperm through a phenomenon known as galvanotaxis. Philosophically, this project is situated within Michel Foucault's notion of biopolitics, and Donna Haraway's theorization of the body as a material-semiotic actor. While Foucault lays the foundation for exposing the body as a battleground for political power, the ideas of Haraway respond to the inherent gender prejudices of this landscape by disassembling the binary of sex through technology to spark a reimagining of new corporeal futures. Artistically, this thesis creates an act of female empowerment, responding to political regimes in which women are losing rights related to procreation within her own body. Investigating the body as a medium of culture, this project raises questions as to how we operate in our politically gendered landscape. Navigating the divergent connections between art and science, this work challenges the viewer to question what is possible. Whereas our biological understanding of sperm is usually deterministic (i.e., as an inherent homing device racing towards the chemical signatures of an egg) or colored by gendered cultural constructs (i.e., its semiotic use in pornography) this project seeks to invert all preconceived notions. By creating a work that is simultaneously technological, functional, and symbolically potent, it seeks to expand our notions of what it is possible, and what is possible to question. / by Anne [Ani] Liu. / 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 105-109). / Present urban planning issues require to involve the public in the urban design process, and this slow and complicated process remains the primary domain of expert planners and consultants. Although there have been many attempts to leverage new mobile tools to engage the community. These tools support the three stages of planning 1. data collection 2. analysis and visualization three solutions. Within these tools, some gather unstructured data that is hard to convert into physical interventions. Also, some applications are not designed to encourage debate and consensus building. This study will consider how a structured integrated tool will help the process of grassroots urban design. This thesis will focus on the development of a bottom-up, crowd-sourced, urban planning tool to improve the quality and safety of urban bike lanes. A mobile application will be developed to enable non-experts to actively participate in the process of real time data collection and feedback, mapping, selection of solutions, and the establishment of priorities. The system will be evaluated using both quantitative and qualitative methods, compared to present methods on bottom up interventions. / by Yasushi Sakai. / S.M.

Thesis: Ph. D., 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 96-106). / Data driven methods based on deep neural networks (DNNs) have ushered in a new era in the field of machine learning computer vision. Conventional algorithmic approaches are being replaced by end-to-end deep learning systems that can leverage big data. Deep learning has begun revolutionizing human centric fields such as health-care and finance, finding its way into automated screening and diagnoses. At present, developing and training artificial neural network architectures requires both human expertise and labor, requiring millions of labeled data-points to train and hours of engineering effort to develop best performing architectures. In this dissertation, my goal is to make deep learning more accessible by developing algorithms for low shot learning (learning from a few examples). This work includes new semi-supervised approaches to learn from unlabeled datasets with only a fraction of labeled examples, deep learning methods to learn from generated data using simulation based techniques, and learning to optimize neural networks for smaller data sets. Specifically, this dissertation focuses on two proposed directions which will contribute towards both technical and conceptual advances in literature. -- How can we use invariant-based approaches when training from small datasets ? -- How to enable training from multiple data sources carrying very small amounts of data ? -- How to use meta-modeling approach to automatically generate high-performing DNNs ? To address these questions, this dissertation describes machine learning algorithms as follows (a) an action recognition autoencoder which learns over very small datasets; (b) an algorithm to train deep neural networks over multiple entities; (c) a meta-modeling approach to automatically generate high-performing architectures. We also provide a dataset of neural network topologies used for predicting accuracy of a deep neural network. / by Otkrist Gupta. / Ph. D.

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 93-95). / This thesis explores how we can help parents learn English through peer-learning experiences with their children. I discuss the design and study of the learning experience facilitated by "Read Out Loud," a mobile application that supports parents who are learning English as they read storybooks with their children. The mobile application and the context of its use provide a new interface to language learning tools such as text-to-speech, word translation, and shared-reading prompts. The thesis develops a set of design and user-testing guidelines that creates affordances for parents who are learning English. It builds on an iterative design process that includes two user studies of parents using the Read Out Loud application while reading with their children. Through observation and interviews, I investigate how technologies like Read Out Loud can be designed to motivate and augment the family reading experience and how familial factors may influence a parent's use of and interaction with the technology. / by Juliana Toni Nazare. / 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 97-100). / Electrodermal activity (EDA) refers to the electrical activity measured on and under the surface of the skin and has been used to study sleep, stress, and mood. While gathering this signal was once confined to the laboratory, it can now be acquired in ambulatory studies through commercially available wearable sensors. In this thesis, we model and analyze electrodermal response (EDR) events (1-5 second peaks in the EDA signal) during sleep in an ambulatory study. In particular, we describe an EDR event detection algorithm and extract shape features from these events to discuss the difference in shape between sleep and wake. We also describe an automatic artifact detection algorithm that we use on over 100,000 hours of EDA data we have collected in the 30-day SNAPSHOT Study from 164 participants. Finally, we model the detected EDR events as a point process using a state-space generalized linear model. We identify a significant influence of recent EDR event history on current EDR event likelihood across different participants. We also use this model to analyze EDR event rates during different periods of the night. / by Sara Ann Taylor. / 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 102-110). / Our sense of smell is perhaps the most pervasive of all senses, as it has the ability to evoke memories and emotions in a vivid and subtle manner. While olfactory communication is one of the most basic methods of communication, it is also one of the least understood and one of the least exploited in human computer interaction. In this thesis we describe the hidden power of scent along with the design and implementation of "Essence", a custom made olfactory wearable device and its stationary version. Essence is a necklace that can be used by any person in their daily life for the purpose of altering one's mood, as well as affecting cognitive and psychological conditions. It can influence the user's behavior through consciously perceivable as well as subliminal bursts of scent released while the person is asleep or awake. The device can be remotely controlled through a smartphone and can vary the intensity and frequency of the released scent. The system can also potentially be triggered by physiological data such as brain signals, heart rate, or galvanic skin response, etc. The types of scent that can be placed in the necklace can range from essential oils to odorless scents like hormones and pheromones. We conducted a set of preliminary studies that show an overall satisfaction, comfort and ease of use of the system. We also demonstrate the effectiveness for mood enhancing and cognitive performance during wakefulness and sleep state. / by Judith Amores Fernandez. / 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 159-167). / There are many health-related risks associated with chronic stress. One approach for addressing this issue comes from strengthening inside self-regulation abilities rather than eliminating outside sources of stress. However, technology has not been explored to its full potential for delivering calming interventions. While many of the persuasive technologies developed for fostering behavior change focus on cognitive processes, little attention has been given to influencing behavior through automatic processes of the brain. Considering the bidirectional relationship between psycho-physiological signals and self-reported emotional states, manipulating the physiological signals that a human being has voluntary control over is promising for achieving a desired emotional state. In this work, we focus on respiration due to the fact that it is both a voluntary and involuntary response of the body. It is a good indicator of stress, but can also be manipulated to induce calmness. This thesis introduces BrightBeat: a set of seamless visual, auditory, and tactile interventions that mimic a calming breathing oscillation with the aim of influencing physiological syncing and consequently bringing a sense of focus and calmness. These interventions are designed to run easily on commonplace personal electronic devices, respect the user's privacy, and to not require constant focus or attention in order to be effective. We have designed BrightBeat interventions iteratively and have examined both objective and subjective measures of impact through a series of studies with N=54 users in total. From an objective perspective, BrightBeat interventions significantly influenced calmer (slower) breathing and had a lasting influence. From a subjective perspective, considering the individual differences, these interventions have been shown to improve self-reported calmness and focus. Also, participants reported high preference for using them in the future. This work has been supported by the Robert Wood Johnson Foundation and the MIT Media Lab Consortium. / by Asma Ghandeharioun. / S.M.