Thesis: Ph. D., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, February, 2021 / Cataloged from the official PDF of thesis. / Includes bibliographical references (pages 218-240). / Synergetic tensions have evolved the dichotomy between the physical and digital design domains into a symbiotic unity. New capabilities in digital fabrication give rise to sophisticated tools of computational design, while new affordances in computational design inspire innovation in digital fabrication. The role of design in this process is that of synthesis through mediation. As designers, we mediate between different principles and fields, and their synergies and conflicts generate new elements of design. The challenge to mediate in a universal language across domains becomes critical as a third domain encompassing biological entities grows more amenable to design. Biological systems offer reproduction, self-organization and growth -- among other features and benefits -- / which in turn enable previously unattainable properties to design systems. At the same time, their own modes of intelligence, expression, and agency demand a promising shift in design thinking. This thesis hypothesizes that the relations across design domains can be established through translational design computation, which is a framework that uses computational design as a language to mediate between physical, digital, and biological entities. We build this framework in two parts -- / Systems and Mediations. The first part, Systems, explores whether computational design can serve as a mediating language between the three entities. The second part, Mediations, examines how these mediations can occur. In Systems, we show that computational design can mediate between living and nonliving matter along the spectrum of biomimetic, biointegrated, and biosynthetic systems. As part of this, we demonstrate three systems of computational mediation: (i) programmable matter applies computational design to physical systems to enable biologically inspired design strategies, (ii) programmable templating applies computational design to the intersection of physical and biological systems to facilitate synergistic relationships, and (iii) programmable growth applies computational design to biological systems to give rise to material architectures. / In Mediations, we present dynamic, synergetic, and emergent strategies for how computational mediations can occur within cocreation systems. The living and nonliving parts of any cocreation system may interact to form synergies. Combined, these synergies produce complexes that give rise to new macro-level organizations -- products of the synergies of the parts and not simply of the parts themselves. Thus, the mediation between physical, digital, and biological entities needs to address the design of dynamic relations guiding synergetic behaviors, the design of the synergetic behaviors themselves or ultimately, the design of emergent self-expression of the system. Throughout this thesis, the framework is developed theoretically and applied in practice. It is documented in publications such as Making Data Matter and Hybrid Living Materials and projects such as Wanderers, Living Mushtari, the Vespers Series, Rottlace, Lazarus, Totems, Fiberbots, and Silk Pavilion II. / by Christoph Bader. / Ph. D. / Ph.D. Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences

Thesis: Ph. D., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, February, 2021 / Cataloged from the official PDF of thesis. / Includes bibliographical references (pages 279-290). / Children learn best when knowledgeable adults support their learning process. Yet many learning technologies have not yet incorporated this vital social dimension. In response, we develop a technology-assisted coaching system, where a new adult collaborator--a coach--uses digital tools to support children and their families as they use children's literacy apps. This coaching system blends in-person and digital coaching in order to preserve the relational elements vital to coaching, while harnessing the power of digital technology to make information easily accessible to coaches, children, and families at their convenience. In our system, as children play with literacy apps, every tap and click of their play is streamed to their coach through our digital coaching platform. Using custom-built digital tools, coaches engage in four core coaching practices. / They analyze children's in-app activity, scaffold their learning, share progress with caregivers, and invite caregivers to engage in literacy learning experiences with their children. To develop this system, we iteratively designed, built, and evaluated it with approximately a hundred children and their families. We conducted in-depth study of two versions of the system's design through a randomized control trial (RCT) and a formative pilot study. From the RCT, we found that the coaching system increased caregivers' awareness of their children's in-app play and children's playtime with the literacy app. We also found that for families with lower formal education levels, the effect of the coaching system was greater across almost all outcome variables investigated. In both studies, we found coaches were able to use our digital coaching tools to effectively engage in the four core coaching practices, and that these tools helped increase coaches' efficiency. / Based on our findings, we discuss changes to the system's design to improve and scale this approach and provide design considerations for building digital coaching systems. Through the creation and in-depth study of a novel sociotechnical system for coaching children's literacy learning, this work contributes to the field of learning technology. We hope this work serves as a helpful guide to designers, developers, and policymakers as they create and scale-up these types of digital networks for children's learning. / by Juliana Toni Nazare. / Ph. D. / Ph.D. Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences

Thesis: S.M., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, February, 2021 / Cataloged from the official PDF version of thesis. / Includes bibliographical references (pages 57-58). / Ethical trade-off surveys have played a key role in building a data-driven understanding of human moral psychology. They have been conducted all over the world for decades, eliciting assessment of ethical dilemma outcomes from populations as diverse as those of rural, tribal settlements, and industrialized, information-age, cosmopolitan cities. While much data has been gathered through these surveys, attempts to compare what people across cultures consider ethically justifiable have been hindered by the fact that the surveys used have been reformulated for different cultures in the scenarios they depict, and in their framing. The objective of this thesis project is to build a survey tool with global reach and internationalized surveys, in order to collect survey data from around the world using consistent scenarios and framing. Building on the precedent and success of the Moral Machine tool for surveying people around the world regarding ethical dilemmas involving autonomous vehicles, I built and deployed a tool for conducting surveys with scenarios of the classic action/omission trolley problem, to collect ethical dilemma survey data internationally, in ten languages, for three variants of the trolley problem - one for remote action/omission with no double effect consideration, one for double effect consideration with direct action/omission, and one for double effect consideration with remote action/omission. Analyzing data from this experiment, I conclude that differences in preferences across the variants are confirmed across populations, and that they are universal across populations in order of preference. / by Sohan Savio Dsouza. / S.M. / S.M. Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences

Thesis: S.M., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, February, 2021 / Cataloged from the official PDF version of thesis. / Includes bibliographical references (pages 49-59). / Negative emissions technologies that can remove carbon dioxide from the atmosphere are a critical tool to limit global temperature rise and ocean acidification. Bioengineering capabilities have not been sufficiently assessed or utilized for the development of negative emissions technologies. Bioengineering holds the potential to improve the efficiency of some existing technologies and to create new methods of carbon removal. I review existing technologies to assess how bioengineering could improve them, focusing on technologies that could achieve at least 1 Gt of CO₂ removal per year. I also investigate and describe potential new methods of carbon removal that leverage bioengineering. Key questions for additional research are identified, as are key engineering targets for the development of improved negative emissions technologies. This evaluation of potential high-impact R&D work is intended to provide an initial roadmap for the development of bioengineered negative emissions technologies that are scalable, sustainable, and can remove gigatons of CO₂ from the atmosphere. / by Sarah Mary Haiken Sclarsic. / S.M. / S.M. Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences

Thesis: Ph. D., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, September, 2019 / Cataloged from the official PDF of thesis. "The Table of Contents does not accurately represent the page numbering"--Disclaimer page. / Includes bibliographical references (pages 117-126). / How can we build machines that collaborate and learn more seamlessly with humans, and with each other? How do we create fairer societies? How do we minimize the impact of information manipulation campaigns, and fight back? How do we build machine learning algorithms that are more sample efficient when learning from each other's sparse data, and under time constraints? At the root of these questions is a simple one: how do agents, human or machines, learn from each other, and can we improve it and apply it to new domains? The cognitive and social sciences have provided innumerable insights into how people learn from data using both passive observation and experimental intervention. Similarly, the statistics and machine learning communities have formalized learning as a rigorous and testable computational process. / There is a growing movement to apply insights from the cognitive and social sciences to improving machine learning, as well as opportunities to use machine learning as a sandbox to test, simulate and expand ideas from the cognitive and social sciences. A less researched and fertile part of this intersection is the modeling of social learning: past work has been more focused on how agents can learn from the 'environment', and there is less work that borrows from both communities to look into how agents learn from each other. This thesis presents novel contributions into the nature and usefulness of social learning as an inductive bias for reinforced learning. / I start by presenting the results from two large-scale online human experiments: first, I observe Dunbar cognitive limits that shape and limit social learning in two different social trading platforms, with the additional contribution that synthetic financial bots that transcend human limitations can obtain higher profits even when using naive trading strategies. Second, I devise a novel online experiment to observe how people, at the individual level, update their belief of future financial asset prices (e.g. S&P 500 and Oil prices) from social information. I model such social learning using Bayesian models of cognition, and observe that people make strong distributional assumptions on the social data they observe (e.g. assuming that the likelihood data is unimodal). / I were fortunate to collect one round of predictions during the Brexit market instability, and find that social learning leads to higher performance than when learning from the underlying price history (the environment) during such volatile times. Having observed the cognitive limits and biases people exhibit when learning from other agents, I present an motivational example of the strength of inductive biases in reinforcement learning: I implement a learning model with a relational inductive bias that pre-processes the environment state into a set of relationships between entities in the world. I observe strong improvements in performance and sample efficiency, and even observe the learned relationships to be strongly interpretable. / Finally, given that most modern deep reinforcement learning algorithms are distributed (in that they have separate learning agents), I investigate the hypothesis that viewing deep reinforcement learning as a social learning distributed search problem could lead to strong improvements. I do so by creating a fully decentralized, sparsely-communicating and scalable learning algorithm, and observe strong learning improvements with lower communication bandwidth usage (between learning agents) when using communication topologies that naturally evolved due to social learning in humans. Additionally, I provide a theoretical upper bound (that agrees with our empirical results) regarding which communication topologies lead to the largest learning performance improvement. / Given a future increasingly filled with decentralized autonomous machine learning systems that interact with humans, there is an increasing need to understand social learning to build resilient, scalable and effective learning systems, and this thesis provides insights into how to build such systems. / by Dhaval D.K. Adjodah. / Ph. D. / Ph.D. Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences

Thesis: S.M., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2019 / Cataloged from PDF of thesis. "The Table of Contents does not accurately represent the page numbering"--Disclaimer page. / Includes bibliographical references (pages 119-124). / Children's creativity - the ability to come up with novel, surprising, and valuable ideas - has been known to contribute to their learning outcomes and personal growth. Standardized ways to measure creativity and divergent thinking reported that as children enter elementary school, their creativity slumps and thinking becomes more convergent, especially around the 4th grade. One cause for this is school curricula become more structured and lose the aspect of creative play. This is especially concerning for kids growing up in the era of Artificial Intelligence, where mechanical and repetitive jobs that require structured thinking move to machines. To be successful in this world of intelligent agents, we must empower children not only to understand how these intelligent agents work, but also to be able to think creatively about generating new artifacts in consort with such agents, which requires imaginative novel thought. / In this thesis, I explore whether a social robot's interaction with children can be an effective way to help children think more creatively. I suggest two ways in which robots used as pedagogical tools can help children think more creatively are: 1. through artificial creativity demonstration, such as showing the use of novel ideas, and 2. through offering creativity scaffolding, such as asking reflective questions, validating novel ideas, and engaging in creative conflict. I designed four collaborative game-based activities that involve child-robot interaction and afford different forms of creative expression: 1. Droodle Game, which affords verbal creativity, 2. Magic Draw, which affords figural creativity, 3. WeDo Construction with Jibo, which affords construction creativity and 4. Escape Adventure, which affords divergent thinking and creative problem solving. / I designed the behavior of the robot such that it either scaffolds the child for creative thinking, or the robot gives the appearance of creative thinking by artificially emulating human creativity. I evaluated the role of the social robot in influencing children's creativity by running comparative studies between children playing these creativity games while interacting with the robot with creativity-inducing behaviors (creative condition), and without creativity-inducing behaviors (non-creative condition). Children who interacted with the creative robot exhibited higher levels of creativity than children who interacted with a non-creative control robot. I conclude that children can model a social robotic peer's creative expression via social emulation. When scaffolded for creativity, children exhibited higher levels of creativity. This enabled me to develop a robot scaffolding paradigm which fosters creativity in young children. / This thesis contributes design guidelines for child-robot interactions which promote creative thinking, and provides evidence that these creativity inducing behaviors exhibited by social robots can foster creativity in young children. / by Safinah Arshad Ali. / S.M. / S.M. Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences

Thesis: Ph. D., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2018 / Cataloged from student-submitted PDF of thesis. / Includes bibliographical references (pages 193-210). / Broadly speaking, time-resolved phenomena refers to three dimensional capture of a scene based on the time-of-flight principle. Since speed and and time are proportional quantities, knowing time-of-flight allows one to estimate distances. This time-of-flight may be attributed to a pulse of light or a wave packet of sound. Depending on the sub-band of the electromagnetic spectrum, the interaction of waves or pulses with the scene of interest results in measurements and based on this proxy of the physical world, one is interested in inferring physical properties of the scene. This may be something simple as depth, or something more involved such as fluorescence lifetime of a biological sample or the diffusion coefficient of turbid/scattering medium. The goal of this work is to develop a unifying approach to study time-resolved phenomena across various sub-bands of the electromagnetic spectrum, devise algorithms to solve for the corresponding inverse problems and provide fundamental limits. Sampling theory, which deals with the interplay between the discrete and the continuous realms, plays a critical role in this work due to the continuous nature of physical world and the discrete nature of its proxy, that is, the time-resolved measurements. / by Ayush Bhandari. / Ph. D. / Ph.D. Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences

Thesis: S.M., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, May, 2020 / Cataloged from the official PDF of thesis. / Includes bibliographical references (pages 51-57). / Machine learning (ML) has increasingly been used to address the growing burden of mental illness and lack of access to quality mental health care. Recently such models have been applied to online data, such as social media postings to augment mental health screening. Despite the potential of these methods, online ML classifiers still perform poorly in multi-class settings. In this thesis, we propose the usage of novel document embeddings and mental health based user embeddings for triaged suicide risk screening. Machine learning to infer suicide risk and urgency is applied to a dataset of Reddit users in which the risk and urgency labels were derived from crowdsource consensus. We show that the document embedding approach outperforms count-based baselines and a method based on word importance, where important words were identified by domain experts. We examine interpretable features and methods that help to discern and explain risk labels. Finally, we find, using a Latent Dirichlet Allocation (LDA) topic model, that users labeled at-risk for suicide post about different topics to the rest of Reddit than non-suicidal users. / by Noah C. Jones. / S.M. / S.M. Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences

Thesis: S.M., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, September, 2020 / Cataloged from the official PDF version of thesis. / Includes bibliographical references (pages 99-102). / This thesis describes the design of a novel modular acoustic communication device for underwater wireless communication. The ocean plays a vital role in the global climate system and biosphere, providing a wealth of biodiversity and resources. Human exploitation, pollutants, and contaminants have already impacted the deepest trenches of the ocean. Yet most of this impact remains invisible - the depth and breadth of the ocean, the opacity of water to light, and the lack of mass-deployable ocean instrumentation, mean that current observations are wildly under sampled in space and time. Even given suitable instruments, a major challenge remains: how to get the data home. Given the opacity of sea water to radio, and the complexity and expense of underwater cabling, audio communication is in may ways the low hanging fruit. Unfortunately, most commercial audio communication systems are both extremely expensive and more powerful than needed for many, if not most, monitoring requirements. / They are also proprietary, a frustrating barrier to development of novel devices. As a result, these off-the-shelf systems are ill-suited to the global multi-scale instrumentation challenges of the future. In what follows we document the design, development, and testing of a low-cost, high-efficiency, modular, and fully open-sourced acoustic communication system. This system is specifically intended for mass deployment, with particular emphasis on standardized hardware and software interfaces. Crucially, this design prioritizes cost and simplicity over performance - as such, this system does not aim to replace the current crop of commercially available systems, but rather to provide a new kind of tool specifically for scalable deployments of low-cost instruments. / Functionally, the system is composed of three core modules: a power management unit; a central processing unit; and a family of interchangeable acoustic transceiver units operating across a spectrum of transmission frequencies. The resulting design should be of valuable use to the scientific community, environmental agencies, citizen scientists, and anyone who needs to transport low-bandwidth data through the ocean at extremely low cost. / by Charlene Xia. / S.M. / S.M. Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences

Thesis: S.M., Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, September, 2020 / Cataloged from the official PDF version of thesis. / Includes bibliographical references. / We explore the vision of closed-loop bio-digital interfaces for human augmentation, where the bio-digital system allows for both sensing and writing biological information to the body. Current-generation wearable devices sense an individual's physiological data such as heart rate, respiration, electrodermal activity, and EEG, but lack in sensing their biological counterparts, which drive the majority of individual's physiological signals. On the other hand, biosensors for detecting biochemical markers are currently limited to one-time use, are non-continuous and don't provide flexibility in choosing which biomarker they sense. We believe that the future for wearable biosensors lies in going beyond specific sensing capabilities and becoming a wearable "lab" on body, where a small device can offer a fully integrated and re-configurable system that mimics several processes usually performed in the laboratory for clinical diagnostics and analysis of human health. To illustrate our vision of having a lab on body, we prototyped "Wearable Lab" a bio-digital platform for sensing biochemical and digital data from saliva. Our platform contains digital sensors such as an IMU for activity recognition, as well as an automated system for continuous sampling of biomarkers from saliva by leveraging existing paper-based biochemical sensors. The platform could aid with longitudinal studies of biomarkers and early diagnosis of diseases. We present example data collected from the device, show a preliminary evaluation, and discuss the limitation of our platform. / by Pat Pataranutaporn. / S.M. / S.M. Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences