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 117-129). / We experience the world through our senses. Our consciousness is the culmination of our recalled perceptual associations. Stimulating the senses can yield remarkable biological effects for cognitive recovery; this is the cornerstone of my thesis. Alzheimer's is currently the 6th leading cause of death in the United States and impacts an estimated 50 million worldwide. Therefore, it is vital to develop an effective cure that goes beyond stalling symptoms, to reverse disease progression. My research contribution to the Aging Brain Initiative, reveals the importance of multi-modal gamma stimulation for human-centered applications. This takes form within a novel sensory taxonomy of the effects of gamma entrainment stimuli in humans. Simultaneously, groundbreaking multisensory testing methodologies are outlined to evaluate subtle cognitive changes in human participants. The amalgamation of this research is physically expressed in the design of a multisensory, interactive Gamma Instrument, designed to effect peak levels of gamma entrainment and sensory congruence. This work provides foundational components for future gamma research and broadens our understanding of the role of music in medicine and multisensory cognition. / by Alexandra Rieger. / 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, 2019 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 82-84). / How different, in a precise sense, is The New York Times from Fox News? Or - Fox from NPR, NPR from CNN, CNN from Breitbart? If we think of news organizations as producers of language, as "speakers" - how similar or different are the voices? This question of the distance between news sources is fundamental to concerns about fragmentation and polarization in the news ecosystem. A number of studies have measured the proximity between outlets in terms of overlap at the level of audience, and then defined content-level differences in terms of the underlying audience composition - for example, the fraction of the readership who have shared content from particular political candidates. The "content graph" of the news ecosystem - the set of similarities and differences at the level of the actual coverage - is often assumed to be tightly linked to the "audience graph"; and the two are even defined in terms of each other. / How exactly do these two systems interact, though? In many ways, our knowledge of the content graph is less precise than knowledge about the audience graph, which in some ways is simpler to measure. A rich line of work has studied the coverage of specific issues in news content, and recent work has started to systematically survey the content produced by a range of outlets, often by way of unsupervised approaches that characterize differences at the level of topic. Building on this, I attempt to precisely quantify the relative similarities among major media organizations from a standpoint of textual discriminabiliy, focusing on a corpus of 1.2 million article headlines from 15 major US news outlets, extracted from an archive of 73 million links posted on Twitter over a 625-day period running from the beginning of 2017 through the summer of 2018. / I formulate the question as a supervised learning problem, in which classifiers are presented with a headline and trained to identify the outlet that produced it. This training objective is used to induce high-quality distributed representations of headlines, and also makes it possible to measure the degree to which different outlets produce similar and dissimilar content. I then contextualize these language-level similarities against two backdrops. First, I examine the degree to which similarities at the level of headlines correlate with similarities at the level of audiences - with specific focus on sites of misalignment, where outlets "speak" in ways that don't match the typical patterns of other outlets that share similar audiences. / Among the news organizations considered in this study, the Associated Press and The Hill are the two most "misaligned" outlets, and we can perhaps look to specific portions of their content as a signal for the types of topics, styles, and stances that might be effective at permeating across axes of political and cultural difference. Second - I study headlines as a historical process. How stable are the linguistic profiles of major news organizations, and to what degree have they evolved into new configurations? I find significant changes over first 18 months of the Trump presidency, with BuzzFeed doubling down on "quiz" articles; Huffington Post moving away from lifestyle content and towards political reporting; The Daily Kos becoming less exclusively focused on politics; and Fox shifting towards a kind of "tabloid" style, with a focus on violent crime, personal misfortune, and socially-charged political issues. / David McClure. / 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, 2018 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 133-141). / Hybrid Living Materials (HLMs) are formed by combining living and non-living materials such that the new material takes on the properties of both. Yet, the integrated control of both material and biological properties and their interactions remains challenging due to the complexity of natural constructs and the lack of standardized infrastructure to control biological function in 3D space. The state of the art remains limited in its range of scale and application scope. We seek to extend the conventions used for the digital design and fabrication of human-made structural materials to harness valuable biotic functionalities-such as sense, response, growth, metabolization, and even adaptation-to enable a diverse new class of multifunctional materials. This thesis establishes a generalizable framework and technical workflow for the creation of HLMs. / The approach integrates (i) computational design, (ii) digital fabrication, and (iii) synthetic biology to generate materials that are programmed to host and template engineered bacterial cells. Specifically, this research employs a multimaterial 3D printing platform in combination with newly developed biochemical signaling resins and volumetric modeling tools to produce additively manufactured objects containing high resolution bioactive diffusion gradients. Computer aided design (CAD) tools allow multiple genetic regulatory signals to be precisely positioned within the cured architectures of these printed shapes. When combined with hydrogel immobilization methods to sustain E. coli and facilitate biosignal transmission on these objects, this approach achieves spatial genetic regulatory control of engineered cells across complex geometries and produces functional hybrid-living materials. / Importantly, the HLM fabrication platform achieves living constructs of up to half a meter in length, and accurate simulation of executable biological patterning and outputs. The platform provides design and fabrication tools that advance the merger between digital manufacturing and biological engineering for a unique level of control and repeatability over spatially-varying material properties and cellular function. Thus, the methodology, hardware, and computational design space established provide an entry point for designers, engineers, and scientists to mediate bacterial functionalities with material technology for broad application. / by Rachel Soo Hoo Smith. / 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, 2019 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 162-180). / Demand for data visualization has exploded in recent years with the increasing availability and use of data across domains. Traditional visualization techniques require users to manually specify visual encodings of data through code or clicks. While manual specification is necessary to create bespoke visualizations, it renders visualization inaccessible to those without technical backgrounds. As a result, visualization recommender systems, which automatically generate results for users to search and select, have gained popularity. Here, I present systems, methods, and data repositories to contextualize and improve visualization recommender systems. The first contribution is DIVE, a publicly available and open source system that combines rule-based recommender systems with manual specification. DIVE integrates state-of-the-art data model inference, visualization, statistical analysis, and storytelling capabilities into a unified workflow. / In a controlled experiment, we show that DIVE significantly improves task performance among a group of 67 professional data scientists. Over 15K users have uploaded 7.5K datasets to DIVE since its release. In response to the limitations of rule-based recommender systems, VizML is a machine learning-based method for visualization recommendation. VizML uses neural networks trained on a large corpus of datasetvisualization pairs to predict visualization design choices, such as visualization type and axis encoding, with an accuracy of over 85%, exceeding that of base rates and baseline models. Benchmarking with a crowdsourced test set, we show that our model achieves human-level performance when predicting consensus visualization type. To support learned visualization systems, VizNet is a large-scale visualization learning and benchmarking repository consisting of over 31M real-world datasets. / To demonstrate VizNet's utility as a platform for conducting crowdsourced experiments with ecologically valid data, we replicate a prior perceptual effectiveness study, and demonstrate how a metric of visualization effectiveness can be learned from experimental results. Our results suggest a promising method for efficiently crowdsourcing the annotations necessary to train and evaluate machine learning-based visualization recommendation at scale. Enabled by the availability of real-world data, Sherlock is a deep learning approach to semantic type detection. We train Sherlock on 686K data columns retrieved from the VizNet corpus by matching 78 semantic types from DBpedia to column headers. We characterize each matched column with 1, 588 features describing the statistical properties, character distributions, word embeddings, and paragraph vectors of column values. / A multi-input neural network achieves a support-weighted F1 score of 0.89, exceeding that of a decision tree baseline, dictionary and regular expression benchmarks, and the consensus of crowdsourced annotations. I conclude by discussing three opportunities for future research. The first describes design considerations for mixed-initiative interactions in AI-infused visualization systems such as DIVE. The second reviews recent work on statistical validity of insights derived from visualization recommenders, which is an especially important consideration with learned systems such as VizML. Lastly, I assess the benefits of learning visualization design from non-experts then present experimental evidence towards measuring the gaps between expert and non-expert judgment. / by Kevin Zeng Hu. / 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, 2019 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 269-284). / Rapidly advancing cognitive technologies, such as artificial intelligence (AI), have the potential to drastically impact modern society and to shape the future of work. Accordingly, policy makers and researchers seek forecasts into technological change and labor trends, including growing job polarization and income inequality as well as decreasing career mobility and spatial mobility for workers. Although a given technology impacts demand for only a narrow set of workplace skills, modern empirical work relies on coarse labor distinctions between cognitive and physical or routine and non-routine work to explain employment trends. In this dissertation, I explore the complex ways that skills and employment undergird aggregate labor dynamics in the US. As a motivating example, I demonstrate how simple measures for skills within a labor market contribute to the differential impact of automation across US cities of different sizes. / I build on this motivation to address methodological barriers through a refined model of workplace skills and their interdependencies, thus connecting microscopic workplace connections to macroscopic labor trends. I perform an unsupervised analysis of specific workplace skills as a skills network whose aggregate and refined topology grant new insights into job polarization and workers' career mobility. Since these inter-skill connections predict career mobility, I construct a map of US occupations that captures worker transition rates between employment opportunities and, in combination with urban employment data, predicts workers' spatial mobility. These refined models that connect workplace skills to both inter-city and intra-city dynamics enable new insights and new input data sources for real-time labor trends at the level of specific technologies and specific workplace skills. / I conclude by exploring one novel and potentially useful source of input information: the evolution of scientific Al research. The analyses in this dissertation provide new tools to policy makers designing viable worker retraining programs, offer new insights to individual workers navigating their careers, and present new measures for economic resilience in the face of changing technology. / by Morgan Ryan Frank. / 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, 2019 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 91-105). / The recent harnessing of microbial adaptive immune systems, known as CRISPR, has enabled genome-wide engineering across all domains of life. A new generation of gene-editing tools has been fashioned from the natural DNA/RNA-targeting ability of certain CRISPR-associated (Cas) proteins and their guide RNA, which work together to recognize and defend against infectious genetic threats. This straight-forward RNA-programmed sequence recognition by CRISPR has facilitated its rapid global impact on genetic research, diagnostics, therapeutics, and bioproduction. An ideal DNA-editing platform would achieve perfect accuracy on any desired cellular and genomic target. CRISPR systems, however, have limited target fidelity and range, in part due to their evolutionary pressures to defend microbes from fast-mutating viruses without self-targeting their own guide RNA. / These natural limitations of CRISPR can especially constrain gene-editing in animals and plants, which are more vulnerable to off-target activity occurring in one of their trillions of cells with genomes that are 1000x larger than those of unicellular microbes that natively harbor CRISPR systems. This thesis overcomes three critical challenges for precise and broad gene-editing of complex organisms: 1) engineering a means of specificity for the type of cells to edit, 2) improving target-matching accuracy, and 3) broadening the editable portion of the genome. / This thesis addresses these challenges by integrating custom developed computational design tools and biological validation of the resulting novel CRISPR systems; 1) To target within multicellular heterogeneity, new oligonucleotide-sensing structural motifs are designed and embed into guides that can potentially control CRISPR nuclease activity based on cell-type transcriptome patterns; 2) To discern among increased similarity between a target and non-target sequences in larger genomes, base-pairing thermostability principles are employed to tune the biochemical composition of guides that can evade subtly mismatched off-target sites; 3) To expand the reach of editing techniques with narrow windows of operation, such as base-editing, bioinformatics workflows that discover previously uncharacterized Cas proteins with novel target scope are created. / This thesis demonstrates the effectiveness of these strategies in the context of in vitro, bacterial, and human cell culture assays, and contributes advancements in the precision and generality for CRISPR gene-editing. / by Noah Michael Jakimo. / 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, 2019 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 266-294). / Children are now growing up with Al-enabled, socially interactive technology. As such, we need to deeply understand how children perceive, interact, and relate to this kind of technology, especially given the many ethical concerns that arise in the context of human-machine interactions, most of which are most contentious with children. To this end, I explore questions about young children's interactions and relationships with one such technology--social robots-during language learning activities. Language learning is a ripe area for exploring these questions because of the social, interactive, interpersonal nature of the activity. In addition, literacy, language, and interpersonal skills are some of the most important skills any child will learn, as they can greatly impact children's later educational and life success. / Through a series of 9 empirical child-robot interaction studies with 347 children and using both teleoperated and autonomous robots, I establish the role of social robots as relational technology-that is, technology that can build long-term, social-emotional relationships with users. I hypothesize that a key aspect of why social robots can benefit children's learning is their social and relational nature. To that end, I demonstrate the capabilities of social robots as learning companions for young children that afford opportunities for social engagement and reciprocal interaction, particularly peer-to-peer mirroring. I discuss how we can understand children's conceptualizations of social robots as relational agents and measure children's relationships over time. I introduce the term relational AI to refer to autonomous relational technologies. / I develop a computational relational Al system to examine how using relational Al in a social robot can impact child-robot learning interactions. Through testing the autonomous system in a longitudinal study with 49 children, I explore connections between children's relationship and rapport with the robot and their engagement and learning. I discuss the ethical use and design implications of relational AL. I show that relational AI is a new, powerful educational tool, unlike any other existing technology, that we can leverage to support children's early education and development. / "Supported by a MIT Media Lab Learning Innovation Fellowship, and by the National Science Foundation (NSF) under Grants CCF-1 3 89 86, IIS-1122886, IIS-11228 4 5, IIS-112308 5 , IIS-1523118, and Graduate Research Fellowship Grant No. 1122374"--Page 6 / by Jacqueline M. Kory-Westlund. / 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, 2019 / Cataloged from PDF version of thesis. "June 2019." Vita. / Includes bibliographical references (pages 118-122). / Programmable Synthetic Hallucinations describe the utilization of the bio-physiological mechanics of hallucination generated in the human brain to display virtual information directly in the visual field. Science fiction films, television shows, and video games have trained audiences to think of holograms as luminous volumetric images that float registered in the viewer's 3D space and require no special glasses or optics to see or interact with them. The ability of users to interact with a floating aerial lightfield without the use of face-worn binocular optics is a difficult challenge and one in which a hallucinatory experience offers a solution. While we do not have the ability to activate individual neurons to recreate an neuro-electrical pattern indiscernible from the perception of reality, this dissertation shows that creating phosphenes within the visual field via the magnetic stimulation of neurons in the visual cortex is a viable first step. / By electrically stimulating the cells in the hypercolumns of V1, one can induce the perception of a pixel of light within the visual field of a user. These magnetophosphenes are visual perceptions described as luminous shapes, which can be created by time-varying magnetic fields. These change the membrane potential and trigger an action potential directly in neurons of the visual cortex. Previous TMS studies have shown evocation of phosphenes in a binary manner, with subjects reporting the presence or absence of a phosphene but not targeted to a specific location. However, to date, no information or example has been found indicating the use of cortical phosphenes, induced magnetically or otherwise, in performance or public display. Presently, commercial transcranial magnetic stimulators can only be focused to an area approaching one square centimeter, a single output channel, and require manual placement of the coil apparatus. / Novel coil designs became a central focus of this research. Further work increased the number of output channels, embedding them in a wearable apparatus with a multichannel array of induction coils. Clinical trials were undertaken at MIT's Clinical Research Center. We were able to evoke visual phenomena in 11 out of 16 test subjects in a known, targeted location. The induced magnetophosphenes were noted above the noise floor of naturally occurring retinal phosphenes and were statistically verified to be a result of the system being tested. / by Daniel Edward Novy. / 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, 2019 / Cataloged from PDF version of thesis. / Includes bibliographical references (pages 136-139). / Between traditional industrial design, which operates at the macro scale (cm to m), and material engineering, which operates at the micro/nano scale ([mu]m to nm), is the emerging design space of the mesoscale. While the definition of mesoscale varies across disciplines, mesoscale materials are usually considered to be in between the molecular and macroscopic length scale. It is the scale of human hair or a grain of sand. It is the scale where material properties meet human perception, and the rational meets intuition. In the past 10 years, additive manufacturing, especially 3D printing, has enabled designers to directly manipulate geometries at this scale. Yet existing design and manufacturing approach have not been able to unleash the full potential of mesoscale materials for the design world. This thesis proposes computational tools and an additive manufacturing apparatus to enable the creation and fabrication of materials at the mesoscale. / The ability to programmably assemble materials with tailored structures at the centimeter, millimeter, and micrometer length scales enables tunable mechanical and electrical properties. Those properties determine not only the static performance, but also, when energized, the dynamic behavior of a material. The emerging material performance and behavior allows us to design unprecedented objects and environments with input (sensing) and output (actuation) capabilities, which can be integrated for the next generation of interaction design. I first introduces three translations to bridge a material's microscopic properties with macroscopic interface design. Four research projects (bioLogic, KinetiX, SensorKnit, and Cilllia) are presented to embody the translation. I then propose an implementation workflow for additive manufacturing of mesoscopic materials. The implementation will be presented based on my ongoing research project Cilllia, 3D printed functional hair structures. / Cilllia investigates a scalable digital representation of hierarchical tunable materials, a CAD software interface for material design, and a DLP-based 3D printer that allows for continuous material production. The tools for creating Cilllia can be expanded to other types of architected mesoscale materials. Four examples will be presented. Together, they support the vision of a general digital description and physical production system for architected mesoscale materials. / by Jifei Ou. / 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 version of thesis. / Includes bibliographical references (pages [95]-104). / If Waste is Information, what can be learned from an understanding of waste management at MIT as a complex information system? Through engaging people with their role within a waste system, can relations to and practices of waste management change? This thesis presents a case study for making and evaluating context-driven and critical civic games in partnership with local organisations. Working with waste management and sustainability efforts on campus, I explore issues with waste at MIT as a 'crisis of representation', drawing on ideas from systems theory, participatiory urbanism, environmental psychology and game design to develop a new 'systems image' of waste on campus. Through two controlled studies, I examine the potentials and limitations of such work for changing attitudes and behaviours, and the pitfalls of attempting to separate educational interventions from their infrastructural context. More generally, this project contributes to an understanding of how we might use participatory, critical and exploratory games to make legible complex civic systems, and the role of that legibility in changing both individuals, and the systems themselves. / by Agnes Fury Cameron. / S.M. / S.M. Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences