Thesis (S.M.)--Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2006. / Includes bibliographical references (p. 103-106). / The need for more effective communication across different countries has increased as the interactions between them have been growing. Communication is often difficult because of both language differences and cultural differences. Although there have been many attempts to meet the communication need on the level of language with machine translators and dictionaries, many problems related to cultural and conceptual differences still remain. To improve traditional machine translators and cross-cultural communication aids, it is necessary to develop automated mechanisms to analyze cultural differences and similarities. This thesis approaches the problems with automatic computation of cultural differences and similarities. This thesis, GlobalMind, provides common-sense databases of various countries and languages and two inference modules to analyze and compute the cultural differences and similarities from the databases. I describe the design of GlobalMind databases, the implementation of its inference modules, the results of an evaluation of GlobalMind, and available applications. / by Hyemin Chung. / S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2003. / Includes bibliographical references (p. 79-82). / We present a computational model that connects gestalt visual perception and language. The model grounds the meaning of natural language words and phrases in terms of the perceptual properties of visually salient groups. We focus on the semantics of a class of words that we call conceptual aggregates e.g., pair, group, stuff, which inherently refer to groups of objects. The model provides an explanation for how the semantics of these natural language terms interact with gestalt processes in order to connect referring expressions to visual groups. Our computational model can be divided into two stages. The first stage performs grouping on visual scenes. It takes a visual scene segmented into block objects as input, and creates a space of possible salient groups arising from the scene. This stage also assigns a saliency score to each group. In the second stage, visual grounding, the space of salient groups, which is the output of the previous stage, is taken as input along with a linguistic scene description. The visual grounding stage comes up with the best match between a linguistic description and a set of objects. Parameters of the model are trained on the basis of observed data from a linguistic description and visual selection task. The proposed model has been implemented in the form of a program that takes as input a synthetic visual scene and linguistic description, and as output identifies likely groups of objects within the scene that correspond to the description. We present an evaluation of the performance of the model on a visual referent identification task. This model may be applied in natural language understanding and generation systems that utilize visual context such as scene description systems for the visually impaired and functionally illiterate. / by Sheel Sanjay Dhande. / S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2012. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 95-100). / This thesis applies network theory to firms, their employees, and various aspects of the employees to understand diversity within an industry at both the firm-level and employee-level. We hypothesize that the interest diversity of a firms' employee can influence that firm's economic performance and growth. Using the LinkedIn API, we are able to collect ~600 employees (past and present) for 43 companies using a keyword search. This data is used to create a visualization of people's interests, an "Interest Space", which is a network graph of how interests are categorized and linked to each other. By analyzing the data from firms associated with the Media Lab, we begin to understand how individual interests affect success among a small network of companies. We research this through case studies of a few companies by analyzing their financial data and interests of their employees. / by Julia Shuhong Ma. / S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2004. / Includes bibliographical references (p. 101-104). / (cont.) theoretical section is accompanied by descriptions of historic and contemporary works that have influenced IMPI. / Shaping collective free improvisations in order to obtain solid and succinct works with surprising and synchronized events is not an easy task. This thesis is a proposal towards that goal. It presents the theoretical, philosophical and technical framework of the Improvisatory Music and Painting Interface (IMPI) system: a new computer program for the creation of audiovisual improvisations performed in real time by ensembles of acoustic musicians. The coordination of these improvisations is obtained using a graphical language. This language is employed by one "conductor" in order to generate musical scores and abstract visual animations in real time. Doodling on a digital tablet following the syntax of the language allows both the creation of musical material with different levels of improvisatory participation from the ensemble and also the manipulation of the projected graphics in coordination with the music. The generated musical information is displayed in several formats on multiple computer screens that members of the ensemble play from. The digital graphics are also projected on a screen to be seen by an audience. This system is intended for a non-tonal, non-rhythmic, and texture-oriented musical style, which means that strong emphasis is put on the control of timbral qualities and continuum transitions. One of the main goals of the system is the translation of planned compositional elements (such as precise structure and synchronization between instruments) into the improvisatory domain. The graphics that IMPI generates are organic, fluid, vivid, dynamic, and unified with the music. The concept of controlled improvisation as well as the paradigm of the relationships between acoustic and visual material are both analyzed from an aesthetic point of view. The / Hugo Solís García. / S.M.

Thesis (Ph.D.)--Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2000. / Vita. / Includes bibliographical references (p. 201-216). / The tendency of block copolymers (BCP's) to microphase separate at the molecular level, producing a wide array of ordered nanostructures, is of particular interest from an engineering standpoint due to the unique mechanical, optical or electrical properties that ensue. Upon considering the potential applications of these materials, however, one limitation arises from the lack of control over bulk thermodynamics and the appearance of order/disorder (solid-like/liquid-like) transitions in these materials. To address this problem, this thesis aims to, firstly, develop a more quantifiable understanding of the molecular factors governing BCP phase behavior, and, secondly, use that knowledge to molecularly engineer new BCP's with enhanced processibility. While most BCP's microphase separate upon cooling through an upper disorder-to-order transition (UDOT), polystyrene-block-poly n-butyl methacrylate, PS-b-PBMA, undergoes ordering upon heating through a lower disorder-to-order transition (LDOT). Preliminary studies on this material revealed a unique pressure sensitivity of this ordering transition. By applying pressure, this material could be forced into the segmentally mixed liquid state, implying "baroplasticity", a highly attractive property from a processing standpoint. To better understand the molecular origin of this behavior, the bulk thermodynamics of a family of BCPs formed from styrene and a homologous series of n-alkyl methacrylates (PS-b-PnAMA, n ranging from 1 to 12) was investigated, both as a function of pressure and temperature. The results of this study reveal an unexpected, though systematic, dependence of the phase behavior of these BCP's on monomer architecture. In short, over a certain range of alkyl side chain length, PS-b-PnAMA block copolymers are marginally compatible and exhibit unexpectedly large pressure coefficients for the ordering transition, ranging from 60 to 150°C/kbar. In an attempt to identify molecular parameters responsible for these thermodynamic trends, as well as those displayed by other systems reported in the literature, combined group contribution/lattice fluid model calculations of the cohesive properties of the corresponding homopolymers are performed. Based on this analysis, the homopolymer mass density is proposed as a macroscopic parameter that appears to govern phase behavior in weakly interacting block copolymers or polymer blends. Using this new criterion, a simple tool for the molecular design of phase behavior into weakly interacting BCP's is identified, which is successfully used to engineer "baroplastic" behavior into several new systems of commercial relevance, including elastomers and adhesives based on styrene and low Tg acrylates. In light of the improved understanding of BCP phase behavior emerging from these studies, a simple phenomenological free energy expression is proposed for compressible polymer mixtures, that can be extended to block copolymers. Its ability to predict qualitative phase diagrams for the systems investigated in this thesis as well as many other polymer pairs is demonstrated. Using this expression, basic principles regarding polymer thermodynamics are outlined. / by Anne-Valérie G. Ruzette. / Ph.D.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Program in Media Arts & Sciences, 1995. / Includes bibliographical references (p. 135-147) and index. / Louis Murray Weitzman. / Ph.D.

Thesis (S.M.)--Massachusetts Institute of Technology, Program in Media Arts & Sciences, 1999. / Includes bibliographical references (p. 97-99). / The digital revolution which has changed so many other aspects of modem life has yet to profoundly affect the working process of visual artists and designers. High-quality digital design tools exist, but they provide the user with an improved traditional design process, not a radically new way of designing. Conventional digital design tools are useful, but when design software emulates a paintbrush or photostudio many powerful possibilities of the computational medium are overlooked. This thesis explores emergent design, a design methodology based on a new process, enhanced interactive genetic programming. The emergent design methodology and tools allow designers to effectively create procedural design solutions (design solutions that take the form of a procedure or program) in a way that requires little or no programming on the part of the designer. The use of preliminary fitness functions in the interactive genetic programming process allows the designer to specify heuristics to guide the search and manage the complexity of the interactive genetic programming task. This document is structured in the form of a case study, in which the enhanced genetic programming process and emergent design methodology are described through their application to the specific problem of developing procedural image filters for still and moving images. Two interactive genetic programming systems for image filter evolution are described, GPI and evolution++, along with the Sol programming language that was used to create them. Results from the implementation and use of GPI and evolution++ are presented, including a number of filtered images and image sequences. These results suggest that fitness-agent enhanced interactive genetic programming and the emergent design methodology may play a useful role in the visual design process, allowing designers to explore a wider range of options with greater ease than is possible through a traditional, procedural, or conventional genetic programming design process. / Richard W. DeVaul. / S.M.

Thesis (Ph. D.)--Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2005. / Leaf 141 blank. / Includes bibliographical references (leaves 135-140). / People collaborating in groups have potential to produce higher-quality output than individuals working alone, due to the pooling of resources, information, and skills. Yet social psychologists have determined that groups rarely harness this potential. This thesis proposes that technology in face-to-face settings can be used to address the social factors that have damaging influence on group decision-making processes. While there is much work in the area of collaborative software and groupware, this work differentiates itself with its specific aim to influence the way a group shares information without mediating the group's communication. By presenting visualizations to the group of individual levels of participation and turn-taking behavior, the technology aims to augment the group's communication ability, by making it more aware of imbalances. A series of dynamic displays positioned peripherally to a discussion were developed and used by a variety of groups during face-to-face meetings. Both observational and experimental results indicate that these displays influence individual participation levels and the process of information sharing used during a decision-making discussion. A display revealing real-time participation levels caused those at the highest levels of participation to decrease the amount they spoke. Viewing a visualization of previous turn-taking patterns caused those who spoke the least to increase the amount they spoke in a subsequent discussion; real-time feedback did not produce this change. Additionally, after reviewing their turn-taking patterns, groups altered their information-sharing strategies. / (cont.) For groups that had poor sharing strategies on an initial task, this change improved their ability to share information related to the decision; for those who did not need intervention, feedback on turn-taking was not beneficial for their subsequent information sharing. The central finding of this research is that displays of social information, viewed during or after a meeting, bring about changes in a group's communication style, highlighting the potential for such displays to improve real-world decision-making. / by Joan Morris DiMicco. / Ph.D.

Thesis (Ph.D.)--Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2000. / Includes bibliographical references (leaves 135-140). / Tomorrow's ubiquitous computing environments will go beyond the keyboard, mouse and monitor paradigm of interaction and will require the automatic interpretation of human motion using a variety of sensors including video cameras. I present several techniques for human motion recognition that are inspired by observations on human gesture, the class of communicative human movement. Typically, gesture recognition systems are unable to handle systematic variation in the input signal, and so are too brittle to be applied successfully in many real-world situations. To address this problem, I present modeling and recognition techniques to adapt gesture models to the situation at hand. A number of systems and frameworks that use adaptive gesture models are presented. First, the parametric hidden Markov model (PHMM) addresses the representation and recognition of gesture families, to extract how a gesture is executed. Second, strong temporal models drawn from natural gesture theory are exploited to segment two kinds of natural gestures from video sequences. Third, a realtime computer vision system learns gesture models online from time-varying context. Fourth, a realtime computer vision system employs hybrid Bayesian networks to unify and extend the previous approaches, as well as point the way for future work. / by Andrew David Wilson. / Ph.D.

Thesis (S.M.)--Massachusetts Institute of Technology, School of Architecture and Planning, Program in Media Arts and Sciences, 2011. / Cataloged from PDF version of thesis. / Includes bibliographical references (p. 93-98). / Computers as we know them are fading into the background. However, interaction modalities developed for "foreground" computer systems are beginning to seep into the day-to-day interactions that people have with each other and the objects around them. The Prototouch system allows user interface designers to quickly prototype systems of ubiquitous computing devices which utilize touch and gesture based interactions popularized by the recent explosion of multitouch-enabled smartphones, enabling the user to act as container, token, and tool for the manipulation and transportation of abstract digital information between these devices. Two versions of the system utilizing different network topologies have been created, and several example applications utilizing the system have been developed. / by David Cranor. / S.M.