Thesis (M.S.)--Massachusetts Institute of Technology, Dept. of Architecture, 1982. / MICROFICHE COPY AVAILABLE IN ARCHIVES AND ROTCH. / Includes bibliographical references (leaves 109-112). / Today's architectural knowledge base is a heterogeneous mixture of empirical wisdom generated in numberless ways by thousands of practitioners through centuries' of trial-and-error. Architecture has never developed a professionally recognized system or research, nor any organized method for continuously developing the body of commonly held knowledge. Traditionally, this lack or organization has not hindered the profession's ability to successfully integrate available means to meet the desired ends. The past decades, however, have produced a myriad of new techniques, materials, building products and. processes, and a similar increase in societal demands. The abundance of knowledge necessary for successful intuitive design has overwhelmed the traditional methods of assimilation. The lack of an organized knowledge base and the means of producing such a base threaten the ability of the profession to fulfill its traditional responsibilities. Research entails the organized production of knowledge, and for over twenty five years, architecture has been struggling to develop an effective form of research. Initial attempts, however, have generally not had any tangible results; borrowed almost directly from science, investigative methods have not been tailored to the needs or nature of architecture. Architectural design is not a pure science; the scientific method, unadapted, in not an appropriate method of inquiry. Traditional methods of architectural inquiry can be updated and combined with appropriately modified scientific attitudes to form an effective system of architectural research. Initially, the relationship between organized inquiry and design must be clarified, and the work of the former directed towards the needs of the latter. To effect the integration between analysis and synthesis, a new type of professional must emerge, the facilitator. With a background in both design and research, the facilitator will be able to fill a variety of roles within the profession, always working to develop a systematic and effective way of adding new information to the knowledge base and bringing this knowledge into professional practice. Intuition is still the foundation of successful design, but intuition can spring only from knowledge. Organized research is a means of providing such knowledge, but only if research is appropriate to both the subject and the user. / by James A. Moore. / M.S.

Form is the manifestation of a process. Since our senses are our only means of perception, then a process to be manifest must necessarily assume a state that our senses can perceive. This state or expression which we can perceive is its form. “Every process has its necessary form which always results in functional forms. They follow the law of the shortest distance between points; cooling occurs only on surfaces exposed to cooling; pressure only on points of pressure; tension on lines of tension; motion creates for itself forms of movement for each energy there is a form of energy.” The functional form of anything is the way in which it acts or functions. Everything has many functional forms and these are perpetually changing. They are constantly being influenced to manifest themselves in different ways by processes which are in perpetual change; and these in turn are influenced by other processes in perpetual change. Functional form at any time is the form in which processes manifest themselves and is the result of all of the changes that have occurred up to that time. / Master of Science

Thesis (M.Arch.)--University of Hong Kong, 1997. / Includes special report study entitled : [1] Flexibility.--[2] Photovoltaics. Includes bibliographical references. Also available in print.

Dynamic hierarchical networks represent an architectural strategy for employing adaptive behavior in applications sensitive to highly variable external demands or uncertain internal conditions. The characteristics of such architectures are described, and the significance of adaptive capability is discussed. The necessity for assessing cost/benefit tradeoffs leads to the use of queueing network models. The general model, a network of M/M/1 queues in a random environment, is introduced and then is simplified so that the links may be treated as isolated M/M/1 queues in a random environment. This treatment yields a formula for approximate mean network delay by combining matrix-geometric results (mean queue length and mean delay) for the individual links. Conditions under which the analytic model is considered valid are identified through comparison with a discrete event simulation model. Last, performance of the dynamic hierarchy is compared with that of the static hierarchy. This comparison establishes conditions for which the dynamic architecture enables performance equal or nearly equal to performance of the static architecture. / Ph. D. / incomplete_metadata

We asked whether the design of a Science Data Repository (SDR) can influence data sharing behaviour in small scientific collaborations. We hypothesised that an SDR can influence data-sharing behaviour when its design considers the context of data-sharing. We proposed an alternative approach to those documented in the literature, employing a combination of socio-technical empirical and analytical methods for context capturing, and choice architecture for context incorporation. To evaluate the approach we applied it to design features in a Scientific Data Repository for a population of small scientific collaborations within the Life Sciences. The application of this thesis' approach consisted of an exploratory case study, a review of factors associated with data sharing, the definition of design claims, and implementation of a set of design features. We collected data using interviews with members of the collaborations and designers of the SDR; as well as obtaining the data-logs from the collaborations' SDR. We evaluated the resulting design features using an asynchronous web experiment. We found that using the empirical approach to context capturing we are able to effectively identify factors associated with data sharing in the small scientific collaborations. Moreover, we identified a number of limitations on the application of the analytical approach to context capturing. Furthermore, we found that the Choice Architecture based procedure for context incorporation can define effective design features in Science Data Repositories. In this work, we show that we can facilitate data-sharing by incorporating context into the design of a Science Data Repository, and identified a set of restrictions to use our approach. The approach proposed in this thesis can be used by practitioners wishing to improve data sharing in an SDR. Contributions, such as the survey of factors associated with data sharing behaviour, can be used by researchers to understand the problems associated with data sharing in small scientific collaborations.

<p dir="ltr">Advances in machine learning (ML) have ignited hardware innovations for efficient execution of the ML models many of which are memory-bound (e.g., long short-term memories, multi-level perceptrons, and recurrent neural networks). Specifically, inference using these ML models with small batches, as would be the case at the Cloud edge, has little reuse of the large filters and is deeply memory-bound. Simultaneously, processing-in or -near memory (PIM or PNM) is promising unprecedented highbandwidth connection between compute and memory. Fortunately, the memory-bound ML models are a good fit for PIM. We focus on digital PIM which provides higher bandwidth than PNM and does not incur the reliability issues of analog PIM. Previous PIM and PNM approaches advocate full processor cores which do not conform to PIM’s severe area and power constraints. This thesis is composed of three major projects: Newton, activation folding (AcF) and ESPIM. Newton is Sk Hynix’s first accelerator-inmemory (AiMX) product for machine learning, AcF improves the performance of Newton by achieving more compute-row access overlap and ESPIM incorporate sparse neural network models to PIM</p>

Indiana University-Purdue University Indianapolis (IUPUI) / Domain-specific models powered using domain-specific modeling languages are traditionally created manually by modelers. There exist model intelligence techniques, such as constraint solvers and model guidance, which alleviate challenges associated with manually creating models, however parts of the modeling process are still manual. Moreover, state-of-the-art model intelligence techniques are---in essence---reactive (i.e., invoked by the modeler). This thesis therefore provides two contributions to model-driven engineering research using domain-specific modeling language (DSML). First, it discusses how DSML semantic and constraint can enable proactive modeling, which is a form of model intelligence that foresees model transformations, automatically executes these model transformations, and prompts the modeler for assistance when necessary. Secondly, this thesis shows how we integrated proactive modeling into the Generic Modeling environment (GME). Our experience using proactive modeling shows that it can reduce modeling effort by both automatically generating required model elements, and by guiding modelers to select what actions should be executed on the model.