483 research outputs found
Utility-Based Abstraction and Categorization
No full textion and Categorization Eric J. Horvitz # and Adrian C. Klein Palo Alto Laboratory Rockwell International Science Center 444 High Street Palo Alto, CA 94301 Abstract We take a utility-based approach to categorization. We construct generalizations about events and actions by considering losses associated with failing to distinguish among detailed distinctions in a decision model. The utility-based methods transform detailed states of the world into more abstract categories comprised of disjunctions of the states. We show how we can cluster distinctions into groups of distinctions at progressively higher levels of abstraction, and describe rules for decision making with the abstractions. The techniques introduce a utility-based perspective on the nature of concepts, and provide a means of simplifying decision models used in automated reasoning systems. We demonstrate the techniques by describing the capabilities and output of TUBA, a program for utility-based abstraction. 1 INTRODUCTION..
Members of the miRNA-200 Family Regulate Olfactory Neurogenesis
No full textMicroRNAs (miRNAs) are highly expressed in vertebrate neural tissues, but the contribution of specific miRNAs to the development and function of different neuronal populations is still largely unknown. We report that miRNAs are required for terminal differentiation of olfactory precursors in both mouse and zebrafish but are dispensable for proper function of mature olfactory neurons. The repertoire of miRNAs expressed in olfactory tissues contains over 100 distinct miRNAs. A subset, including the miR-200 family, shows high olfactory enrichment and expression patterns consistent with a role during olfactory neurogenesis. Loss of function of the miR-200 family phenocopies the terminal differentiation defect observed in absence of all miRNA activity in olfactory progenitors. Our data support the notion that vertebrate tissue differentiation is controlled by conserved subsets of organ-specific miRNAs in both mouse and zebrafish and provide insights into control mechanisms underlying olfactory differentiation in vertebrates.National Institutes of Health. (U.S.) (grant NS049319)National Institutes of Health. (U.S.) (grant GM56211)Wellcome Trust (London, England) (grant 066790/B/02/Z
Turn-Taking and Coordination in Human-Machine Interaction
No full textAndrist S, Bohus D, Horvitz E, Mutlu B, Schlangen D, eds. Turn-Taking and Coordination in Human-Machine Interaction. AAAI Spring Symposium. 2015
Principles and applications of continual computation
No full textAbstractAutomated problem solving is viewed typically as the allocation of computational resources to solve one or more problems passed to a reasoning system. In response to each problem received, effort is applied in real time to generate a solution and problem solving ends when a solution is rendered. We examine continual computation, reasoning policies that capture a broader conception of problem by considering the proactive allocation of computational resources to potential future challenges. We explore policies for allocating idle time for several settings and present applications that highlight opportunities for harnessing continual computation in real-world tasks
Predictive Approaches for Acute Adverse Events in Electronic Health Records
No full textThesis (Ph.D.)--University of Washington, 2019Medical errors have been cited as the third leading cause of death in the United States in 2013. Failure to rescue (FTR) is a subtype of medical errors and refers to the loss of an opportunity to save a patient’s life after the development of one or more preventable and treatable complications. Focusing on detecting early signs of deterioration may therefore provide opportunities to prevent and/or treat an illness in a timely manner, which may in turn reduce the number of FTR cases. When implementing a data-driven model to predict the risk of potential FTR onsets in a supervised setting, gold standard information for the target FTR onset is often not directly retrievable in electronic health records (EHR) so that it requires to manually annotate clinical observations with corresponding labels. This method results in a bottleneck to scalability and the full utilization of the clinical observations available in EHRs for model training. In this dissertation, I propose a machine learning framework that can be used to derive a risk prediction model using proxy events of the disease of interest, the administration of relevant clinical interventions, as a noisy label via a distant supervision approach. Moreover, this study evaluated the effects of considering the temporal progression of FTR risk estimates calculated using myopic evidence. Lastly, a case study is presented to demonstrate that the proposed prediction models can be deployed to quantify the adverse effects of clinical interventions with regard to the target disease of interest. This dissertation demonstrates 1) the feasibility of using proxy events of the target disease as a label for supervised model training, 2) the performance improvement when temporal progression is considered in the risk prediction model design, and 3) the applicability of the proposed risk prediction model to quantify the adverse effects of clinical interventions regarding the target disease. Suggestions are also provided on how the proposed model could be further improved by integrating experts’ knowledge with the proposed framework
Recommended from our members
Automated Workflow Synthesis
No full textBy coordinating efforts from humans and machines, human computation systems can solve problems that machines cannot tackle alone. A general challenge is to design efficient human computation algorithms or workflows with which to coordinate the work of the crowd. We introduce a method for automated workflow synthesis aimed at ideally harnessing human efforts by learning about the crowd's performance on tasks and synthesizing an optimal workflow for solving a problem. We present experimental results for human sorting tasks, which demonstrate both the benefit of understanding and optimizing the structure of workflows based on observations. Results also demonstrate the benefits of using value of information to guide experiments for identifying efficient workflows with fewer experiments.Engineering and Applied SciencesAccepted Manuscrip
Axons Degenerate in the Absence of Mitochondria in C. elegans
No full textMany neurodegenerative disorders are associated with mitochondrial defects [1, 2 and 3]. Mitochondria can play an active role in degeneration by releasing reactive oxygen species and apoptotic factors [4, 5, 6 and 7]. Alternatively, mitochondria can protect axons from stress and insults, for example by buffering calcium [8]. Recent studies manipulating mitochondria lend support to both of these models [9, 10, 11, 12 and 13]. Here, we identify a C. elegans mutant, ric-7, in which mitochondria are unable to exit the neuron cell bodies, similar to the kinesin-1/unc-116 mutant. When axons lacking mitochondria are cut with a laser, they rapidly degenerate. Some neurons even spontaneously degenerate in ric-7 mutants. Degeneration can be suppressed by forcing mitochondria into the axons of the mutants. The protective effect of mitochondria is also observed in the wild-type: a majority of axon fragments containing a mitochondrion survive axotomy, whereas those lacking mitochondria degenerate. Thus, mitochondria are not required for axon degeneration and serve a protective role in C. elegans axons
Machine Learning, Reasoning, and Intelligence in Daily Life: Directions and Challenges
No full textTechnical developments and trends are providing a fertile substrate for creating and integrating machine learning and reasoning into multiple applications and services. I will review several illustrative research efforts on our team, and focus on challenges, opportunities, and directions with the streaming of machine intelligence into daily life. 1 Reflections on Trends and Directions Over the last decade, technical and infrastructural developments have come together to create a nurturing environment for developing and fielding applications of machine learning and reasoning—and for harnessing automated intelligence to provide value to people in the course of their daily lives. These developments include (1) technical advancements in machine learning and reasoning, (2) the growth in CPU and memory capabilities within commonly available devices and platforms, (3) the connectivity, content, and services provided by the evolving Web, and (4) the increasing availability of data resources, including corpora of behavioral data collected via inexpensive sensors, and through ongoing interaction with software and services. 1.1 Panoply of Applications and Services Opportunities for integrating applications of machine intelligence into the daily lives of people are growing with the increasing popularity of computing systems, the widening diversity of web services, the growing popularity of portable devices that contain general-purpose operating systems, and ongoing innovations in human-computer interaction— including the increasing prowess of speech recognition, handwriting, and sketch-understanding interfaces. Various examples of the integration of automated learning and reasoning into daily life have been appearing as implicit and explicit extensions to traditional systems and services, and also in prototypes and systems that provide qualitatively new kinds of experiences. I will review several projects and efforts undertaken by our team that highlight directions and approaches to introducing potentially valuable machine learning and reasoning into the daily lives of people
Busybody: creating and fielding personalized models of the cost of interruption
No full textInterest has been growing in opportunities to build and deploy statistical models that can infer a computer user’s current interruptability from computer activity and relevant contextual information. We describe a system that intermittently asks users to assess their perceived interruptability during a training phase and that builds decision-theoretic models with the ability to predict the cost of interrupting the user. The models are used at run-time to compute the expected cost of interruptions, providing a mediator for incoming notifications, based on a consideration of a user’s current and recent history of computer activity, meeting status, location, time of day, and whether a conversation is detected
- …
