Search

Your search keyword '"Huggins M"' showing total 393 results
Start Over Author "Huggins M"
393 results on '"Huggins M"'

4. Practical guidelines for intent recognition: BERT with minimal training data evaluated in real-world HRI application

Author

Massachusetts Institute of Technology. Media Laboratory, Huggins, M, Alghowinem, S, Jeong, S, Colon-Hernandez, P, Breazeal, C, Park, HW, Massachusetts Institute of Technology. Media Laboratory, Huggins, M, Alghowinem, S, Jeong, S, Colon-Hernandez, P, Breazeal, C, and Park, HW

Abstract

Intent recognition models, which match a written or spoken input's class in order to guide an interaction, are an essential part of modern voice user interfaces, chatbots, and social robots. However, getting enough data to train these models can be very expensive and challenging, especially when designing novel applications such as real-world human-robot interactions. In this work, wefi rst investigate how much training data is needed for high performance in an intent classification task. We train and evaluate BiLSTM and BERT models on various subsets of the ATIS and Snips datasets. Wefi nd that only 25 training examples per intent are required for our BERT model to achieve 94% intent accuracy compared to 98% with the entire datasets, challenging the belief that large amounts of labeled data are required for high performance in intent recognition. We apply this knowledge to train models for a real-world HRI application, character strength recognition during a positive psychology interaction with a social robot, and evaluate against the Character Strength dataset collected in our previous HRI study. Our real-world HRI application results also confirm that our model can produce 76% intent accuracy with 25 examples per intent compared to 80% with 100 examples. In a real-world scenario, the difference is only one additional error per 25 classifications. Finally, we investigate the limitations of our minimal data models and offer suggestions on developing high quality datasets. We conclude with practical guidelines for training BERT intent recognition models with minimal training data and make our code and evaluation framework available for others to replicate our results and easily develop models for their own applications.

Published
2021

19. CHAPTER REPORTS

Author

Hagan, W. A., Hill, Albert F., Buchta, J. W., Andersen, E. V., Ott, P. W., Huggins, M. L., Hanke, Martin E., Bevan, Arthur, Randall, D. P., Brink, R. A., Gay, H. J., Edington, Will E., Wichers, Edward, Osterberg, A. E., Couch, John N., Park, O. W., Keys, D. A., States, M. N., Cady, L. C., Creighton, H. Jermain, Roberts, Joseph K., Pullen, M. W., Stunkard, H. W., Arenson, S. B., Devereux, E. D., Leonard, H. B., Smail, Lloyd L., Haring, M. M., Colver, C. W., Welker, Wm. H., Schriever, William, de Forest, H., Osborn, R. A., Crofutt, Charles B., Emerson, Alfred, Ott, W. P., Adams, Thomas C., and Fleisher, Moyer S.

Published
1929

28. Schiff base oligopyrrolic macrocycles as ligands for lanthanides and actinides

Author

Sessler, J. L., Melfi, P. J., Tomat, E., Callaway, W., Huggins, M. T., Gordon, P. L., Keogh, D. W., Date, R. W., Bruce, D. W., Bertrand Donnio, Masson, Beatrice, Institut de Physique et Chimie des Matériaux de Strasbourg (IPCMS), and Université Louis Pasteur - Strasbourg I-Centre National de la Recherche Scientifique (CNRS)

Published
2006

32. Preliminary speaker recognition experiments on the NATO N4 corpus

Author

Zissman, M., Buuren, R. van, Grieco, J., Reynolds, D., Steeneken, H., Huggins, M., and TNO Technische Menskunde

Subjects
Defence
Abstract

The NATO N4 corpus contains speech collected at naval training schools within several NATO countries. The speech utterances comprising the corpus are short, tactical transmissions typical of NATO naval communications. In this paper, we report the results of some preliminary speaker recognition experiments on the N4 corpus. We compare the performance of three speaker recognition systems developed at TNO Human Factors, the US Air Force Research Laboratory, Information Directorate and MIT Lincoln Laboratory on the segment of N4 data collected in the Netherlands. Performance is reported as a function of both training and test data duration. We also investigate the impact of cross-language training and testing.

Published
2001

33. Accreditation benefits nation's jails, juvenile detention centers

Author

Huggins, M. Wayne and Kehoe, Charles J.

Subjects
Jails -- Standards, Juvenile detention homes -- Standards, Correctional institutions -- Standards, Law
Abstract

The management of jails and juvenile detention facilities has been improved with the usage of standards developed by the American Correctional Assn (ACA). The standards recommend policies and procedures for detention management and criteria for assessing the quality of services offered. Application of ACA standards in juvenile detention centers have helped these centers define their objectives.

Published
1992

39. Predictive testing for Huntington disease

Author

Huggins, M and Hayden, M R

Subjects
Moral Obligations, Social Responsibility, Letter, Huntington Disease, Genetic Diseases, Inborn, Humans, Ethics, Medical, Female, Physician's Role, Trust, Truth Disclosure, Risk Assessment, Confidentiality
Abstract

A doctor has duties towards his patients of both confidentiality and veracity and at times these may conflict, as in the following case. A mother who has the symptoms of Huntington's chorea does not wish her daughters to know. The doctor must try to make her realise how valuable the information can be to the daughters, and thus obtain her consent to inform them. If the mother's consent cannot be obtained, then the doctor must tell the mother that he cannot allow her attitude to deprive the daughters of this information, especially at this crucial time as they plan to have children. The well-being of the daughters' future families must take precedence over the mother's desire for secrecy regarding her condition.

Published
1992

43. Linkage disequilibrium and modification of risk for Huntington disease

Author

Adam, S., Theilmann, J., Buetow, K., Hedrick, A., Collins, C., Weber, B., Huggins, M., and Michael Hayden

Subjects
Adult, Male, Genetic Linkage, Middle Aged, United Kingdom, Pedigree, Huntington Disease, Gene Frequency, Risk Factors, Mutation, Humans, Female, Deoxyribonucleases, Type II Site-Specific, Alleles, Polymorphism, Restriction Fragment Length, Research Article
Abstract

The major limitation in performing predictive testing for Huntington disease (HD) is the unavailability of DNA from crucial family members. In our program approximately 20% (36/183) of persons have been excluded from predictive testing because of this reason. The major aim of this study was to examine whether data derived from linkage disequilibrium could modify risk analysis for persons at risk for HD. As a first step, we assessed whether the previously reported linkage disequilibrium between alleles recognized by probe pBS674E-D at locus D4S95 remained significant in a much larger data set. A total of 1,150 chromosomes from 622 individuals--200 affected and 422 unaffected--from 118 families were assessed. Significant haplotype association was detected with AccI and MboI RFLPs at the locus D4S95, with all the families (P = .00003), as well as for a subset from the United Kingdom (P = .0037). Data derived from linkage disequilibrium studies using D4S95 modifies the risk for HD, especially in persons of U.K. descent. Utilization of this approach for risk modification of HD awaits both validation of these data and additional information concerning ethnic-specific alleles at the D4S95 locus.

Published
1991

49. Attitudes toward direct predictive testing for the Huntington disease gene. Relevance for other adult-onset disorders. The Canadian Collaborative Group on Predictive Testing for Huntington Disease.

Author

Babul R, Adam S, Kremer B, Dufrasne S, Wiggins S, Huggins M, Theilmann J, Bloch M, Hayden MR, Babul, R, Adam, S, Kremer, B, Dufrasne, S, Wiggins, S, Huggins, M, Theilmann, J, Bloch, M, and Hayden, M R

Abstract

Objective: To assess attitudes toward, and projected utilization of, direct mutation testing by individuals at risk for Huntington disease (HD).Design: Prior to the cloning of the gene for HD, a questionnaire concerning the use of a definitive test was constructed and mailed to 354 participants in the Canadian Collaborative Study for HD. Completed questionnaires were received from 250 participants (response rate, 71%). Persons were asked to indicate whether they would participate in a new predictive test that was either 100% accurate (the definitive test, requiring blood only from the proband) or only 99% accurate.Results: Most (72%) of the persons who had previously received a result in a predictive testing program said they would request testing in either situation. Significantly more persons would request the definitive test than the 99% accurate test (72% vs 58%; P < .02). Respondents for whom testing was uninformative in the linkage test program or who had previously received an increased-risk result were more likely to indicate they would use the test than those who received a decreased-risk result or chose not to have the original test (P = .0003). Less than half (46%) of the participants who initially chose not to have the linkage test said they would return for the new direct test. The major factor that has limited acceptance of predictive testing for this group is the concern about receiving an increased-risk result in the absence of any therapy to alter progression of the disease.Conclusions: A direct mutation test for HD will most readily be accepted by persons who wanted but could not previously receive a result in the linkage test program and those who previously received an increased-risk result. In the absence of therapy, the majority of persons who previously chose not to have predictive testing are unlikely to participate in a new test despite improved accuracy. This has implications for the expected demands for testing services for other adult-onset genetic disorders. [ABSTRACT FROM AUTHOR]

Published
1993
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results