Browsing by Subject "learning"
Now showing 1 - 20 of 21
Results Per Page
Sort Options
Item Automatic Discovery of Semantic Relations using MindNet(2010-05-19) Syed, Zareen; Viegas, Evelyne; Parastatidis, SavasInformation extraction deals with extracting entities (such as people,organizations or locations) and named relations between entities (such as "People born-in Country") from text documents. An important challenge in information extraction is the labeling of training data which is usually done manually and is therefore very laborious and in certain cases impractical. This paper introduces a new “model” to extract semantic relations fully automatically from text using the Encarta encyclopedia and lexical-semantic relations discovered by MindNet. MindNet is a lexical knowledge base that can be constructed fully automatically from a given text corpus without any human intervention. Encarta articles are categorized and linked to related articles by experts. We demonstrate how the structured data available in Encarta and the lexical semantic relations between words in MindNet can be used to enrich MindNet with semantic relations between entities. With a slight trade off of accuracy a semantically enriched MindNet can be used to extract relations from a text corpus without any human intervention.Item Cleaning Noisy Knowledge Graphs(CEUR Workshop Proceedings, 2017-10-22) Padia, AnkurMy dissertation research is developing an approach to identify and explain errors in a knowledge graph constructed by extracting entities and relations from text. Information extraction systems can automatically construct knowledge graphs from a large collection of documents, which might be drawn from news articles, Web pages, social media posts or discussion forums. The language understanding task is challenging and current extraction systems introduce many kinds of errors. Previous work on improving the quality of knowledge graphs uses additional evidence from background knowledge bases or Web searches. Such approaches are di cult to apply when emerging entities are present and/or only one knowledge graph is available. In order to address the problem I am using multiple complementary techniques including entity linking, common sense reasoning, and linguistic analysis.Item Deep Understanding of a Document's Structure(ACM, 2017-12-05) Rahman, Muhammad Mahbubur; Finin, TimCurrent language understanding approaches focus on small documents, such as newswire articles, blog posts, product reviews and discussion forum discussions. Understanding and extracting information from large documents like legal briefs, proposals, technical manuals and research articles is still a challenging task. We describe a framework that can analyze a large document and help people to locate desired information in it. We aim to automatically identify and classify different sections of documents and understand their purpose within the document. A key contribution of our research is modeling and extracting the logical structure of electronic documents using machine learning techniques, including deep learning. We also make available a dataset of information about a collection of scholarly articles from the arXiv eprints collection that includes a wide range of metadata for each article, including a table of contents, section labels, section summarizations and more. We hope that this dataset will be a useful resource for the machine learning and language understanding communities for information retrieval, content-based question answering and language modeling tasks.Item Detecting Spam Blogs: A Machine Learning Approach(AAAI, 2006-07-16) Kolari, Pranam; Java, Akshay; Finin, Tim; Oates, Tim; Joshi, AnupamWeblogs or blogs are an important new way to publish information, engage in discussions, and form communities on the Internet. The Blogosphere has unfortunately been infected by several varieties of spam-like content. Blog search engines, for example, are inundated by posts from splogs – false blogs with machine generated or hijacked content whose sole purpose is to host ads or raise the PageRank of target sites. We discuss how SVM models based on local and link-based features can be used to detect splogs. We present an evaluation of learned models and their utility to blog search engines; systems that employ techniques differing from those of conventional web search engines. We evaluate the effectiveness of a combination of features, and finally report our informal analysis of a blog search engine index.Item Development and preliminary testing of tablet application to increase reading motivation and summarization for adolescent students with ADHD(2015-08) Pinna, Joanne E.; Holman, Lucy; University of Baltimore. School of Information Arts and Technologies; University of Baltimore. Master of Science in Information Design and Information ArchitectureChildren with ADHD have a variety of difficulties with reading, including phonetics, reading comprehension, distractibility, lack of reading organizational skills and a low ability to summarize. This study created a tablet-based reading application designed to enhance their capabilities in developing a multimodal approach to reading. Participants who demonstrated difficulty in completing a reading task in a book exhibited a positive outcome on wanting to complete the reading and tasks in the application and complete their summary writing. The application encourages participants to read, answer questions about what was read about the text, record the answers, access notes written, and it aids in summarization of collected sequential information. A rubric score were used to compare summary writing differences after reading from the book and reading with the application. It was determined that there wasn't significant total score differences between the two, but the rubric score demonstrated areas of improvement.Item Echoes of Diversity: Using Maker Activities to Support Children’s Learning and Empowerment(Universidad Iberoamericana, 2017-03-07) Hamidi, Foad; Saenz, Karla; Baljko, MelanieItem Effect of Exam Wrappers on Student Achievement in Multiple, Large STEM Courses(NSTA) Hodges, Linda C.; Beall, Lisa C.; Anderson, Eric C.; Carpenter, Tara; Cui, Lili; Feeser, Elizabeth; Gierasch, Tiffany; Nanes, Kalman M.; Perks, H. Mark; Wagner, CynthiaMetacognition, the ability to think about and regulate one’s thinking, is an important factor in effective student learning. One intervention to promote student metacognition is the exam wrapper—a reflection students complete after an exam noting how their performance related to their preparation. Results are mixed on the effect of the exam wrapper use on student achievement in single STEM courses. In this study, we implemented exam wrappers in five large science and math courses and examined their impact on students’ course outcomes, as well as students’ self-reported behaviors on the Metacognitive Awareness Inventory (results for over 1,100 distinct individuals). Our data include a subset of students who completed exam wrappers in multiple courses simultaneously. We observed a modest but statistically significant positive relation between exam wrapper use and course grades in each course. The relation between exam wrapper use in multiple courses and cumulative grade point average was also statistically significant for male students. These results did not correlate with students’ metacognitive awareness, however. These findings have important implications for how instructors construct and implement wrappers to maximize their potential usefulness.Item "From Laurels to Learners: Leadership with Virtue"(Journal of Management Development **Found at The University of Liverpool Repository**, 2019-01-09) Antonacopoulou, Elena; Bento, ReginaThe purpose of this paper is to present a new approach to leadership development founded on the principle of the Leader-as-Learner: a reflective human who pursues the 4C – virtues of courage, commitment, confidence and curiosity, rather than the laurels of traditional approaches of heroic leadership. Design/methodology/approach: Exploring art-based methods and fostering a new approach to leadership development: Leaders-as-Learners. Findings: In this paper, studies and theoretical findings from the literature are discussed. Research limitations/implications: This paper includes extending life stories and modes of learning by projecting possible selves as leaders, to learn the daily practice of leadership. Practical implications: Leadership involves not only the art of judgment but refines it through a learning orientation to confront volatility, uncertainty, complexity and ambiguity conditions. Social implications: Leadership is not limited to organizations and in relation to work practices. It is a central aspect in all social affairs and integral to building societies which serve, through leaders, the common good. Originality/value: An approach to leadership development that supports human flourishing and locates leadership among ordinary people who do extra-ordinary things.Item Hiragana: A Visual Introduction(2017-05) Yasuda, Ayumi; Lempke, Paul; Scott-Nelson, Michael; Taylor, Ben; Digital ArtsHiragana is the basic and most fundamental of three Japanese writing systems—hiragana, katakana, and kanji. Hiragana is used in all forms of written communication in combination with the other two systems, or by itself for young readers and beginners. There is a corresponding katakana for each hiragana. Katakana is used primarily for foreign words. Hiragana is also used to clarify pronunciation of kanji. Kanji consists of approximately 4000 characters that are usually modified or nuanced by hiragana. Hiragana is made up of 46 phonetic letters (kana). Each of the kana is a complete syllable with its own sound, and has a set order and direction for writing strokes, which often differ from Western convention. The capstone project is an illustrated book to introduce all 46 kana. In this visual introduction, the audience—both children and adults—will meet each of the 46 kana that have been created with various elements—objects, animals, and items of Japanese culture and heritage—that start with the same sound. For example, あ [a] is made of items such as あり (ari, ants) and あめ (ame, candy). With charming, playful, and stylized illustrations, this book offers a unique, fun way of learning Japanese kana.Item A Hybrid Approach to Unsupervised Relation Discovery Based on Linguistic Analysis and Semantic Typing(Association for Computational Linguistics, 2010-06-06) Syed, Zareen; Viegas, EvelyneThis paper describes a hybrid approach for unsupervised and unrestricted relation discovery between entities using output from linguistic analysis and semantic typing information from a knowledge base. We use Factz (encoded as subject, predicate and object triples) produced by Powerset as a result of linguistic analysis. A particular relation may be expressed in a variety of ways in text and hence have multiple facts associated with it. We present an unsupervised approach for collapsing multiple facts which represent the same kind of semantic relation between entities. Then a label is selected for the relation based on the input facts and entropy based label ranking of context words. Finally, we demonstrate relation discovery between entities at different levels of abstraction by leveraging semantic typing information from a knowledge base.Item Improving Binary Classification on Text Problems using Differential Word Features(ACM, 2009-11-02) Martineau, Justin; Finin, Tim; Joshi, Anupam; Patel, ShamitWe describe an efficient technique to weigh word-based features in binary classification tasks and show that it significantly improves classification accuracy on a range of problems. The most common text classification approach uses a document's ngrams (words and short phrases) as its features and assigns feature values equal to their frequency or TFIDF score relative to the training corpus. Our approach uses values computed as the product of an ngram's document frequency and the difference of its inverse document frequencies in the positive and negative training sets. While this technique is remarkably easy to implement, it gives a statistically significant improvement over the standard bag-of-words approaches using support vector machines on a range of classification tasks. Our results show that our technique is robust and broadly applicable. We provide an analysis of why the approach works and how it can generalize to other domains and problems.Item Learning Assisted Side Channel Delay Test for Detection of Recycled ICs(2020-10-23) Vakil, Ashkan; Niknia, Farzad; Mirzaeian, Ali; Sasan, Avesta; Karimi, NaghmehWith the outsourcing of design flow, ensuring the security and trustworthiness of integrated circuits has become more challenging. Among the security threats, IC counterfeiting and recycled ICs have received a lot of attention due to their inferior quality, and in turn, their negative impact on the reliability and security of the underlying devices. Detecting recycled ICs is challenging due to the effect of process variations and process drift occurring during the chip fabrication. Moreover, relying on a golden chip as a basis for comparison is not always feasible. Accordingly, this paper presents a recycled IC detection scheme based on delay side-channel testing. The proposed method relies on the features extracted during the design flow and the sample delays extracted from the target chip to build a Neural Network model using which the target chip can be truly identified as new or recycled. The proposed method classifies the timing paths of the target chip into two groups based on their vulnerability to aging using the information collected from the design and detects the recycled ICs based on the deviation of the delay of these two sets from each other.Item A Machine Learning Approach to Linking FOAF Instances(AAAI, 2010-01-23) Sleeman, Jennifer; Finin, TimThe friend of a friend (FOAF) vocabulary is widely used on the Web to describe individual people and their properties. Since FOAF does not require a unique ID for a person, it is not clear when two FOAF agents should be linked as coreferent, i.e., denote the same person in the world. One approach is to use the presence of inverse functional properties (e.g., foaf:mbox) as evidence that two individuals are the same. Another applies heuristics based on the string similarity of values of FOAF properties such as name and school as evidence for or against co-reference. Performance is limited, however, by many factors: non-semantic string matching, noise, changes in the world, and the lack of more sophisticated graph analytics. We describe a supervised machine learning approach that uses features defined over pairs of FOAF individuals to produce a classifier for identifying co-referent FOAF instances. We present initial results using data collected from Swoogle and other sources and describe plans for additional analysis.Item Mobile, Collaborative, Context-Aware Systems(AAAI, 2011-08-07) Zavala, Laura; Dharurkar, Radhika; Jagtap, Pramod; Finin, Tim; Joshi, AnupamWe describe work on representing and using a rich notion of context that goes beyond current networking applications focusing mostly on location. Our context model includes location and surroundings, the presence of people and devices, inferred activities and the roles people fill in them. A key element of our work is the use of collaborative information sharing where devices share and integrate knowledge about their context. This introduces a requirement that users can set appropriate levels of privacy to protect the personal information being collected and the inferences that can be drawn from it. We use Semantic Web technologies to model context and to specify high-level, declarative policies specifying information sharing constraints. The policies involve attributes of the subject (i.e., information recipient), target (i.e., the information) and their dynamic context (e.g., are the parties co-present). We discuss our ongoing work on context representation and inference and present a model for protecting and controlling the sharing of private data in context-aware mobile applications.Item Platys: From Position to Place- Oriented Mobile Computing(AAAI Press, 2015-06-01) Zavala, Laura; Murukannaiah, Pradeep K.; Poosamani, Nithyananthan; Finin, Tim; Joshi, Anupam; Rhee, Injong; Singh, MunindarThe Platys project focuses on developing a high-level, semantic notion of location called place. A place, unlike a geospatial position, derives its meaning from a user's actions and interactions in addition to the physical location where it occurs. Our aim is to enable the construction of a large variety of applications that take advantage of place to render relevant content and functionality and, thus, improve user experience. We consider elements of context that are particularly related to mobile computing. The main problems we have addressed to realize our place-oriented mobile computing vision are representing places, recognizing places, and engineering place-aware applications. We describe the approaches we have developed for addressing these problems and related subproblems. A key element of our work is the use of collaborative information sharing where users' devices share and integrate knowledge about places. Our place ontology facilitates such collaboration. Declarative privacy policies allow users to specify contextual features under which they prefer to share or not share their information.Item Polarization contrast vision in Octopus(The Company of Biologists Ltd, 1996-04-01) Shashar, N.; Cronin, T. W.While the ability to analyze polarized light is widespread among animals, its contribution to form vision has not yet been documented. We tested the hypothesis that polarization vision can be used for object discrimination, by training octopuses to distinguish between targets on the basis of the presence or absence of a pattern produced by a 90 ° polarization contrast within the target. Octopuses recognized a 90 ° contrast pattern within a single target, when presented either on a horizontal/vertical axis or on a 45 °/135 ° axis. They were able to transfer their learning to new situations and to detect a polarization contrast when the orientations of the e-vector of light passing through the target center and background differed by as little as 20 °. Polarization vision may provide information similar to that available from color vision and thus serve to enhance the detection and recognition of objects.Item Taming Wild Big Data(AAAI Press, 2014-11-13) Sleeman, Jennifer; Finin, TimWild Big Data is data that is hard to extract, understand, and use due to its heterogeneous nature and volume. It typically comes without a schema, is obtained from multiple sources and provides a challenge for information extraction and integration. We describe a way to subduing Wild Big Data that uses techniques and resources that are popular for processing natural language text. The approach is applicable to data that is presented as a graph of objects and relations between them and to tabular data that can be transformed into such a graph. We start by applying topic models to contextualize the data and then use the results to identify the potential types of the graph’s nodes by mapping them to known types found in large open ontologies such as Freebase, and DBpedia. The results allow us to assemble coarse clusters of objects that can then be used to interpret the link and perform entity disambiguation and record linking.Item Turning Doctoral Students into Faculty in Gerontological Social Work: The AGESW Experience(Taylor & Francis, 2019-10-29) Kusmaul, Nancy; Wladkowski, Stephanie; Hageman, Sally; Gibson, Allison; Mauldin, Rebecca L.; Greenfield, Jennifer C.; Fields, Noelle L.Developing faculty interested in aging may help social work meet the needs of our growing aging population. However, doctoral students need a variety of supports to complete PhDs and become gerontological social work faculty. This study explored one program’s role in supporting the development of social work doctoral students to faculty in gerontology. An e-mail invitation was sent to all former participants (2010–2016 cohorts) of the Association for Gerontology Education in Social Work (AGESW) Pre-Dissertation Fellows Program (PDFP). The 38-question online survey consisted of Likert-type scales, multiple answers, and one open-ended question per section about the program’s impacts on their academic career development in teaching, research, mentoring, and support. Forty-five respondents, representing all six cohorts, completed the survey. More than half reported that the PDFP contributed to their ability to publish their research (64.4%, n = 29), grow their professional network (86.7%, n = 39, and teach (55.5%, n = 25). Doctoral programs provided different experiences than the PDFP, including mentoring, methodological training, professional development, networking, and peer support. Results suggest the PDFP provides content recipients value that supplements instruction received in their institutions. The program’s ability to connect students to each other and to national leaders enhances their career development and socialization into academic roles.Item Understanding the Logical and Semantic Structure of Large Documents(SIAM, 2017-04-27) Rahman, Muhammad MahbuburUp-to-the-minute language understanding approaches are mostly focused on small documents such as newswire articles, blog posts, product reviews and discussion forum en- tries. Understanding and extracting information from large documents such as legal documents, reports, proposals, technical manuals and research articles is still a challenging task. The reason behind this challenge is that the documents may be multi-themed, complex and cover diverse topics. For example, business opportunities may contain information on the background of the business, product or service of the business, plan, team management, financial or budget related data, competitors, logistics, compliance, legal information and boilerplate content that is repeated across documents. The content can be split into multiple files or aggregated into one large file. As a result, the content in the whole document may have different structures and formats. Furthermore, the information is expressed in different forms such as paragraphs of text, headers, data forms, tables, images, mathematical equations, lists or a nested combination of these structures.Item A Unified Bayesian Model of Scripts, Frames and Language(AAAI Press, 2016-02-12) Ferraro, Francis; Durme, Benjamin VanWe present the first probabilistic model to capture all levels of the Minsky Frame structure, with the goal of corpus-based induction of scenario definitions. Our model unifies prior efforts in discourse-level modeling with that of Fill-more's related notion of frame, as captured in sentence-level, FrameNet semantic parses; as part of this, we resurrect the coupling among Minsky's frames, Schank's scripts and Fill-more's frames, as originally laid out by those authors. Empirically, our approach yields improved scenario representations, reflected quantitatively in lower surprisal and more coherent latent scenarios.