To manage the information flood in the Internet, we usually navigate specific information using the provided search engines. Search engines are convenient but with limited functions. For example, it is impractical and impossible to browse through the entire collected information for us to gain an overall picture about what the navigated information stands for. To do so, we need an appropriate approach to automatically extracting concepts from the navigated information to assist users to easily and quickly gain the primary understanding toward a topic that interests users. In this research, we propose an approach to extracting concepts from the navigated web information and detecting the concept changes over time. It basically includes two stages. In the first stage, information is decomposed into paragraphs and they are clustered with key terms identified through the aid of latent semantic indexing method. Concepts are represented in the form of paragraph summary and associated key terms, which allows the user to easily comprehend what they describe. The second stage is to adaptively modify the concept structure to detect concept changes. With new information added, the concepts could be merging, splitting, or even emerging with time. Three experiments are conducted in this research to verify the proposed approach. Results of the first and second experiments show both high recall and high precision that matches the predefined concept categories. The last one is an illustrated real case application on the tsunami event. It shows that we can easily grasp different concepts of the tsunami reports and realize their changes by using our approach. The feasibility of employing our approach is thus justified.

With the rapid expansion of e-commerce, the Web has become an excellent source for gathering customer opinions (or so-called customer reviews). Customer reviews are essential for merchants or product manufacturers to understand general responses of customers on their products for product or marketing campaign improvement. In addition, customer reviews can enable merchants better understand specific preferences of individual customers and facilitates making effective marketing decisions. Prior data mining research mainly concentrates on analyzing customer demographic, attitudinal, psychographic, transactional, and behavioral data for supporting customer relationship management and marketing decision making and did not pay attention to the use of customer reviews as additional source for marketing intelligence. Thus, the purpose of this research is to develop an efficient and effective opinion summarization technique. Specifically, we will propose a semantic-based product feature extraction technique (SPE) which aims at improving the existing product feature extraction technique and is desired to enhance the overall opinion summarization effectiveness.

Interoperability between Geographic Information System (GIS) of different infrastructure companies is still a problem to be handled. Infrastructure companies deal with many operations as a part of their daily routine such as a regular maintenance, or sometimes they deal with unexpected situations such as a malfunction due to natural event, like a flood or an earthquake. These situations may affect all companies and affected infrastructure companies response to these effects. Responses may result in consequences and in order to model these consequences on GIS, GISs are able to share information, which brings the interoperability problem into the scene. The present research, aims at finding an answer to interoperability problem between GISs of different companies by considering contextual information. During the study, the geographical features are handled as the major concern and interoperability problem is examined by targeting them. The model constructed in this research is based on the ontology and because the meaning of the terms in the ontology depends on the context, ontology based context modeling is also used. v In this research, a system implementation is done for two different GISs of two

Universiẗat, Diss.--Duisburg-Essen, 2005.

Univ., Diss.--Saarbrücken, 2006.

The objective of this Thesis research is to develop algorithms for temporally consistent semantic segmentation in videos. Though many different forms of semantic segmentations exist, this research is focused on the problem of temporally-consistent holistic scene understanding in outdoor videos. Holistic scene understanding requires an understanding of many individual aspects of the scene including 3D layout, objects present, occlusion boundaries, and depth. Such a description of a dynamic scene would be useful for many robotic applications including object reasoning, 3D perception, video analysis, video coding, segmentation, navigation and activity recognition. Scene understanding has been studied with great success for still images. However, scene understanding in videos requires additional approaches to account for the temporal variation, dynamic information, and exploiting causality. As a first step, image-based scene understanding methods can be directly applied to individual video frames to generate a description of the scene. However, these methods do not exploit temporal information across neighboring frames. Further, lacking temporal consistency, image-based methods can result in temporally-inconsistent labels across frames. This inconsistency can impact performance, as scene labels suddenly change between frames. The objective of our this study is to develop temporally consistent scene descriptive algorithms by processing videos efficiently, exploiting causality and data-redundancy, and cater for scene dynamics. Specifically, we achieve our research objectives by (1) extracting geometric context from videos to give broad 3D structure of the scene with all objects present, (2) Detecting occlusion boundaries in videos due to depth discontinuity, (3) Estimating depth in videos by combining monocular and motion features with semantic features and occlusion boundaries.

The overall purpose of this study is to investigate lexical semantic representation in bilinguals who speak typologically different languages, specifically, Mandarin and English. Three questions are posed about semantic representation: 1) Do bilingual speakers demonstrate greater heterogeneity in semantic knowledge than monolingual speakers; 2) To what extent do bilingual speakers use paradigmatic and syntagmatic relations to organize their semantic knowledge; and 3) What is the cross- linguistic overlap in bilingual speakers' semantic representation. Thirty Mandarin- English bilingual adults and 30 monolingual English-speaking adults participated in a repeated word association task and generated three associations to each of 36 stimuli. The bilingual speakers completed the same task in their two languages on two different days whereas the monolingual speakers responded to the same 36 stimuli on two different days. Results indicated that 1) the bilingual speakers produced a more heterogeneous set of responses in English than monolingual speakers; heterogeneity was greater in English than Mandarin among the bilingual speakers; 2) the bilingual speakers produced more paradigmatic associations (e.g., happy-sad, spoon-chopsticks, catch-throw) and fewer syntagmatic associations (e.g., happy-smile, spoon-eat, catch-ball) than the monolingual speakers; and 3) approximately 48% of the bilingual speakers' responses were cross- linguistic synonyms, whereas approximately 76% of the monolingual speakers' responses were identical from session 1 to session 2. These findings suggest that late bilinguals (second language learners) use categorical relations to organize their semantic knowledge to a greater extent than monolingual speakers and that reduced experience with a second language can lead to greater heterogeneity in semantic knowledge in that language. The findings also suggest that bilingual speakers have more distributed semantic representations than monolingual speakers. Additional research is needed to explore the areas of heterogeneity, categorical organization, and cross-linguistic overlap in order to further our understanding of bilingual speakers' semantic knowledge representation. / text

Temporal relation classification is one of the most challenging areas of natural language processing. Advances in this area have direct relevance to improving practical applications, such as question-answering and summarization systems, as well as informing theoretical understanding of temporal meaning realization in language. With the development of annotated textual materials, this domain is now accessible to empirical machine-learning oriented approaches, where systems treat temporal relation processing as a classification problem: i.e. a decision as per which label (before, after, identity, etc) to assign to a pair (i, j) of event indices in a text. Most reported systems in this new research domain utilize classifiers that make decisions effectively in isolation, without explicitly utilizing the decisions made about other indices in a document. In this work, we present a new strategy for temporal relation classification that utilizes global models of temporal relations in a document, choosing the optimal classification for all pairs of indices in a document subject to global constraints which may be linguistically motivated. We propose and evaluate two applications of global models to temporal semantic processing: joint prediction of situation entities with temporal relations, and temporal relations prediction guided by global coherence constraints. / text

Some persons with aphasia exhibit a selective verb deficit, which results in a reduced ability to produce verbs in most contexts. A functional level (Bock & Levelt, 1994) impairment may result in impaired sentence production because the verb serves as the semantic-syntactic interface of a sentence. This interface is related to a verb’s relationship with its arguments/thematics. Arguments fill the syntactic slots of subject and object, and those same words serve as thematic roles by referring to who does what to whom. The current study investigates the effect of Verb Network Strengthening Treatment (VNeST) on sentence production using a single subject experimental design across subjects in 4 participants, 2 with nonfluent aphasia and 2 with fluent aphasia. Participants received semantic treatment aimed at re-strengthening the connections between a verb (e.g., measure) and related thematic pairs that refer to the doer and receiver of the action (e.g., carpenter/lumber, chef/sugar). The ability to produce thematic role pairs for trained verbs was tested during treatment while generalization to the ability to produce sentences containing a subject, verb, and object in a picture description task with trained verbs (e.g., The carpenter is measuring the stairs.) and semantically related untrained verbs (e.g., The nurse is weighing the baby.) was monitored. In addition, pre- and post-treatment single word retrieval of verbs (The Northwestern Verb Production Battery (NVPB) (Thompson, 2002)) and nouns (The Boston Naming Test (Goodglass & Kaplan, 1983)) was examined as well as sentence production abilities in unrelated picture description (NVPB) and constrained connected speech tasks. All participants met treatment criteria and exhibited generalization to sentence production with sentences containing trained and semantically related untrained verbs. Participants 1, 2, and 3 exhibited improvements on all pre- and post-treatment measures, including connected speech. Participant 4 exhibited gains on multiple measures but did not show improvement in connected speech. These findings indicate that treatment aimed at strengthening the verb network results in improved word retrieval in naming and sentence production across multiple tasks. Theoretical and clinical implications regarding the impact of using VNeST on rehabilitation of sentence production deficits in aphasia are discussed. / text

Diamond is a SPARQL query engine for linked data. Linked data is a sub-topic of the Semantic Web where data is represented as a labeled directed graph using the Resource Description Framework (RDF), a conceptual data model for web resources, to affect a web-wide interconnected, distributed labeled graph. SPARQL graph patterns entail portions of this distributed graph. Diamond compiles SPARQL queries into a physical query plan based on a set of newly defined operators that implement a new variant of the Rete match, a well known artificial intelligence (AI) algorithm used for complex pattern-matching problems. / text