Automatic detection of nocuous coordination ambiguities in natural language requirements.

Folding%2Daware and Structure%2Dconscious 3D Substructures in Folding Data Identification and Applications.

On Building a Quantitative Food%2DDisease%2DGene Network.

Community discovery and sentiment mining for Chinese BLOG.

Research on cooperative distributed geographic modeling control pattern based on Holon.

Collecting high quality overlapping labels at low cost.

Strategies for adaptive smoothing and rebuffering via dynamic network calculus.

A %284n %2D 9%29%2F3 diagnosis algorithm for generalised cube networks.

A Semi%2Dsupervised Topic%2DDriven Approach for Clustering Textual Answers to Survey Questions.

Modern Agricultural Digital Management Network Information System of Heilongjiang Reclamation Area Farm.

Case Analysis of Farm Agriculture Machinery Informatization Management Network System.

Multiple models adaptive feedforward decoupling controller using HDBD focus.%

Multiple models adaptive feedforward decoupling controller using HDBD focus.%

An evolutionary approach to system%2Dlevel fault diagnosis.

Feature selection for automatic taxonomy induction.
Most existing automatic taxonomy induction systems exploit one or more features to induce a taxonomy; nevertheless there is no systematic study examining which are the best features for the task under various conditions. This paper studies the impact of using different features on taxonomy induction for different types of relations and for terms at different abstraction levels. The evaluation shows that different conditions need different technologies or different combination of the technologies. In particular, co-occurrence and lexico-syntactic patterns are good features for is-a, sibling and part-of relations; contextual, cooccurrence, patterns, and syntactic features work well for concrete terms; co-occurrence works well for abstract terms.
Path selection based on service curve measurement in MANETs.

Structural stability and robustness to bounded rationality for non%2Dcompact cases.

Assigning roles to protein mentions The case of transcription factors.

Multivariable direct adaptive decoupling controller using multiple models and a case study.

An Effective Approach for Identifying Evolving Three%2DDimensional Structural Motifs in Protein Folding Data.

XMAS An Experiential Approach for Visualization





Network Fault Diagnosis An Artificial Immune System Approach.

A Nonlinear Hierarchical Multiple Models Neural Network Decoupling Controller.

CRO a system for online review structurization.

Metric%2Dbased ontology learning.
Ontology learning is an important task in Artificial Intelligence, Semantic Web and Text Mining. This paper presents a novel framework for, and solutions to, three practical problems in ontology learning. An incremental clustering approach is used to solve the problem of unknown group names. Learned models at each level of an ontology address the problem of no control over concept abstractness. A metric learning module moves beyond the limitation of traditional use of features and incorporates heterogeneous semantic evidence into the learning process. The metric-based learning framework integrates these separate components into a single, unified solution. An extensive evaluation with Word-Net and Open Directory Project data demonstrates that the method is more effective than a state-of-the-art baseline algorithm.
Learning the distance metric in a personal ontology.
Personal ontology construction is the task of sorting through relevant materials, identifying the main topics and concepts, and organizing them to suit personal needs. Automatic construction of personal ontologies is difficult in part because measuring the semantic distance between two concepts is difficult. Knowledge-based approaches use either knowledge bases, such as WordNet, or lexico-syntactic patterns to induce the differences between concepts. However, these techniques are only applicable for a subset of concepts and leave the majority unmeasurable. On the other hand, statistical approaches are able to induce the differences between any concept pair but lack of human knowledge involvement and hence suffer from low precision. In the context of personal ontology construction, semantic distances between concepts need to reflect personal preferences. Based on that, this paper presents a supervised hierarchical clustering framework to incorporate personal preferences for distance metric learning in personal ontology construction. In this framework, periodic manual guidance provides training data for learning a distance metric and the learned metric is used during automatic activities to further construct the ontology. A detailed user study demonstrates that the approach is effective and accelerates the construction of personal ontologies.
Guaranteed Cost Control of Networked Control Systems with Long Time Delay.


Stabilization of Networked Control Systems with Long Varying Time Delay Based on Two Steps Transformation.


Patterns in Spatio%2Dtemporal Data.
Abstract Moving point object data can be analyzed through the discovery of patterns. We consider the computational efficiency of detecting four such spatio-temporal patterns, namely flock, leadership, convergence, and encounter, as defined by Laube et al., 2004. These patterns are large enough subgroups of the moving point objects that exhibit similar movement in the sense of direction, heading for the same location, and/or proximity. By the use of techniques from computational geometry, including approximation algorithms, we improve the running time bounds of existing algorithms to detect these patterns. 1 Introduction Moving point object data is becoming increasingly more available since the development of GPS andradio transmitters. One of the objectives of spatio-temporal data mining [16, 23] is to analyze such data sets for interesting patterns. For example, a group of caribou with radio collars gives rise to the positionsof each caribou at a sequence of time steps. Analyzing this data gives insight into entity behavior, in particular, migration patterns [22]. The analysis of moving objects also has applications in sports (e.g.,soccer players [12]) and in socio-economic geography [9]. There is ample research on data mining of moving objects (e.g., [13, 25, 26, 28]) in particular, on thediscovery of similar trajectories or clusters. In general the input is a set of
Identification of transcription factor contexts in literature using machine learning approaches.

DELAY  A Lazy Approach for Mining Frequent Patterns over High Speed Data Streams.

A spatio%2Dtemporal mining approach towards summarizing and analyzing protein folding trajectories.
A spatio-temporal mining approach towards summarizing and analyzing protein folding trajectories
A fast diagnosis algorithm for locally twisted cube multiprocessor systems under the MM%2A model.

Next steps in near%2Dduplicate detection for eRulemaking.
Large volume public comment campaigns and web portals that encourage the public to customize form letters produce many near-duplicate documents, which increases processing and storage costs, but is rarely a serious problem. A more serious concern is that form letter customizations can include substantive issues that agencies are likely to overlook. The identification of exact- and near-duplicate texts, and recognition of unique text within nearduplicate documents, is an important component of data cleaning and integration processes for eRulemaking. This paper presents DURIAN (DUplicate Removal In lArge collectioN), a refinement of a prior near-duplicate detection algorithm DURIAN uses a traditional bag-of-words document representation, document attributes (&amp;quot;metadata&amp;quot;), and document content structure to identify form letters and their edited copies in public comment collections. Experimental results demonstrate that DURIAN is about as effective as human assessors. The paper concludes by discussing challenges to moving near-duplicate detection into operational rulemaking environments.
DURIAN a demo for near%2Dduplicate detection.
Recently, the move from paper to electronic public comments makes it much easier for individuals to customize form letters while harder for agencies to identify substantive information since there are many near-duplicate comments that express the same viewpoint in slightly different language. The identification of exact- and near-duplicate texts, and recognition of unique text within near-duplicate documents, is an important component of data cleaning and integration processes for eRulemaking.
Mining Spatial and Spatio%2DTemporal Patterns in Scientific Data.
This report summarizes the Ph.D. Workshop held in conjunction
Multiple Models Fuzzy Decoupling Controller for a Nonlinear System.

Mining Spatial and Spatio%2DTemporal Patterns in Scientific Data.
This report summarizes the Ph.D. Workshop held in conjunction
Multiple Models Adaptive Decoupling Controller Using Dimension%2DBy%2DDimensionTechnology.

Hierarchical Multiple Models Neural Network Decoupling Controller for a Nonlinear System.

Chaos Optimization Strategy on Fuzzy%2Dimmune%2DPID Control of the Turbine Governing System.

Analysis and Comparison of Typical Reversible Watermarking Methods.

Near%2Dduplicate detection by instance%2Dlevel constrained clustering.
For the task of near-duplicated document detection, both traditional fingerprinting techniques used in database community and bag-of-word comparison approaches used in information retrieval community are not sufficiently accurate. This is due to the fact that the characteristics of near-duplicated documents are different from that of both “almost-identical ” documents in the data cleaning task and “relevant ” documents in the search task. This paper presents an instance-level constrained clustering approach for near-duplicate detection. The framework incorporates information such as document attributes and content structure into the clustering process to form near-duplicate clusters. Gathered from several collections of public comments sent to U.S. government agencies on proposed new regulations, the experimental results demonstrate that our approach outperforms other near-duplicate detection algorithms and as about as effective as human assessors.
Knowledge Transfer and Opinion Detection in the TREC 2006 Blog Track.
The paper describes the opinion detection system developed in Carnegie Mellon University for TREC 2006 Blog track. The system performed a two-stage process: passage retrieval and opinion detection. Due to lack of training data for the TREC Blog corpus, online opinion reviews provided in other domains, such as movie review and product review, were used as the training data. Knowledge transfer was performed to make the cross-domain learning possible. Logistic regression ranked the sentence-level opinions vs. objective statements. The evaluation shows that the algorithm is effective in the task.
Two%2Dstage statistical language models for text database selection.

Near%2Dduplicate detection for eRulemaking.
respond to public comments before issuing regulations. In recent years, agencies have begun to accept comments via both email and Web forms. The transition from paper to electronic comments makes it much easier for individuals to customize &#034;form&#034; letters, which they do, creating &#034;near-duplicate&#034; comments that express the same viewpoint in slightly different languages. This paper explores the use of simple text clustering and retrieval algorithms for identifying near-duplicate public comments. Experiments with public comments about a recent regulation proposed by the Environmental Protection Agency (EPA) demonstrate the effectiveness of the algorithms.
A Dual%2DMode Fuzzy Model Predictive Control Scheme for Unknown Continuous Nonlinear System.

Component Content Soft%2DSensor Based on Adaptive Fuzzy System in Rare%2DEarth Countercurrent Extraction Process.

Ontology%2DBased Resource Descriptions for Distributed Information Sources.

Intelligent Optimal Control in Rare%2DEarth Countercurrent Extraction Process via Soft%2DSensor.

Mining Spatial Object Associations for Scientific Data.
In this paper, we present efficient algorithms to discover spatial associations among features extracted from scientific datasets. In contrast to previous work in this area, features are modeled as geometric objects rather than points. We define multiple distance metrics that take into account objects ’ extent. We have developed algorithms to discover two types of spatial association patterns in scientific data. We present experimental results to demonstrate the efficacy of our approach on real datasets drawn from the bioinformatic domain. We also highlight the importance of the discovered patterns by integrating the underlying domain knowledge. 1
A generalized framework for mining spatio%2Dtemporal patterns in scientific data.
In this paper, we present a general framework to discover spatial associations and spatio-temporal episodes for scientific datasets. In contrast to previous work in this area, features are modeled as geometric objects rather than points. We define multiple distance metrics that take into account objects ’ extent and thus are more robust in capturing the influence of an object on other objects in spatial neighborhood. We have developed algorithms to discover four different types of spatial object interaction (association) patterns. We also extend our approach to accommodate temporal information and propose a simple algorithm to derive spatio-temporal episodes. We show that such episodes can be used to reason about critical events. We evaluate our framework on real datasets to demonstrate its efficacy. The datasets originate from two different areas: Computational Molecular Dynamics and Computational Fluid Flow. We present results highlighting the importance of the identified patterns and episodes by using knowledge from the underlying domains. We also show that the proposed algorithms scale linearly with respect to the dataset size.
Dynamic Classification of Defect Structures in Molecular Dynamics Simulation Data.
In this application paper we explore techniques to classify anomalous structures (defects) in data generated from abinitio Molecular Dynamics (MD) simulations of Silicon (Si) atom systems. These systems are studied to understand the processes behind the formation of various defects as they have a profound impact on the electrical and mechanical properties of Silicon. In our prior work we presented techniques for defect detection [11, 12, 14]. Here, we present a two-step dynamic classifier to classify the defects. The first step uses up to third-order shape moments to provide a smaller set of candidate defect classes. The second step assigns the correct class to the defect structure by considering the actual spatial positions of the individual atoms. The dynamic classifier is robust and scalable in the size of the atom systems. Each phase is immune to noise, which is characterized after a study of the simulation data. We also validate the proposed solutions by using a physical model and properties of lattices. We demonstrate the efficacy and correctness of our approach on several large datasets. Our approach is able to recognize previously seen defects and also identify new defects in real time. 1
Toward Unsupervised Correlation Preserving Discretization.

Discovering Spatial Relationships Between Approximately Equivalent Patterns in Contact Maps.

Correlation Preserving Discretization.

Association%2DRule Based Information Source Selection.

Effectiveness of web page classification on finding list answers.
List question answering (QA) offers a unique challenge in effectively and efficiently locating a complete set of distinct answers from huge corpora or the Web. In TREC-12, the median average F 1 performance of list QA systems was only 6.9%. This paper exploits the wealth of freely available text and link structures on the Web to seek complete answers to list questions. We employ natural language parsing, web page classification and clustering to find reliable list answers. We also study the effectiveness of web page classification on both the recall and uniqueness of answers for web-based list QA.
FADA find all distinct answers.
The wealth of information available on the web makes it an attractive resource for seeking quick answers. While web-based question answering becomes an emerging topic in recent years, the problem of efficiently locating a complete set of distinct answers on the Web is far from being solved. We introduce our system, FADA, which relies on question event analysis, web page clustering, and natural language parsing, to find reliable distinct answers with high recall. The method has been found to be effective in strengthening state-of-the-art Web question answering techniques by emphasizing on answer completeness and uniqueness.
Hierarchical Classification for Multiple%2C Distributed Web Databases.

VideoQA question answering on news video.
When querying a news video archive, the users are interested in retrieving precise answers in the form of a summary that best answers the query. However, current video retrieval systems, including the search engines on the web, are designed to retrieve documents instead of precise answers. This research explores the use of question answering (QA) techniques to support personalized news video retrieval. Users interact with our system, VideoQA, using short natural language questions with implicit constraints on contents, context, duration, and genre of expected videos. VideoQA returns short precise news video summaries as answers. The main contributions of this research are: (a) the extension of QA technology to support QA in news video; and (b) the use of multi-modal features, including visual, audio, textual, and external resources, to help correct speech recognition errors and to perform precise question answering. The system has been tested on 7 days of news video and has been found to be effective. Categories and Subject Descriptor H.3.1 [Content Analysis and Indexing]: linguistic processing, thesaurus.
Intelligent Search for Distributed Information Sources Using Heterogeneous Neural Networks.

A Language Modeling Approach to Search Distributed Text Databases.

Hierarchical Classification for Multiple%2C Distributed Web Databases.

QUALIFIER Question Answering by Lexical Fabric and External Resources.
One of the major challenges in TRECstyle  question-answering (QA) is to overcome  the mismatch in the lexical representations  in the query space and  document space. This is particularly severe  in QA as exact answers, rather than  docmnents, are required in response to  questions. Most current approaches overcome  the mismatch problem by employing  either data redundancy strategy  through the use of Web or linguistic resources.
Modeling Web Knowledge for Answering Event%2Dbased Questions.

Structured use of external knowledge for event%2Dbased open domain question answering.
One of the major problems in question answering (QA) is that the queries are either too brief or often do not contain most relevant terms in the target corpus. In order to overcome this problem, our earlier work integrates external knowledge extracted from the Web and WordNet to perform Event-based QA on the TREC-11 task. This paper extends our approach to perform event-based QA by uncovering the structure within the external knowledge. The knowledge structure loosely models different facets of QA events, and is used in conjunction with successive constraint relaxation algorithm to achieve effective QA. Our results obtained on TREC-11 QA corpus indicate that the new approach is more effective and able to attain a confidence-weighted score of above
QUALIFIER In TREC%2D12 QA Main Task.
This paper describes a question answering system and its various modules to solve definition, factoid and list questions defined in the TREC12 Main task. In particular, we tackle the factoid QA task by Event-based Question Answering. Each QA event comprises of elements describing different facets like time, location, object, action etc. By analyzing the external knowledge from pre-retrieved TREC documents, Web documents, WordNet and Ontology to discover the QA event structure, we explore the inherent associations among QA elements and then obtain the answers. There are three subsystems working parallel to handle definition, factoid, and list questions separately. We highlight the shared modules, fine-grained named entity recognition, anaphora resolution and canonicalization co-reference resolution, among the three subsystems as well.
Modeling Web Knowledge for Answering Event%2Dbased Questions.

Query Expansion with Naive Bayes for Searching Distributed Collections.

The Integration of Lexical Knowledge and External Resources for Question Answering.
For the short, factoid questions in TREC, the query terms we get from the original questions are either too brief or often do not contain most relevant information in the corpus. It will be very difficult to find the answer (especially exact answer) in a large text document collection because of the gap between the query space and the document space. In order to bridge this gap, there is a need to expand the original queries to include the terms in the document space. In this research, we investigate the integration of both the Web and WordNet in performing local context and lexical correlations to bridge the gap. In order to minimize the noise introduced by the external resources, we explore detailed question classes, fine-grained named entities, and successive constraint relaxation.
On the Use of Constrained Associations for Web Log Mining.
this paper, we first present an approach based on association rule mining. Our algorithm discovers association rules that are constrained (and ordered) temporally. The approach relies on the simple premise that pages accessed recently have a greater influence on pages that will be accessed in the near future. The approach not only results in better predictions, it also prunes the rule-space significantly, thus enabling faster online prediction. Further refinements based on sequential dominance are also evaluated, and prove to be quite effective. Detailed experimental evaluation shows how the approach is quite effective in capturing a web user&#039;s access patterns; consequently, our prediction model not only has good prediction accuracy, but also is more efficient in terms of space and time complexity. The approach is also likely to generalize for e-commerce recommendation systems
Performance of ATM networks under hybrid ARQ%2FFEC error control scheme.

