Creating the ultimate research assistant

Looks at the custom tool developed by the author that leverages the Google Web search API (or a similar search service) to discover a list of Web pages matching a given topic; identify and extract trends and patterns from these Web pages' text; and transform those trends and patterns into an understandable, useful, and well-organized information resource. The tool accomplishes these tasks using four main components. First, a search engine client discovers a list of relevant Web pages using the Google Web search API. An information extraction engine then mines concepts and associated text passages from these Web pages. Next, a clustering engine organizes the most significant concepts into a hierarchical taxonomy. Finally, a knowledge base generator uses this taxonomy to generate a hypertext knowledge base from the extracted concepts and text passages.     

Information retrieval in the World-Wide Web: Making client-based searching feasible

Finding specific information in the World-Wide Web (WWW, or Web for short) is becoming increasingly difficult, because of the rapid growth of the Web and because of the diversity of the information offered through the Web. Hypertext in general is ill-suited for information retrieval as it is designed for stepwise exploration. To help readers find specific information quickly, specific overview documents are often included into the hypertext.Hypertext systems often provide simple searching tools such as full text search or title search, that mostly ignore the “hyper-structure” formed by the links.In the WWW, finding information is further complicated by its distributed nature. Navigation, often via overview documents, still is the predominant method of finding one's way around the Web.Several searching tools have been developed, basically in two types:

• •A gateway, offering (limited) search operations on small or large parts of the WWW, using a pre-compiled database. The database is often built by an automated Web scanner (a “robot”).
• •A client-based search tool that does automated navigation, thereby working more or less like a browsing user, but much faster and following an optimized strategy.

This paper highlights the properties and implementation of a client-based search tool called the “fish-search” algorithm, and compares it to other approaches. The fish-search, implemented on top of Mosaic for X, offers an open-ended selection of search criteria.Client-based searching has some definite drawbacks: slow speed and high network resource consumption. The paper shows how combining the fish search with a cache greatly reduces these problems. The “Lagoon” cache program is presented. Caches can call each other, currently only to further reduce network traffic. By moving the algorithm into the cache program, the calculation of the answer to a search request can be distributed among the caching servers.     

超文本结构化转换算法的研究与实现

超文本是一种非结构化的文档.它虽然不支持跨页查询和全文检索,但却是Internet上信息组织与存储的重要方式.提出了一种将超文本转换为结构化数据库的算法.分析了超文本结构化转换的需求,运用图论分析并描述了超文本的转换模型与实现算法.该算法在鲁迅数字图书馆系统中得到了实际应用和验证.     

Introduction: In-Use,In-Situ: Extending Field Research Methods—Part 2

This study used protocol analyses and user drawings of their models of the system to investigate the “getting lost” problem in hypertext navigation. The “getting lost” problem is viewed as occurring when routine expectations of naive users, concerning appropriate linear sequences, are violated. Several ways in which users persistently attempt to work within a linear model, despite its inapplicability, are examined. The transition to more hierarchical user models is described.     

Task-based method for creating usable hypertext

Although large hypertext documentation systems have many benefits in the commercial world, they can be difficult to build and use. To help overcome both these obstacles, a method under development at Hewlett-Packard assists authors in creating usable hypertext. A Wizard of Oz experiment was done with traditional on-line documentation and hypertext documentation to see what users liked and disliked. The experiment showed the need for a set of hypertext construction rules to ensure usability. Following these rules, and with computer assistance, authors are expected to do a task analysis of users' activities, to identify critical objects and create content nodes for them, to create links between nodes within clearly defined and cognitively justified limits, and to conduct usability tests on the resulting hypertext volume. Whilst this method recognises the importance of good clear writing, the rules and suggested prectices are not primarily about writing or screen design, but about creating a coherent navigational web that ensures success among commercial users of hypertext. The authoring system described in this paper has been developed for the HP Help system, which has been adopted as the common help delivery system for developers of Common Open Systems Environments on Unix-like platforms.     

A Predictive Model of Human Performance With Scrolling and Hierarchical Lists

ABSTRACT

Many interactive tasks in graphical user interfaces involve finding an item in a list but with the item not currently in sight. The two main ways of bringing the item into view are scrolling of one-dimensional lists and expansion of a level in a hierarchical list. Examples include selecting items in hierarchical menus and navigating through “tree” browsers to find files, folders, commands, or e-mail messages. System designers are often responsible for the structure and layout of these components, yet prior research provides conflicting results on how different structures and layouts affect user performance. For example, empirical research disagrees on whether the time to acquire targets in a scrolling list increases linearly or logarithmically with the length of the list; similarly, experiments have produced conflicting results for the comparative efficacy of “broad and shallow” versus “narrow and deep” hierarchical structures. In this article we continue in the human–computer interaction tradition of bringing theory to the debate, demonstrating that prior results regarding scrolling and hierarchical navigation are theoretically predictable and that the divergent results can be explained by the impact of the dataset's organization and the user's familiarity with the dataset. We argue and demonstrate that when users can anticipate the location of items in the list, the time to acquire them is best modeled by functions that are logarithmic with list length and that linear models arise when anticipation cannot be used. We then propose a formal model of item selection from hierarchical lists, which we validate by comparing its predictions with empirical data from prior studies and from our own. The model also accounts for the transition from novice to expert behavior with different datasets.     

Is a hypertext a book or a space? The impact of different introductory metaphors on hypertext construction

This study examines the impact of different metaphors on the process of hypertext construction. Two groups of 20 college students with no experience in hypertext construction received introductory explanations on the text format “hypertext” based on either a book or a space metaphor. Then they had to construct hypertexts by linking prepared nodes on the topic of the “Internet”. The different metaphors had significant effects on the constructed hypertexts, the construction process, and knowledge acquisition. The book metaphor encouraged a more linear way of viewing hypertexts that conflicted with the complexity of the contents to be processed. The space metaphor permitted a correspondence between complex semantic structures and complex hypertext structures. Hence, the space metaphor seems to be more appropriate for explaining the text format hypertext to students.     

采用捕鱼策略的优化方法

引入一种采用渔夫捕鱼策略的新的优化方法。该优化方法主要采用移动搜索、收缩搜索和加速搜索三种搜索技术。设初始时在搜索域中随机分布有若干个点,每个点看作一个“渔夫”,每个“渔夫”通过移动、收缩和加速三种搜索方式在搜索空间中独立开展寻优活动,以搜寻全局的最优解或最优点。测试结果表明,该算法具有较好的全局搜索能力,因而该优化方法是有效的和可行的。     

Inductive hierarchical nonnegative graph embedding for “verb–object” image classification

Most existing image classification algorithms mainly focus on dealing with images with only “object” concepts. However, in real-world cases, a great variety of images contain “verb–object” concepts, rather than only “object” ones. The hierarchical structure embedded in these “verb–object” concepts can help to enhance classification. However, traditional feature representation methods cannot utilize it. To tackle this problem, we present in this paper a novel approach, called inductive hierarchical nonnegative graph embedding. By assuming that those “verb–object” concept images which share the same “object” part but different “verb” part have a specific hierarchical structure, we integrate this hierarchical structure into the nonnegative graph embedding technique, together with the definition of inductive matrix, to (1) conduct effective feature extraction from hierarchical structure, (2) easily transfer each new testing sample into its low-dimensional nonnegative representation, and (3) perform image classification of “verb–object” concept images. Extensive experiments compared with the state-of-the-art algorithms on nonnegative data factorization demonstrate the classification power of proposed approach on “verb–object” concept images classification.     

Internet-based decision support for evidence-based medicine

The Protocol Assistant is a knowledge-based system, developed by the Department of Artificial Intelligence and AIAI at the University of Edinburgh, which advises on the treatment of parotid tumours. It has been developed to support both adherence to a clinical protocol based on the latest evidence and the use of clinical judgment where the evidence is weak or inconsistent. It was developed using a knowledge modelling technique named PROforma, which is specifically designed for representing best practice guidelines; the PROforma models were used as the basis for a user interface, which was implemented in HTML. A set of rules were developed in JESS (the Java Expert System Shell) which were capable of “running” the protocol; a simple method of reasoning with certainties, based on the “goodness” of each relevant item of published evidence, was used to recommend which path to follow at choice points. However, the user is also supplied with access to the abstracts of all relevant published articles, using the hypertext facilities of HTML. The Protocol Assistant can thus be used either as a “wizard” which guides users through the decision making process, or as a “hypertext manual” which leads them to the information relevant to the decision they are making. This dual-role capability is crucial for the acceptance of KBS in the real world.     

Learning perceptually grounded word meanings from unaligned parallel data

In order for robots to effectively understand natural language commands, they must be able to acquire meaning representations that can be mapped to perceptual features in the external world. Previous approaches to learning these grounded meaning representations require detailed annotations at training time. In this paper, we present an approach to grounded language acquisition which is capable of jointly learning a policy for following natural language commands such as “Pick up the tire pallet,” as well as a mapping between specific phrases in the language and aspects of the external world; for example the mapping between the words “the tire pallet” and a specific object in the environment. Our approach assumes a parametric form for the policy that the robot uses to choose actions in response to a natural language command that factors based on the structure of the language. We use a gradient method to optimize model parameters. Our evaluation demonstrates the effectiveness of the model on a corpus of commands given to a robotic forklift by untrained users.     

Extending the effectiveness of simulation-based DSS through genetic algorithms

Many real life ill-structured problems involve high uncertainty and complexity preventing application of analytical optimization techniques in building effective decision support systems (DSS). These systems may employ simulation method and search for a “good” solution through “what-if” analysis. However, this method is very time consuming and often overlooks the consideration of many promising alternative solutions. A genetic algorithm (GA) automates the search for “good” solutions by finding near-optimal solutions and increases effectiveness of DSS. This paper introduces a hybrid method based on the combination of Monte-Carlo simulation and genetic algorithms. The combined method is illustrated through application to the marketing mix problem to improve the process for searching and evaluating alternatives for decisional support. The paper compares two methods: MC and MC+GA. It also discusses ways for dealing with crisp and soft constraints contained in the example problem. A business game environment is chosen for experiments. The results of the experiments show that the GA-based approach outperforms human “what-if” method in terms of effectiveness and efficiency.     

From text to hypertext: A post-hoc rationalisation of LaTeX2HTML

The conventional hypertext authoring framework compels authors to represent their material as an interconnected network of nodes and links. Apart from the difficulties that this alone entails, the situation with HTML is even more problematic since the author is also responsible for mapping the abstract network model onto the computer file system. This is likely to hinder the widespread adoption of HTML by information owners who are finding it difficult not only to create but also to maintain coherent documents with complex interconnection topologies.In this paper it is argued that familiar document forms such as books, manuals, articles, reports, etc., often contain sufficient structural and cross-referential cues with which to build a rich hypertextual structure. It is shown how this structure can be automatically extracted and then realised as a collection of HTML files which can be explored using generated navigation panels. The conversion process and the advantages of this approach are illustrated with interactive examples using the LaTeX2HTML converter. Other unique features of LaTeX2HTML — mathematical equations and “conditional text” — are also discussed.Allowing authors to work with familiar metaphors and tools without compromising the flexibility afforded to them by the target hypertext system and delivery mechanism is perhaps the main reason for the growing popularity of text to hypertext conversion tools.     

An efficient algorithm for finding a collision-free path among polyhedral obstacles

The use of graph search algorithms on a so called “visibility graph” is a common approach to finding a minimum-distance collision-free path among polyhedral obstacles in a 2D environment. Complexity of the search can be greatly reduced by reducing the size of the graph. The focus of this article is to provide an algorithm aimed at constructing a subvisibility graph using only “necessary” obstacles, i.e., excluding as many obstacles as possible whose vertices are never via points of the shortest collision-free path.     

GDDT—A graphical design and documentation tool for software development

This paper describes “The Graphical Design and Documentation Tool” (GDDT), a general purpose support system for design and documentation. Self documentation and ease of use are presented as important criteria for such tools. Interactive graphics is introduced as a methodology for creating and maintaining hierarchical representations. The graphical representations of various requirements specification techniques and design tools are supported by an extended network model which supports hierarchical decomposition of structures. GDDT is proposed as a general purpose design tool which supports the most common logical constructions of both requirements specification and design methodologies and yet has a high level human interface which makes it easy to learn and use.     

A concept-based information retrieval approach for engineering domain-specific technical documents

Technical documents, which often have complicated structures, are often produced during Architecture/Engineering/Construction (A/E/C) projects and research. Applying information retrieval (IR) techniques directly to long or multi-topic documents often does not lead to satisfactory results. One way to address the problem is to partition each document into several “passages”, and treat each passage as an independent document. In this research, a novel passage partitioning approach is designed. It generates passages according to domain knowledge, which is represented by base domain ontology. Such a passage is herein defined as an OntoPassage. In order to demonstrate the advantage of the OntoPassage partitioning approach, this research implements a concept-based IR system to illustrate the application of such an approach. The research also compares the OntoPassage partitioning approach with several conventional passage partitioning approaches to verify its IR effectiveness. It is shown that, with the proposed OntoPassage approach, IR effectiveness on domain-specific technical reports is as good as conventional passage partitioning approaches. In addition, the OntoPassage approach provides the possibility to display the concepts in each passage, and concept-based IR may thus be implemented.     

Communication support system for operators in emergency of large-scale plant

It is very important to support communications among operators in the emergency of large-scale plants. This paper deals with a communication support system for emergent situations in a hierarchical management structure for large-scale plants. Our system is based on the communication model that considers human-related factors such as “competence,” “duty,” “responsibility,” and “knowledge.” A prototype of a group communication support system is developed based on the model, and the system is evaluated for a plant example. © 1999 John Wiley & Sons, Inc.     

Organizing and presenting geospatial tags in location-based augmented reality

Recent improvements in the capabilities of smart phones are making the location-based augmented reality services a reality. When widely used, such a system is expected to produce many user-created geospatial tags concentrated at popular and important sites. In this paper, we describe a way to organize and group such geospatial tags (manually or automatically using a nearest neighbor algorithm) and how to efficiently interact to search and find the tag that the user might be interested in. The proposed method was implemented on an Apple iPhone, and an experiment was carried out to verify the improved usability. The results do indicate the advantage of the principle of hierarchical organization of data. We also further found that the “automatic-but-less-accurate” approach is more suitable than “precise-but-manual” due to the dynamic nature of the mobile interaction and less than perfect sensing.