HCI Bibliography Home | HCI Conferences | HCIR Archive | Detailed Records | RefWorks | EndNote | Hide Abstracts
HCIR Tables of Contents: 07080910111213

Proceedings of the Workshop on Human-Computer Interaction and Information Retrieval

Fullname:Proceedings of the Workshop on Human-Computer Interaction and Information Retrieval
Editors:Daniel Tunkelang; Ryen White; Bill Kules
Location:Redmond, Washington
Dates:2008-Oct-23
Standard No:hcibib: HCIR08
Papers:24
Pages:88
Links:Conference Website and Proceedings | Papers | Posters | Full Proceedings
  1. Papers
  2. Posters/Demos

Papers

Helping Users Provide Explicit Context-aware Feedback To Measure Search Experience Satisfaction BIBA 5-8
  Raman Chandrasekar; Matthew Scott; Dean Slawson; A. R. D. Rajan; Daniel Makoski
There are many reasons to evaluate the goodness of search engines. We take a quick look at some measures of goodness used today, and list requirements for an additional metric that goes beyond these. We present the Search Experience Satisfaction (SES) metric as a vital addition, filling an evaluation niche, to be used along with result-quality methods (which provide a basic goodness measure) and implicit measures (which provide a sense of user satisfaction without necessarily identifying the causes of satisfaction or dissatisfaction).
   We describe a prototype that makes it easy for users to provide explicit feedback without taking them away from their tasks. A light-weight, 'always available' feedback bar is used to collect such feedback along with the user's context. This can be used to compute an SES metric with subscores that help diagnose specific issues or identify desirable features. We present findings from a user study conducted with this prototype.
Polestar: Assisted Navigation for Exploring Multi-dimensional Information Spaces BIBA 9-12
  Davor Cubranic
We describe a system for interactive exploration of multi-dimensional information spaces with which user may be relatively unfamiliar. Our tool, named Polestar, assists the user through a novel combination of several techniques: guided faceted browsing, multiple summarization perspectives of data in the information space during the navigation, and a flexible interaction model that provides both an overview of available choices at each navigation step and an intuitive interface for navigation. We report the details of each of these three components and how they are used for interactive exploration of business intelligence (BI) content.
UIs for Faceted Navigation: Recent Advances and Remaining Open Problems BIBA 13-17
  Marti Hearst
Faceted navigation is a proven technique for supporting exploration and discovery within an information collection. The underlying data model is simple enough to make navigation understandable while at the same time rich enough to make navigation flexible in a wide range of domains. Nonetheless, there remain issues in both the presentation of navigation options in the interface and in how to extend the model to allow more flexible discovery while still retaining understandability. This paper explores both of these issues.
Creating Exploratory Tasks for a Faceted Search Interface BIBA 18-21
  Bill Kules; Robert Capra
In this paper we describe a process for creating and evaluating exploratory tasks for a faceted search interface. We used the tasks in an eye tracking study of a faceted library catalog search interface. We report on user perceptions of the tasks. The method is intended to be extensible to generate exploratory tasks for other types of interfaces and domains.
Personal Information Organization and Retrieval Using an Activity-Based Desktop Interface BIBA 22-25
  Stephen Voida
The venerable desktop metaphor is beginning to show signs of strain in supporting modern knowledge work. In this position paper, I examine how the desktop metaphor can be re-framed, shifting the focus away from a low-level (and increasingly obsolete) focus on documents and applications to an interface based upon the creation of and interaction with manually declared, semantically meaningful activities. In this position paper, I present the information organization and retrieval aspects of the Giornata desktop interface in detail and describe how I implemented the system to support a longitudinal deployment. I conclude with a sampling of the findings from the user study and propose opportunities for future work based on the experience.
Human-Guided Ontology Learning BIBA 26-29
  Hui Yang; Jamie Callan
This paper leverages human knowledge and understanding in machine learning algorithms for constructing ontologies. Ontology construction is a highly subjective task where a human user builds a data model which represents a set of concepts within a domain and the relationships between those concepts. Personal preferences have crucial impact on manuallybuilt ontologies, however are inadequately captured by traditional supervised machine learning approach. This paper proposes a human-guided machine learning approach, which incorporates periodical manual guidance into a supervised clustering algorithm, for the task of ontology construction. A user study demonstrates that guided machine learning is able to generate ontologies with manually-built quality and less costs. It also shows that periodical manual guidance successfully directs machine learning towards personal preferences.

Posters/Demos

Augmenting Faceted Exploration with ResultMaps BIB i
  Edward Clarkson; James Foley
Investigating the Effects of Enhanced Search Facets BIB ii
  Abdigani Diriye; Ann Blandford; Anastasios Tombros
Beyond the Search Box: Helping Users Find Health Information on the Web BIBA 30-33
  Kevin Duh; Shawn Medero
Internet users are increasingly relying on the Web for health information. Their information needs can often be quite complex, ranging from researching a personal illness to comparing the pros and cons of various treatments. We believe that a search interface beyond the traditional search box is necessary to support users in making informed health decisions. In this paper, we describe the search interface of Healia, a consumer health search engine, which contains advanced search features such as personalization, faceted browsing, and query suggestion. We present some analyses of the query logs to seek to understand how users interact with our search interface.
Collaborative Query Term Suggestion BIBA 34-37
  Gene Golovchinsky; Pernilla Qvarfordt; Jeremy Pickens
Query term suggestion has been an important component of information seeking support tools. It has been used for automatic query expansion and re-ranking operations as part of relevance feedback, manually during exploratory search, and interactively through user selections of suggested terms. Term suggestion has been driven by document analysis and through collaborative filtering algorithms. In this work, we describe a novel approach to generating query term suggestions based on activities of a coordinated search team. Terms extracted from documents based on the actions of one team member and suggested as possible query terms to another member. We evaluated the effectiveness of this approach and found a significant correlation between the use of suggested terms and improvements in recall.
Lightweight Additions to the Web Search Interface Supporting Exploratory Web Search BIBA 38-41
  Orland Hoeber
In this paper, the features of T heHotMap:com that support exploratory Web search processes are described. This system grew out of two academic research projects that explored the use of visualization and interaction as a means for supporting users as they conduct Web search tasks. In T heHotMap:com, three lightweight interface extensions have been added to the commonly used list-based representation of Web search results. These can be used independently or together to support users as they craft queries and explore search results. A scenario of using the system for exploratory Web search is described in this paper; a live demonstration will be provided at the workshop.
Viewing Searching Systems as Learning Systems BIBA 42-45
  Bernard Jansen
Investigating whether users of a searching system are engaged in a learning environment, the results of this research show that information searching is a cognitive learning process with unique searching characteristics specific to particular learning levels. In a laboratory experiment, we studied the searching characteristics of 72 participants engaged in 426 searching tasks. We developed the searching tasks according to Anderson and Krathwohl's categories of the cognitive learning domain. Research results indicate that applying and analyzing, the middle two of the six categories, generally take the most searching effort in terms of queries per session, topics searched per session, and total time searching. The lowest two learning categories, remembering and understanding, exhibit searching characteristics similar to the highest order learning categories of evaluating and creating. These results suggest that users applied simple searching expressions to support their higher level information needs. These findings points to the need for searching system features that engage the user in a learning process.
Focus on Results: Personal and Group Information Seeking Over Time BIBA 46-49
  Gary Marchionini; Robert Capra; Chirag Shah
Information seeking is a fundamental human activity that is applied to an enormous range of information needs and exhibits diverse sets of individual behavioral nuances. Information needs range from fact retrieval to life-long interests in complex constructs and information-seeking behaviors range from brute force exhaustive search to sophisticated heuristics (e.g., building block, successive fraction, pearl growing, e.g., Hawkins & Wagers, 1982) and stochastic estimations. Today's search engines leverage content, links, metadata, and context such as time and place to return information based on searcher queries or selections. It is left to the information seeker to examine, interpret, and manage results independent of the search system, a condition that we aim to address here.
SocialRank: An Ego- and Time-centric Workflow for Relationship Identification BIBA 50-52
  Jaime Montemayor; Chris Diehl; Mike Pekala; David Patrone
From instant messaging and email to wikis and blogs, millions of individuals are generating content that reflects their relationships with others in the world. Since communication artifacts are recordings of life events, we can gain insights into the social structure, attributes, and dynamics from this communication history. To help an analyst explore, discover and identify important social structures in these online communication archives, we have developed SocialRank, an ego- and time-centric work flow for identifying social relationships in an email corpus. This work flow includes four high-level tasks: discovery, validation, annotation and dissemination. Given the volume of data and complex relationship structures that confront the analyst, an effective analytic process must dramatically accelerate the discovery of relevant relationships, facilitate the recordings of assertions and validations of these discoveries, and produce reports for the dissemination of an analyst's findings. SocialRank supports these tasks, through the integration of relationship ranking algorithms with timeline, social network diagram, and multidimensional scaling visualization techniques.
SketchBrain: An Interactive Information Seeking Interface for Exploratory Search BIBA 53-56
  Hogun Park; Sung Hyon Myaeng; Gwan Jang; Jong-wook Choi; Sooran Jo; Hyung-chul Roh
As the Web has become a commodity, it is used for a variety of purposes and tasks that may require a great deal of cognitive efforts. However, most search engines developed for the Web provide users with only searching and browsing capabilities, leaving all the burdens of manipulating information objects to the users. In this paper, we focus on an exploratory search task and propose an underlying framework for human-Web interactions. Based on the framework, we designed and implemented a new information seeking interface that helps users to relieve cognitive burden. The new human-Web interface provides a personal workspace that can be created and manipulated cooperatively with the system, which helps the user conceptualize his information seeking tasks and record their trails for future uses. This interaction tool has been tested for its efficacy as an aid for exploratory search.
Search: From Information to Knowledge BIBA 57-60
  Yan Qu
As more and more people use the Web as a knowledge base or a learning environment, it is important to provide easy access to existing knowledge structures on the web. This article advocates a new type of information seeking system that supports both topical search and the search for knowledge structure. Challenges and opportunities in designing such systems are discussed.
Geography and Networks BIBA 61-62
  Robert Reich
Searching for relevant content on the public Internet has become an arduous task for many reasons, including but not limited to spam, poor content quality and information overload. Thus, a user searching "LCD monitor" might be overloaded with dozens of results of stores and price-comparison sites -- significant time is then required by the user to sift through the content and locate what is relevant to their task.
   In most approach's today, the user must expend significant effort to seek out and identify relevant content. It would be desirable to build a system that could facilitate locating relevant content in a more natural an intuitive fashion. Me.dium has built such a system and this paper address a few of the key concepts and learning's.
What might users be learning from the system? BIBA 63-66
  Catherine Smith
Perhaps the most reliable characteristic of any common web search system is the correlation between an item's position on a results list and the probability that the item will be useful. Several recent studies suggest that search system users have learned this property of ranked lists, and have developed routine procedures (habits) for interaction during search. This paper briefly reviews those studies. The paper then presents additional experimental evidence illustrating that while searchers rely on position as an indicator of an item's value, they also alter their behavior when that evidence becomes unreliable. The paper concludes by arguing that effective mechanisms for assisting searchers will invoke and guide the development of new procedures (habits) for non-routine, complex search. Such a system would, in effect, help its user learn how to search.
Challenges for Supporting Faceted Search in Large, Heterogeneous Corpora like the Web BIBA 67-69
  Jaime Teevan; Susan Dumais; Zachary Gutt
Faceted search systems help people find what they are looking by allowing them to specify not just keywords related to their information need, but also metadata. While such systems hold great potential and have been successfully used in vertical domains, there are many challenges in extending them to large, heterogeneous collections like the Web, corporate intranets, or federated search engines that access many different data silos. In this position paper we discuss the challenges in greater detail. Those that we have identified stem from the fact that such datasets are 1) very large, making it difficult to assign quality meta-data to every document and to retrieve the full set of results and associated metadata at query time, and 2) heterogeneous, making it difficult to apply the same metadata to every result or every query.
Novel User Interfaces via Model-Mediated Information Retrieval BIBA 70-73
  Earl J. Wagner; Larry Birnbaum
Using content-specific models to guide information retrieval can provide richer interfaces to end-users in both navigating news articles and learning the context of news events. We present Brussell, a system that uses semantic models of news event situations to perform anticipatory information retrieval, organize extraction results and present a novel interface for navigating among the milestone events of a situation.
Summarization and Refinement Tags in Folksonomies BIBA 74-76
  Joyce Wang; Vladimir Zelevinsky; Daniel Tunkelang
Folksonomies improve search and navigation of documents by allowing users to collaboratively tag documents. Unfortunately, the number of tags can be overwhelming to users who are seeking information, even when the tags are restricted to those that occur in the search results. In this paper, we describe a novel approach for highlighting tags of interest for users, based on the premise that tags can be useful because they either summarize or refine the current set of results. We also present a treemap interface that visually communicates both kinds of tags to users. Finally, we present the results of a user study designed to test the validity of our approach.
Site Metadata on the Web BIBA 77-80
  Erik Wilde
The navigation structure of Web sites can be regarded as metadata that can be used for interesting applications in User Interface (UI) design and Human-Computer Interaction (HCI), as well as for Information Retrieval (IR) tasks. However, there currently is no established format for site metadata, which makes it hard for Web sites to publish their structure in a machine-readable way, which could then be used by HCI and/or IR applications. We propose a model and a format for site metadata that is built on top of an existing format and thus could be deployed with little overhead by publishers as well as consumers. Making site metadata available as machine-readable data can be used for improving user interfaces (informing user agents about the context of the page they are displaying) and better information retrieval (allowing search engines to use sitemap information for better ranking and display of the results).
Improving Exploratory Search Interfaces: Adding Value or Information Overload? BIBA 81-84
  Max Wilson; mc schraefel
One method for supporting more exploratory forms of search has been to include a compound of new interface features, such as facets, previews, collection points, synchronous communication, and note-taking spaces, within a single search interface. One side effect, however, is that some compounds can be confusing, rather than supportive during search. Faceted browsing, for example, conveys domain terminology and supports rich interaction, but can potentially present an abundance of information. In this paper we focus on the faceted example and conclude with our position that Cognitive Load Theory can be used to estimate and thus manage the potential complexities of adding new features to search interfaces.
Supporting Exploratory Search for the ACM Digital Library BIBA 85-88
  Vladimir Zelevinsky; Joyce Wang; Daniel Tunkelang
The Association for Computing Machinery (ACM) is the world's largest educational and scientific computing society, providing the computing field's premier digital library. Many of its articles are tagged by authors with key words and phrases. Unfortunately, the tagging is sparse and inconsistent. As a result, the use of tags for article retrieval leads to high precision but low recall. The alternative of performing full-text search on the tags leads to unacceptably low precision. We have developed a system to bootstrap on author-supplied tags, thus improving tagging across the collection. Preliminary testing suggests we have achieved an order of magnitude increase in recall without perceptibly sacrificing precision. The system can thus leverage the automatically assigned tags to support exploratory search.