Corporations increasingly use and retain information only in the form of electronically held data and documents. As a result, the production and sharing of information in legal proceedings throughout the U.S. will depend heavily on techniques for accessing, searching, organizing and analyzing electronic data -- the principal focus of E-Discovery. Large corporations may have terabytes of e-mail and other files spanning many years that are potentially relevant to a case. In response to a court order, an E-Discovery team must identify, assemble, individuate and categorize an organization's files, segregate all "privileged" material (which may be withheld legally), and deliver a minimally comprehensive and exhaustive set of data to the opposing party -- all in a relatively short amount of time. The techniques needed to accomplish such a task necessarily include search, clustering, classification, filtering, social network analysis, extraction, and more -- and no one of these is sufficient. Such requirements challenge our traditional models for search. In particular, the appropriate user models do not reflect the standard "web" or "enterprise" conditions. This presentation explicates the requirements and types of solutions that dominate E-D.

David Milward
Linguamatics, Cambridge, UK
Accessible Knowledge Discovery Using Agile Natural Language-Based Text Mining

This presentation reviews the challenges faced by the pharmaceutical industry and other knowledge-intensive industries in answering business-critical questions using diverse text resources. It discusses a selection of case studies where an NLP-based approach for discovering relevant facts and relationships from unstructured text is delivering significant value - both in terms of improved productivity and in discovering new knowledge by combining information extracted from different sources.

Brian J Buck
RiverGlass, Illinois
The Science of Search: How the Enterprise Intelligence Cycle (EIC) is Critical to Success

For the knowledge professional, there is more to ‘search’ than simply looking up a specific piece of information; t is about information discovery, relevance assessment, analytical summary and reporting in support of the entire intelligence-gathering process.
   The Enterprise Intelligence Cycle (EIC) is one of the most important information processes within an organization. It encompasses everything from the identification of critical information needs, the entire process of information seeking and collection, the incorporation of new information into organizational and individual knowledge models, and the application of human judgment to create the essential intelligence for timely business decision making. Recognizing and optimizing the EIC will be a prerequisite for success in a world increasingly overloaded with the volume, variety, and velocity of unstructured data – as will the adoption on search techniques that truly take into account user context and purpose.

Miles Kehoe and Mark Bennett
New Idea Engineering, California
Search Security Issues for the Enterprise

Enterprise search must factor into account access control and privacy issues, in particular sensitive documents need to be searchable so that they can be shared with the appropriate audience, but not visible to everyone behind the firewall. For example: nobody wants 401K account summaries to display except after appropriate access has been granted, corporate strategy documents relating to outsourcing or layoff plans should not be viewable by all until they are announced. Security must be handled at the document, sub-document and sub-field levels. Here are our best practices for these leading search engines. And here are actual "gotchas" that we have seem at consumer sites and that you can learn from.

Sid Probstein
Attivio, Massachusetts
Intelligent Integration: Combining Search and BI Capabilities for Unified Information Access

Enterprise search technologies are efficient in filtering unstructured content such as emails and documents. While corporate reports and dashboards display transactional database information, there is a disconnect between technologies. How do you integrate these two sources of data to make them more useful? What about important content that exists outside your organization? There is a new generation of innovative technologies that enable the integration of unstructured content with structured data, bringing together enterprise search with business intelligence capabilities. By enabling automatic updates and alerts in real time, these technologies can affect business processes when it matters: at the convergence of business decisions and actions. For example, drug companies could be alerted whenever a product is mentioned in connection with any terms implying “adverse effects.” Consumer goods companies could search blogs for comments on their products by tying their structured product catalog with the ability to analyze unstructured content and apply sentiment analysis.
  This presentation provides real-world examples and explores new tools that combine enterprise search with business intelligence capabilities that provide faster time to value by:

• Enabling decisions based on an intuitive complete view of your information landscape: both structured and unstructured content including databases, web pages, office documents, email, and media files.
• Providing a single repository: eliminating the need for jumping from application to application based on the type of question, or format of the information you examine.
• Enabling users to use a simple search interface to access all your information assets, rather than learning complex BI applications to access structured data.
• Offering comprehensive connectivity and language support, easy installation and being linearly scalable.
• Integrating data with key business processes in real-time to affect enterprise-wide processes.

David Seuss
Northern Light, Massachusetts
Using Text Analytics for the Automated Analysis and Discovery of Meaning From Large Stores of Market Intelligence

There has been much recent coverage at places like the Search Engine Meeting on using text analytics for reputation management in brand metric tracking applications, but how do you create systems that assist in business analysis, strategic research and competitive intelligence from volumes of news and market research reports?  For example, rather than merely tracking mentions of your brands and measuring the sentiment toward them, you could find out which technologies your competitors are working on, uncover where your competitors are using pricing to gain market share, and identify what product marketing tactics are being employed in your target markets. Text analytics can greatly assist in this process, but using text analytics for strategic research is different from using  it for reputation management, and requires completely different solutions.  This presentation describes the opportunities and challenges in creating systems for the automated analysis and discovery of strategic meaning from market intelligence content.  It describes what it takes to create such systems and outlines the pitfalls needed to be avoided in developing and deploying them.

Jeff Catlin
Lexalytics, Massachusetts
Taking Search to the Next Stage with the Power of Text Analytics

Enterprise search is still growing, evolving and enhancing, and its main purpose continues to be to help users find the answers to their business questions hidden in a complex myriad of sources.  But many people are beginning to ask "what's next?" for the enterprise search industry.  The answer:  combining search technologies with the fast-evolving area of entity extraction and sentiment analysis.  Extracting important metadata and providing insight to the sentiment of those data compliments enterprise search by helping the user uncover the questions they may not think to ask. In fact, those familiar with text analytics would argue that enterprise search is more important than ever to maintain a competitive edge, and that text analytics will play an increasingly large part in that equation.

Christian Reuschling and Andreas Dengel
German Research Center for Artificial Intelligence, Germany
DynaQ - Dynamic Queries for Document-Based, Personal Information Spaces

The paradigms of common, keyword-based document search engines are often not sufficient for the natural searching attitude of human beings. In most systems, the only possibility for searching is to formulate a query from scratch and obtain the results.
  If we have not found what we were looking for, we usually have to start again, reformulating our query. While it is hard for humans to explain an entity completely, it is easier for us to 'navigate' through the document space step-wise, to have an overview of the current state of the search, and, having several tools at hand to support us, to refine the initial query.
  DFKI has developed an inquiry system called 'DynaQ'. Its aim is to enable searchers to explore their personal information space, supporting them with this step-wise searching paradigm called 'orienteering'. For that, the system offers several tools in order to fulfill the Visual Information-Seeking Mantra "overview, zoom & filter, details-on-demand". Some key features of DynaQ will be demonstrated:

• Birds eye view of the result list
• Dynamic query sliders allowing search terms to be weighted, thus dynamically  re-sorting the results on-the-fly
• Thumbnail generation for indexed documents (pdf, office, etc)
• Relevance feedback: Queries can be contextualized by marking one or more documents as relevant. Documents that are similar to them will be ranked higher in the result list. Users can choose between two kinds of similarity:

1. Textual content similarity
2. Image similarity (for text represented in bitmaps or for image files)

• Push search dialogue showing details and related documents according to the attribute similarity (e.g., same author, similar full text)
• Indexing of  all common file formats (e.g., pdf, MS Office, rtf, gif, jpeg) and Emails)
• Availability of the complete Wikipedia index.

Daniel Tunkelang
Endeca, Massachusetts
Enabling the Information Seeking Process

In the early days of information science, the process of finding information was conceived as precisely that: a process. But the success of commercial search engines had the unfortunate side effect of reducing this process to a guessing game of relevance. Given that search engines are not mind readers and cannot reliably infer a person's intent from a two-word query, we need to remember that information seeking calls for a process, a dialogue between the user and the system.
  This presentation outlines the principles of information seeking as a dialogue and walks though concrete examples that illustrate the principles of human-computer information retrieval (HCIR), a vision that is reshaping approaches to information access. Specifically, the presentation shows how designing an application in terms of bi-directional communication between the user and the system addresses the inherent limitations of conventional search engine approaches.

Peter Noerr
MuseGlobal, California
The Underground Information Ecosystem:  Connectors

How these vital, but fragile, items allow different systems to connect and interact, and how they need to be maintained and supported for the information to flow. Like plumbing, they stay out of sight, but are critical for system integration and services, such as federated searching, content harvesting, semantic mapping, and any activity which requires information from more than one source.