Wednesday, November 6, 2013

Search engines, distributed processing and content processing pipelines are not new. However enabling technologies of mature search engines, powerful content processing pipelines and cheap distributed processing are coming together to empower a next generation of information access, analysis and presentation much closer to the holy grails of knowledge management. Hear from the founder of Search Technologies how modern search engines are currently being combined with powerful independent content processing pipelines and the distributed processing technologies from big data to form new and exciting enterprise search architecture, delivering results only available to the biggest companies with the deepest pockets in the past.

Pattern Analysis & Categorization: Big Metadata Toolkits

The exploding volume, complexity, and velocity of structured and unstructured data and their interactions presents challenges and opportunities to derive valuable insights. Among the challenges of managing massive datasets are gathering, validating, preserving, analyzing, and maintaining linkages from those analyses to the source dataset. Identifying patterns in datasets using information retrieval methods and writing out the results as metadata are well-established information management processes that should be adopted by organizations working with today’s Big Data sets. This presentation provides an overview of pattern analysis and categorization methods, including keyword and regular expression matching, business rules, pattern categorizers, entity extraction, and trained categorizers that are the key building blocks of analytics toolkits for big metadata applications.

Using Example-Based Auto-Categorization to Tame Big Data

There are at least two schools of thought regarding Big Data. Part of the organization wants to take advantage of it.The other part views it as “Dark Data”—undiscovered, unanalyzed, and unreachable without the proper analysis tools and skills. It’s both. Concept-based auto-categorization,which automatically categorizes documents based on their actual content, not keywords or terms, is the fastest, easiest, and most repeatable way to pinpoint only the most important documents and emails among libraries spanning millions of files and messages. It is an established standard in legal e-discovery and U.S. intelligence, having proved defensible and highly scalable. Learn how companies are beginning to step up to big data analytics using example-based auto-categorization in order to take advantage of all their data, no matter where—or how—it resides. Analytics for Big Data can bring great value to many business applications—social media, market analysis, internal information analysis. It is also being applied toward governance issues and regulatory and legal compliance matters. We’ll also look at empowering the entire information life cycle; how the “infomediaries” gain business advantages; and why traditional search engines have not been able to keep up.

Benefits of Integrated UX-IM Design

Customers today face market-driven design challenges that span multiple content formats and sources, multiple application platforms, and multiple user platforms. These challenges often require the services of two types of designers: user experience (UX) designers and the information modelers (IM) of taxonomies and metadata. An integrated UX-IM design team is one approach to accelerating solution implementation. When approaching the overall design with a collaborative team, the project can have a common methodology and shorter design cycles.This presentation examines how collaboration between these teams ultimately benefits the entire design cycle while resonating with executive sponsors.

Ensuring Consistent & Accurate Tagging: Interface Design & Metadata Application

As taxonomists, we spend a lot of energy designing vocabularies that are browsable and user-friendly, but we often overlook the system interfaces that taggers use to access and apply taxonomy terms to objects. The irony is, even the best vocabulary is of little value if a poorly designed tagging interface impedes a user from tagging consistently and accurately. This presentation looks at how tagging interfaces impact the metadata application process, exploring examples from content management systems, product information management systems, and crowdsourcing sites from around the web. It discusses how interface design can influence data quality, findability, and the long-term viability of a taxonomy project and concludes with some high-level principles to consider if you’re evaluating, building, or revamping a tagging interface.

Applying User Research to Designing Info Models

User experience design isn’t the only practice that can benefit from user research. Exploring the way users natively organize information in their minds gives taxonomists a means to precisely identify and define information structures, as well as a powerful guiding mechanism for difficult information model design decisions. This talk demonstrates how to construct an effective research plan that blends quantitative and qualitative methods to explore and analyze user mental models and how to transform research insights into viable, flexible, and sustainable information models.

Utilizing Ontologies for Taxonomy & Content Organization

This presentation focuses on the design and implementation of ontologies and how they are leveraged to implement taxonomies and provide better content organization. It demonstrates through case studies how this implementation improves unstructured content search and retrieval. Case studies illustrate how this approach to content organization has improved “findability” and reuse of content and knowledge in several organizations. Along with examples of ontology and taxonomy adoption, an underlying view of the card sort results, keyword, and controlled vocabulary building used to meet the business expectations of the KM solution are shared.

Enhancing Searches With Taxonomies & Semantic Tech

We’ve all seen how the major search engines sometimes second-guess—often correctly—what we meant to search for; taxonomies using standards-based semantic technology can help your own applications do this and more.While semantic technologies rarely try to store the complete meanings of words, data about the relationships of words and phrases to other words and phrases (for example, “broader than” or “alter- native label”) can often store enough semantics to automate search enhancement.When this data is stored using the W3C SKOS standard, it can more easily be aggregated, queried, and used by a variety of tools. Because SKOS is based on RDF, the growing amount of publicly available RDF data about terms and term relationships can be an especially big help to drive improved searches with your own systems. This talk looks at how taxonomies based on semantic technology can help with focusing and augmenting searches, correcting terms, disambiguation, and using other vocabulary metadata sources to improve the data driving your search enhancements.

Developing a Semantic Search Application: A Pharma Case Study

Adding semantics to search can be a daunting task since it involves dealing with language (messy and not really an IT core competency) and relatively new technologies such as text analytics, and, to work well, requires an interdisciplinary team. Given the complexity and uncertainty of developing semantic search, it usually makes sense to start with a small, focused pilot or POC.This is what one pharmaceutical company decided to do. This talk describes an initial pilot carried out by a varied and diverse team using three different search and text analytics products. It discusses the pluses and problems this diversity caused, what worked best, and what didn’t work so well and covers the key issues and approaches that were needed for success.