Conference program chair Stephanie Lemieux welcomes attendees to the start of Taxonomy Boot Camp 2018!

Taxonomies and ontologies are seeing a resurgence of interest and usage as Big Data proliferates, machine learning advances, and integration of data becomes more paramount. The previous models of labor-intensive, centralized vocabulary construction and maintenance do not mesh well in today’s interdisciplinary world. Learn about how information professionals can play a starring role in this new world. McGuinness gives a real-world view of building and maintaining large collaborative, interdisciplinary vocabularies along with the data repositories and services they empower, such as the National Institutes of Environmental Health Sciences’ Child Health Exposure Analysis Resource.

This introductory taxonomy tutorial covers key concepts to get you up-to-speed for the rest of the conference or helps prepare you to take on a role in a taxonomy project. Topics include comparisons and suitable applications of different types of taxonomies/controlled vocabularies (hierarchical, faceted, and thesauri), standards as applied to taxonomies, the relationship of taxonomies to metadata, best practices for developing terms and their relationships, and taxonomy management software.

Our engaging speaker shares how to deliver insight that is targeted and personalized wherever you are. Explore how and why the search box is evolving with your organization’s needs, content services and application context, to meet the priorities of your changing workforce and communication needs.

This session provides an overview of the principles of design thinking, including collective intelligence and human-centered design. Learn some tactical, easy, and fun ways to use these principles to build a taxonomy, including rapid ideation, card sorting, dot-voting, user personas and journey maps, and rapid prototyping, with the support and consensus of end users. Learn about these principles and how they can be applied to build consensus and buy in amongst your stakeholders.

Leading with the best foot forward. How context, connection, and collaboration will guide your team to successful development and implementation of taxonomies. This talk focuses on how to set your team up for success when building a taxonomy. 

To remain relevant, taxonomies must be maintained. They must change to account for new kinds of material being created over time, changes in the world or organization (new products, partners, etc.), and lessons learned through experience with earlier versions of the taxonomy. Learn the rationale and best practices for taxonomy governance processes based on the premise that a team should be established to oversee its maintenance. A sample framework for a taxonomy team is described, including a charter and suggested team roles.

Not everyone has the budget or resources to implement complex text analytics software, but there is a lot you can do with tools you already have on your computer! Spreadsheets and basic builtin command-line tools can normalize text, analyze phrase length, determine occurrence frequency, and more. Get introduced to the most useful tools and programmatic shortcuts for different stages of taxonomy development. Come away with a cheat sheet of formulas and command-line utilities that you can use right away.

Is it creepy or helpful when Spotify recommends a playlist you absolutely love? This talk covers the basics of machine learning, how taxonomies enhance it, and how to recognize machine learning opportunities. The XO Group’s taxonomy team relates machine learning techniques you’ve probably encountered in everyday apps to demonstrate how they’re integrated into products and applications. They share the major role taxonomies play in The Knot’s personalization and recommendation algorithms and how they adjusted and adapted taxonomies and taxonomy management to enhance our machine learning solutions.

Is a burrito a sandwich? This innocuous-seeming question is great fodder for soliciting opinions—and thinking about categories. But when we make inferences based on semantics (as in a machine learning environment), putting things into categories becomes an important ethical gesture. And as taxonomists name things, make categories, and name categories and put things into categories, we are therefore making ethical decisions as part of our job description.

Automatic categorization, automatic taxonomy construction, machine learning, artificial intelligence: These are some of the terms that can make a taxonomist very nervous for his or her future career. Should taxonomists fear automation or embrace it? Is there a place for taxonomy construction and ongoing development as more processes are automated? Hear about the role of taxonomies and the taxonomist in a world of increasing automation, strategies for making the case to the business for the continued need for taxonomies and taxonomist roles, and more.

When Gartner acquired CEB in 2017, it began the complex process of integrating people, products, and platforms. Each company had a fully developed taxonomy of topics as well as different auto-tagging and taxonomy management applications. To enable the migration of hundreds of thousands of content objects and expand the client experience, the team had to reconcile the two vocabularies and reorganize the hierarchy. Find out about its integration process, as well as how Gartner translated auto-classification rules from one application to the other.

Content recommendation engines have become very popular. But naïve approaches to recommendation yield poor results. This session showcases the unique approach SAGE developed to overcome this issue, based on combining text analytics with its newly minted social sciences taxonomy to effectively reveal connections between concepts across disciplines and content repositories and link journal articles, case studies, reports, podcasts, and even numerical data.

It doesn’t matter how good or complete taxonomies are if they can’t be used in enterprise systems. HBO’s metadata and taxonomy team has developed successful partnerships across the enterprise with its IT architects, developers, and project managers on ways to vet, document, get sign-off and implement taxonomies. Hear how HBO implemented a global standard to describe language-related content (subtitles, audio tracks, dubbing, and forced narratives) to accommodate a growing overseas market—contending with proprietary internal systems, siloed workflows, and a need for increased multi-language specificity when describing products and their components.

Our engaging speaker shares how to deliver insight that is targeted and personalized wherever you are. Explore how and why the search box is evolving with your organization’s needs, content services and application context, to meet the priorities of your changing workforce and communication needs.

As organizations implement the next generation of KM, traditional search efforts have now integrated recommendations, analytics, and question/answering machines. The holy grail of cognitive solutions is a combination of robust IT architectures and information models that are shaped in an agile way. Learn about the importance of knowledge graphs in providing entity-centric views on enterprise information, high-quality training data for machine learning in more cost-efficient ways, and linking text with structured data to unify enterprise information. Hear about several real-world examples, illustrating the case for Semantic AI.

How can you tell if your taxonomy is performing well, and how can you best communicate that to stakeholders? Facing extreme competition in ecommerce, Target is answering these questions and adapting its item taxonomy in response. Gain insight into Target’s process for assessing its taxonomy effectiveness after a months-long process of defining metrics, setting goals, and establishing reporting. Key lessons include the importance of quantitative measures, how to avoid subjective “guesses,” and how to talk about taxonomy with leadership.

Indeed faces the challenge of building a complete occupational taxonomy, spanning dozens of countries and languages. Retail associates are easy to classify, but what about Zamboni drivers and lighthouse keepers? Hear insights on how to keep the process of populating an iterative taxonomy at scale manageable, productive, and efficient using auto-classification with the human understanding of more than 30 taxonomists. Also hear about Indeed’s latest project operationalizing a natural-language processing tool and the challenges of cross-cultural and remote collaboration between taxonomists, developers, and the Texas and Tokyo offices.

Developing an auto-categorization system using a taxonomy requires rigorous testing to evaluate the tags that the system assigns. Testing requires not only a substantial number of sample documents to produce an accurate result set for reviewing, it also requires coordinating the reviews of the tagging output by subject matter experts. Creating a strategy for testing the auto-categorization system will allow you to coordinate the review efforts that will fine-tune your auto-categorization system over time. Leidos explains how you can create a testing strategy that can be repeated throughout the development of the auto-categorization model. Learn about setting baselines for tagging accuracy, choosing sample documents, maintaining records of results, reviewing those results, and managing an on-going testing strategy.

Controlled vocabularies can be used for more than classifying content. Lehnert and Sweeney present case studies in which taxonomies and ontologies are used for managing projects, product information, product development using the Agile software development lifecycle, and lead and customer tracking. Though not designed specifically for these use cases, users have stretched the use of the software and vocabulary structures. Learn what worked well, what did not, and how pushing the use cases for ontologies helps reveal functional needs and product improvement.

What do a consumer goods manufacturer and a credit insurance group have in common? Both are subject to a variety of risks which, if not detected, may dramatically impact their operations and bottom lines. Delve into the challenges of putting together a semantic, technology-based business solution that monitors and reacts to a large amount of consumer feedback in real time, providing insights on consumer product quality. Hear how this approach assists credit risk analysts in the early detection of signals and events affecting companies’ solvency to anticipate default risks of targeted companies. Walk through this journey to solve real-world problems with business intelligence solutions based on semantic data and technologies.

The ability to access and share linked data across a variety of KOS-based technologies provides a tremendous opportunity to expand the sphere of knowledge that each of us has access to beyond our own domains and expertise. In this talk we review the principals of how one accesses linked data using SPARQL queries across one or more LD sources to identify concepts and properties which may then be appended to internally managed data. We also talk about the use of URIs (uniform resource identifiers) that act as key identifiers for content and allow one to publish taxonomies and thesauri, or segments of them, to be accessed to by other linked data users.

BARTOC.org is the most comprehensive terminology registry for knowledge organization systems (KOS) curated by an international group of experts. It offers a search interface for currently more than 2,700 vocabularies based on their metadata. The Skosmos Browser provides a user interface for SPARQL queries for SKOS concepts and terms and enables full-text search across included SKOS vocabularies simultaneously. Hear more about BARTOC.org and what makes it a valuable resource for indexing, information retrieval, and taxonomy development.

Sharing knowledge for enterprise success requires entrepreneurial skills, new ways of thinking and operating, continuous learning, and change. There are many new tools available to help, but it is the people and the culture of an organization that determines its ultimate success. Wilkinson interviewed 200 of today’s top entrepreneurs, including the founders of Airbnb, LinkedIn, eBay, PayPal, Yelp, Dropbox, Tesla Motors, SpaceX, Chipotle, Under Armour, Spanx, Jetblue, and Revolution Foods, to distill what it takes to go from startup to scale in our rapidly changing economy. As leaders reinvent their approaches to digital transformation for organization survival in this economy, they can learn these fundamental skills, practice them, and pass them on. Join our accomplished researcher and speaker as she shares her framework and provides ways to master the skills that underlie entrepreneurial success.

The information industry is abuzz with excitement about emerging and maturing technologies often grouped under the ubiquitous yet nebulous rubric of AI, or with nascent disruptive technologies such as blockchain, which may be poised to migrate from the confines of cryptocurrencies and cultivate business applications in other diverse activities such as records management. The growth of big data applications has produced a tectonic shift across the information landscape, challenging enterprises to rationalize the collision between information science and data science. Within some industries linked data, ontologies and the semantic web have had a major impact, whereas in other industries these technologies are perceived as academic or over-engineered. Clarke surveys industry trends before honing-in on common practical applications and tools that taxonomists and others need to do their job.

Seneca once noted, “Our plans miscarry because they have no aim,” and yet today, people are building taxonomies without a purpose. Fortunately, we’re seeing resurgence in enterprise information architecture as executives seek to make sense of the mess left behind by a “Move fast and break things” mindset. Is the intent to boost customer satisfaction with more powerful search, browse, and personalization? Is the goal to improve staff productivity by creating infrastructure, standards, and a governance model to align vocabularies and processes? Morville draws upon his experiences with ecommerce and enterprise information architecture projects to illustrate how to plan with purpose, holistically and inclusively.

Hear how Ubisoft developed a strategy to define a new common corporate taxonomy from disparate departmental vocabularies such as KM, MDM and HR. The team defined a minimum metadata standard for high-value content and a method to systematically apply the new taxonomy to high-value content. Discover how they manage and provide access to these taxonomies “as-a-service” in a globally distributed environment and plan to go further in developing a corporate ontology for content auto-classification and recommendation.

The Associated Press (AP) has a robust hierarchical taxonomy built for news, covering more than 230,000 topics, events, people, organizations, and places, as well as an auto-classification system that applies our taxonomy to content. Given the multitude of tags that make it on to a news article, which are the most important? Should the salience of each type of tag be measured the same way? Hear how AP has approached the problem of news classification relevance, ranging from statistical methods for topical multi-label classification and entity salience to mapping of our taxonomy to basic categories.

For the past 3 years, Claussen has worked on a cognitive computing project utilizing IBM’s Watson question-answering capability. The goal was to develop a global data privacy research product across different jurisdictions, legal systems, and customer types. What began as a standard taxonomy exercise developed into a much broader project, for which her skills in content organization, taxonomies, and metadata provided the foundation. Hear concrete recommendations for a taxonomist’s successful involvement in a cognitive computing project.

Organizations across industries are deploying semantically enriched search solutions for unified access to information stored in different repositories, formats, and with differing metadata. This talk describes case studies that demonstrate a practical, scalable process to achieve semantic search enrichment. It employs auto-classification based on machine learning and taxonomies with humans in the loop to efficiently and accurately tag content. This hybrid approach combines the best capabilities of human experts and AI (machine learning) in an integrated, governed process.

Chatbots and other AI infrastructures are dependent on the quality of the questions they receive and the quality of the datasets they are utilizing. Taxonomies and ontologies are essential components in creating high-quality questions and datasets for the chatbot to use. Hear about five different ways Factor used ontology to support a new AI chatbot in question refinement and the information retrieval processes as well as the ways that the requirements of the chatbot influenced the design of the ontology.

As voice command interfaces become more accurate and useful, how does taxonomy play a role in the further development of the invisible interface? How do taxonomy and use of metadata benefit from improved AI? How do taxonomists rethink the user functionality and cognitive load? The answer may come from the worlds of linguistics and metadata. Doane reviews current knowledge and proposes initial solutions for using a taxonomy in the new age of the voice command interface, including some real-world examples of building taxonomies in voice command software.

Interested in industry trends? Stymied by a taxonomy design challenge at work? Bring your toughest, crunchiest taxonomy issues and challenges to our panel of seasoned, full-time taxonomists, who compete to answer your questions with insight, entertainment, and perhaps even controversy! The best questions (as voted by the audience) will bring home prizes!