Programme 2012 – Overview
Programme as at January 2012. Not necessarily in final speaker order
Monday 16 April 2012
09:00. Start of Meeting
Roger Bradford - Agilex, USA
Getting beyond Keywords: Using Conceptual Representations of Text in Search, Analysis and Discovery
Describes how the use of conceptual representations of text can aid users in finding and using information. The emphasis of the discussion is on new developments in semantic processing, especially in the area of discovery applications.
David Milward - Chief Technology Officer, Linguamatics, UK
Making Knowledge Discoverable: The Role of Agile Text Mining
Explores the role of agile text mining in establishing text mining as a mainstream technology. Agile text mining provides a comprehensive and flexible toolbox for extracting information from millions of documents, incorporating linguistic processing, regular expression matching, chemical structure searching and large-scale terminologies . It also allows users to pose their own questions and get answers back in real time. This enables rapid turn around of projects, with the chance to involve stakeholders early on.
** Vendor Short Presentations: Search Technology * Minesoft
Daniel Mayer - TEMIS, France
From Discovery to Delivery : The Emergence of Mainstream Applications of Semantic Content Enrichment Tools in the Enterprise
Demonstrates how some of the leading Content Management Systems can now take advantage of off-the-shelf semantic content enrichment features, and outlines the key benefit areas that enterprises can expect from such integrations now and in the near future.
10:45-11.15 Break & Exhibition
Bob Stembridge - Thomson-Reuters, UK
Bringing Together Tools and Techniques for Patent Analysis: The Harmony of the Spheres
Discusses with examples how patent analysis tools and techniques can be used together effectively to provide technological and competitive intelligence for critical business decision support.
** Vendor Short Presentations: Digital Science * LexisNexis
Jay Ven Eman and Marjorie Hlava - Information Access, USA
Text Mining, Term Mining and Visualization - Improving the Impact of Scholarly Publishing
Describes “term mining” and how it differs from text mining. While the two approaches are not mutually exclusive, term mining can provide unique insights into large bodies of information with a degree of precision and persistence not possible in alternative approaches.
Siaw Ling Lo - Nanyang Polytechnic, Singapore
Mining Opinion from Twitter
An influential index has been formulated to extract lists of Twitter users who are influential in a particular domain. This index together with sentiment analysis has shown promise to filter relevant information from the vast volume of contents. The algorithms developed correspond with the various events and is suitable for product and brand monitoring on Twitter.
13:00-14:30. Lunch & Exhibition
Panel Session - Chaired and Animated by Susan Feldman - IDC, USA
Two industry user practitioners from two large companies -- Wolfgang Thielemann, Head of Information Retrieval and Analysis at Bayer Pharma, and Heinz-Gerd Kneip, Head of Specialty Chemicals and Biotechnology Information at BASF -- discuss with Sue Feldman and the audience information tools they have chosen, and how their choices were made and evaluated.
** Vendor Short Presentations: Lighthouse IP Group * Intellixir * Linguamatics
Laurent Hill - Questel, France
Dealing with Large Data Volumes in Statistical Analysis and Text Mining Tools
Many tools already exist for patent analysis and text mining that can deal with up to a few 10,000s of patents. However some landscaping analyses require the analysis of hundreds of thousands, if not millions of documents. In this presentation we look at new techniques allowing analysts to process up to millions of documents, both for classical business graphics and for more sophisticated text mining techniques. Real use cases in large organisations will show how these new tools can support better patent-based decision making.
16:20-16:50 Break & Exhibition
** Vendor Short Presentations: Questel * STI Centre, Moscow * Dotmatics
Jon Goldhill - Zanran, UK
What is the future of data search?
Data search can be discussed from various angles:
- the method of access (search vs specialist repositories)
- the type of data (structured vs un-structured)
- paid for or free data
- the underlying web technology that supports it (semantic web...)
This presentation reviews these various issues and attempts to answer what data search will look like in the future.
Patrick Beaucamp, - Founder and CEO, BPM-Conseil, France
Open Source Platform & Cloud Platform for Information Analysis
Examines how and why you can use the Open Source BI Platform to make your data available for everyone, with an overview of available Open Source BI platforms and of the kind of information that can be managed and made accessible.
18:15-19:30
Conference Mixer Cocktail
-- END OF DAY ONE --
Tuesday 17 April 2012
09:00 Start of Meeting
Joseph Parry - Cambridge Intelligence, UK
Visualization for Decision-Makers
How should we communicate the results of our analysis to decision-makers? This presentation argues that visualizations and infographics can play an important role, not only for analytical processes of data analysts, but also for explaining our analytical results to decision-makers at the highest of levels. Some care must be taken to avoid various common pitfalls when designing such visuals: the presentation will cover bad examples as well as good in order to uncover design guidelines and practical advice for those wishing to pursue a more visual approach.
Nils C. Newman (Search Technology, USA), Ismael Rafols (University of Sussex, England), Alan L. Porter (Georgia Tech, USA)
Patent Overlay Mapping
We are now developing “patent overlay mapping” to aid in visualizing topical concentrations in observed patent activity. Others have explored such mapping, but we believe our approach offers advantages in both production and interpretation. To produce the maps we have devised a multi-tier treatment of International Patent Classification (IPC) categories to achieve more commensurate groupings of patent activity. We apply a thesaurus to associate a given set of IPCs with these categories and then overlay the activity over 466 nodes categorised into 35 topic areas.
Patrice François - Associate Director, Digimind, France
Leveraging the Power of Visualization for Real Time Analysis
Intelligence managers have to cope with increasing volumes of demands from their internal clients, and an overwhelming amount of data. We demonstrate how the latest visualization technologies can help intelligence managers identify more quickly the potential threats and opportunities hidden in the data.
10:30-11:00 Break & Exhibition
Anton Heijs - CEO, Treparel Information Solutions, The Netherlands
Big Data Analytics and the Importance of Visualization
Big data analytics means that structured (table) and unstructured (text) data can be analysed in combination. Using data mining and text mining in combination with the appropriate visualisation techniques is powerful. Big data analytics is also about capturing the context of data in the analysis. This can lead to the analysis of patent data together with research/legal and news data. The requirements and constraints of such an approach will be discussed together with examples.
Gregory Grefenstette – Chief Science Officer, Exalead, France
Merging Information from Structured and Unstructured Information Sources in Search Based Applications
Merging information means typing entities in the same semantic space, recognising equality, and recognising relationships between things. Databases, metadata and text all have different ways of encoding this semantic information. Search Based Applications can now merge these different information sources into intuitively usable and agile platforms. Search Based Applications are possible because of advances in natural language processing, and in search technology which can handle rich semantic features (entities, numbers, dates, GPS, hierarchical categories, affect analysis ...). We describe the semantic processes that allow information to be merged from different sources, and give examples of a number of deployed Search Based Applications.
Susan Feldman - IDC, USA
Towards Unified Access Systems for Data Exploration
The combination of too much information and increasing IT complexity makes it difficult for businesses and IT departments alike to understand and react to customers, trends or competition in a timely manner. Information is scattered across transactional systems, email archives, call centre records, social media and the Internet. Gathering and compiling it quickly — and then analysing it and acting on it -- has become a daunting task. What is needed is a single point of information access and management that can quickly gather, process, find and analyse information from all sources: structured or unstructured. Unified access systems that combine the features of BI and search software are now beginning to supplant these earlier systems.
Peter Noerr - MuseGlobal, USA
A Tale of Unified Access: Content at your Fingertips?
Presents a cautionary tale of the pitfalls and problems on the road to making unified and easy access to information. Various methodologies have been tried and currently there are a variety of approaches to enable access to information from the many places and ways in which it exists. Current trends are to move towards “one stop shopping” for information access, and this is coming known as unified access. There is good and ill in this movement and we discuss both in an intent to show the progress that has been made and the work still to be done.
13.00-14:15. Lunch & Exhibition
Martin Hofmann-Apitius - Fraunhofer Institute for Algorithms and Scientific Computing (SCAI), Germany
From (Text) Mining to Models: Applying Large-Scale Text Mining on Patents and Electronic Patient Records
On the side of the semantic representation of pharmaceutical knowledge, the scientific community is still lagging behind the proprietary, commercial sector. We have started to work on a public, open, free accessible pharma ontology that is composed of smaller, high-resolution ontologies covering well-defined topics such as “protein-ligand interaction”.
Stuart McLean - LexisNexis, USA
Patent Prior-Art Searching with Latent Semantic Analysis
Semantic Search recognises concepts within users' search input and offers alternative phrases to expand the search and raise related documents to the top of the search results. Instead of using dictionaries and thesauri to perform term expansion, Semantic Search returns words that have appeared in patents and Non-Patent Literature (NPL) in blocks of text related to the identified concepts. In this way, Semantic Search can see through obfuscating language so common in patents and seek out those related patents missed in the examination process.
Alexander (Alex) Butler - Executive Vice President, IPVision, USA
Expert System Driven Insights into Patent Quality and Competitive Positions
Discusses how Patent Claims Analytics are used by leading patent holders and legal experts as a business intelligence tool to evaluate the quality and strength of patent positions, the quality returned from patent investment and counsel, and technology leaders / followers within a specific technologies. Examples will demonstrate patent breadth and quality trends for representative large patent holders, recent, well-publicised patent portfolio transactions, and a case study where Patent Claims Analytics and complimentary analyses lead to an agreed 75% price modification in a significant patent portfolio transaction.
15:45-16:00 Short Break
Kalpana Ravi (Philips India – Intellectual Property and Standards), B. Ravi Kiran (LIGM, ESIEE, Université Paris Est, France), Mannan Bakthavathsalu (Philips India – Intellectual Property and Standards)
Automatic Query Re-Ranking in a Patent Database by Local Frequency and Adjacency Distribution
When results are ranked by databases for a given query, they do not always rank relevant patents at the top and in order later on. The problem with current approaches is the lack of control with the resolution in the frequency count of the literals and the proximity or adjacency of search literals with respect to each other. Our tool gives a new method by which users of a database are able to re-score or re-rank the patent based on our proposed heuristic.
2012 Conference Final Endnote
Willem Geert Lagemaat - CEO, Lighthouse IP Group, The Netherlands
What Information Tools can you Imagine for the Future?
In the rapid developing world of new features that are out and open on the Internet, many techniques are becoming available in the information retrieval world. However, many users still like to save data in Excel format and then use Excel to create graphs. This illustrates our inability to look further and think of other ways of working with large quantities of data. If the question were to be asked to you as a user, would you then be able to say what tools your imagination projects for you in the future? How creative can you be in supporting industry in the direction in which we should all be going?
2012 Conference Highlights
Pierre Buffet (ex-Questel) and Robert Stembridge (Thomson-Reuters), two information industry experts, give their personal reactions to the highlights of the meeting and reflect on what they have learned over the two days.
END OF 2012 Conference at approximately 17:30
