Wednesday 26 Oct

Search Based Applications (SBAs) use search engine technology as a backbone to aggregate  intelligently large volumes of unstructured data (like Web pages) and structured data (like database content), and to make those data available in a highly contextual, quasi real-time manner to a wide base of users for a varied range of purposes. We present the general convergences under way in search and database disciplines, convergences that make SBAs possible, and which serve as harbingers of information management paradigms and technologies to come. The presentation will be illustrated with a number of already implemented SBA case studies.

Information services have been in a constant state of change over the past decade. Where in the past the information centre was the only gateway to information, this has now changed and the centre needs to shift focus to stay relevant.

Based on his own experience, Dennie Heye will explain how he achieved this at Shell:

• He is embedded in a department to create and execute an information / knowledge management strategy. Information is an asset and should be managed as such. He has developed a clear, measurable strategy to increase maturity for information and knowledge management.
• Development of SharePoint solutions: using the SharePoint features, knowledge of business problems and stakeholder management, Heye has developed fit-for-purpose information management solutions for portfolio management and team working.
• Change agent for virtual collaboration: Heye works with teams to collaborate effectively and efficiently and to manage information across regions, business and time zones using tools such as Wikis, blogs, SharePoint and videoconferencing.

Dennie Heye is convinced that information professionals have a unique set of skills and experience that are relevant in a constantly changing and demanding world.

During 2010, the Danish Patent and Trademark Office (DKPTO) fundamentally changed the structure of its internal resources for search and examination. As expected, through budding and branching, during its first ten years, the first generation intranet grew into a shape that did not match the current work flow, and a continuously growing effort was necessary to convey the information at hand to the examiners using it. Consequently, in 2009 the DKPTO decided to re-design an intranet that was to contain its resources for search and examination. The new intranet should match the work flow to ensure high quality requirements, while being as transparent and convenient as possible for users. Also, an intranet that matched the most up-to-date ways to search was considered a key to success, and speedy delivery was a factor. Therefore, an in-house solution was chosen. Totally, 300 man-hours were allocated for the job. Starting point was an analysis and sorting of existing material, and corresponding needs for documentation. A logical architecture that is ultimately simple, transparent and maintainable was chosen and adhered to. Internet links are centralised in one independent library. Maintenance documentation is arranged so editors are expendable. Quality of search has improved since the reform.

No two patent searches are ever the same and it becomes more complex with medicinal chemistry patents. The search complexity lies in not having used systematic terminology. Text searching is essentially a game of broadening and narrowing. It is important to have a strategic approach to text searching instead of haphazardly entering keywords into a search engine.

Search engines do not understand scientific synonyms and objects embedded in the primary literature which leads to loss of information that is locked in the data silos of pharma/biotech industry. This presentation will discuss a search implementation on a platform where we initially employed IPC code classification to filter and restrict the search space for eliminating false positives. We then employed a dictionary with synonyms, acronyms and textual variance for the target names. A proper modulation of chemical keywords and protein names was used to identify the relevant patents outside the patent classification. Our system follows the “all but only” principle - it retrieves all the associated patents but supplies only the patents relevant to the query, enabling us to minimise the false positives and false negatives while retrieving the relevant patents within the acceptable level of precision and recall. The talk will focus on showing the importance of indexing and classification while searching and analysing patents.

The number of reported targets for marketed drugs is about 500 and the potentially “druggable” targets fall in the range of 6,000 to 10,000. Retrieval of a complete list of patents for these targets would allow (i) scientists to build actionable thematic databases that would help in designing better experiments, (ii) patent analysts to perform authoritative prior art searches, and (iii) business analysts to create technology from diligence reports. In this presentation, samples of disease-centric, target-centric and assignee-centric thematic databases will be discussed. Such an interoperable database will enable the concepts of network pharmacology, facilitate Chemgenomics ideas as well as aid in repurposing of drugs.

Patent filing in Asian languages has seen an upward trend for the past few years. Japan and Korea have been patent powerhouses already for a few years. Since 2001, China has also seen a significant increase in patent filing resulting in more than a million Chinese patents (patents, utility models and designs) filed in 2010 alone, and plans to double this by 2015. Consequently, the patent databases for each of these Asian countries have become large repositories of patented information. These databases could serve as an important source of prior art. Further, these databases are also crucial to stay on top of the latest technology trends in an ever-evolving global marketplace. However, it may not be easy to traverse these patent databases owing to a language barrier. Machine translations could help in a superficial analysis but cannot be relied on for detailed analysis. There are also sources which provide English language search in annotated/indexed English text. However, it is essential to explore native language searches and analysis in Japanese, Chinese and Korean languages. This presentation i) examines the effectiveness of information sources in native languages, ii) compares search outputs of native language searches against English, and iii) provides the best practices in carrying out a truly global prior art search and patent analysis.

LexisNexis TotalPatent combines smart search and compare functionality with comprehensive patent content to provide information and intellectual property specialist with a clear view on what is happening in the field of Innovation and Intellectual Property.

In this workshop we will show the latest IP content developments at LexisNexis such as the extension of the Indian patent collection, quality and timeliness of the Japanese and Chinese MachineTranslations. We will also show you new functionality to easily compare and visualize multiple search result sets and the Semantic Search functionality. Our experts are available to answer any questions you may have and you can see demo’s at your convenience.