Tuesday 26 Oct

Several new sophisticated tools have been recently released or are currently being developed. Most developers are claiming that they are able to accurately rank and value patents. Can we trust them? What about their impact on the Chicago patent stock exchange? What is the ideal rating tool? Where are we going? 

The primary purpose of the Pistoia Alliance is to streamline pre-competitive elements of the life science research workflow by the specification of common business terms, vocabularies, relationships and processes to underpin the delivery and provision of information services.
   Every Life Science company, software and services supplier is challenged by the technical inter-conversion, collation and interpretation of life science discovery data and as such, there is a vast amount of duplication, conversion and testing that could be reduced if a common foundation of data standards, vocabularies and services could be promoted and ideally agreed within a non-proprietary and pre-competitive framework. This would allow inter-operability between a traditionally diverse set of technologies and approaches to benefit the life science sector. Through global collaboration, this pragmatic community will derive and instantiate open service standards and specifications that can be built on to provision an open information framework while continuing to work with existing standards groups.
   We describe current progress in our working groups, learnings and how companies, academics and others can participate in this approach.

Henry Chesbrough first coined the phrase “open innovation” to describe a new business paradigm for technology transfer through collaboration in his 2003 book “Open innovation: the new imperative for creating and profiting from technology”. The central idea behind open innovation is that in a world of widely distributed knowledge, companies cannot afford to rely entirely on their own research, but should instead buy or license processes or inventions (eg, patents) from other companies. In addition, internal inventions not being used in a firm's business should be taken outside the company (eg, through licensing, joint ventures and spin offs).
  Open innovation highlights the need for better mechanisms to manage collaboration and sharing of information and how best to manage the time shift effect of researchers around the globe working on the same issue following the sun. This presentation discusses the issues around ownership of ideas in an open innovation situation, the strategic use of IP in this environment and managing the information flow between collaborating partners.

Agile text mining approaches have been shown to be valuable for answering questions and extracting knowledge across life sciences research and discovery.  This presentation focuses on particular techniques relevant to patent mining. These include use of linguistics for filtering noise and for terminology development, flexible categorisation and the ability to plug in domain-specific thesauri.  By adopting these techniques we show how it is possible to accelerate the systematic and comprehensive analysis of a series of patents relating to a compound family, and combine the results with insights from other literature sources such as an organisation's internal documents or external scientific literature.

The presentation will describe some of the experiences BASF Group Information Center experienced when applying semantic technologies to (patent) information.
We will discuss use cases for:

• vectorizing to cluster and categorize documents,
• extraction of meaningful entities from text and
• developing, maintaining and using ontologies to enhance searching and analyzing.

Semantic technologies are widely used in pharma, medicine and biology - but in the area of chemistry we have to uncover and solve a lot of chemistry- and patent- specific issues - here we talk about hard stuff.

Traditional keyword-based document retrieval systems such as PubMed and ClinicalTrials.gov dominate the area of biomedical information search. They are easy to use, simple and fast. However, these existing search systems are poorly suited for finding biological relationships in text as they merely return a collection of documents matching user queries which need to be further read by a user.
   We have developed a novel Natural Language Processing algorithm (ConceptScan) which decomposes sentences into a set of primitive semantic relationships. We show that in combination with entity detection, it can be used as a text indexing engine in a search system which allows finding specific biological and clinical information. We demonstrate how ConceptScan may be used in several different use cases in the biomedical domain to improve the search experience. These applications include  

• a) a semantic search solution which accepts natural language queries to return focused results,
• b) a means of finding materials and methods within scientific literature (currently a time-consuming procedure) by understanding the context of a search,and 
• c) a capability for extracting clinical information from free-text sources. In this application we have designed a data model for presenting clinical study information which includes the details on study design, targeted disease, patient group, treatment strategy and drug dose, measured clinical parameters, and numerical clinical outcome results.

With the advent of Web 2.0 -based platforms, open access journals and
other web resources that provide a wealth of information to the end
user are proliferating. The "new age" publishing consumers'
preferences are fast changing. It is becoming essential that we gather
information from multiple heterogeneous sources, and map the content
to ontologies to enable semantic searches that can integrate the
diverse content from the web with publishers' and data providers' own
content.
  This presentation discusses research and technologies designed to
derive significant value from freely available information and to
gather business intelligence reports from published literature and
patents.
 

FIZ Frankfurter Innovationszentrum Biotechnologie GmbH is a competence centre in the area of Inflammation and Proteomics research. A fundamental aspect of any research is to understand and keep track of progress made by peer groups in terms of scientific discoveries and intellectual property. Patents form a definitive source of this information. Annually, thousands of patents are published in the space of Inflammation from a Biological, Pharmacological, and Chemical perspective. At first glance, the problem of finding relevant patents of interest and then organising the patent information into a format which is easily analysed, stored and efficiently retrieved seems to be difficult and chaotic as there are no patterns by which a process can be defined. Furthermore, patent data are highly fragmented and non-standardised specially in case of chemical structures.
   This presentation discusses a hybrid approach, wherein a Machine Learning based text-extraction software is coupled with assisted expert annotations by human editors. A third-party Machine Learning software system is used in the first stage wherein the patent batch is classified based on keywords, segmented and converted into standardised format.
   The third-party software then uses a proprietary, heuristic based, learning algorithm to extract relevant data from the segments. Since it is well known that any automated approach cannot be 100% accurate, in this step the software is assisted by a team of expert human editors who analyse the extracted and segmented data and perform necessary corrections, if any. In the third step, the software then pushes each segment to a team of expert human editors who analyse the segment, extract information relevant to area of research at FIZ and store the information in our internal databases.

Conference Dinner (Heurigen) - sponsored by Dr. Haxel Congress and Event Management GmbH