We are always amazed and impressed at the inventiveness of Linguamatics customers, in their applications of text analytics to address their information challenges. Our annual Linguamatics Spring Users Conference showcased some examples of their innovation, with presentations on text mining used for patent analytics, chemical pharmacokinetics and pharmacodynamics data extraction, creating value from legacy safety reports, and integrating open source tools for advanced entity recognition. We had a record-breaking number of attendees this year, representing over 20 organizations, ranging from our most experienced I2E users to text mining novices.
Patent analytics featured in two of the presentations, demonstrating the value of NLP in extracting critical information from obtuse and lengthy patent documents. Julia Heinrich (Senior Patent Analyst, Biotechnology at Bristol-Myers Squibb, Princeton, New Jersey) asked the question: “Can the infoglut of biotech patent publications be quickly reviewed to enable timely business decisions?”. She admirably demonstrated that with smart use of I2E’s NLP queries, BMS have been able to search the patent body for information on antibody-drug conjugates and convert “unstructured data” into user-friendly, analysis-ready data sets. Thorsten Schweikardt (Senior Information Scientist, Boehringer Ingelheim) gave an overview of workflows developed using KNIME to create patent landscapes for specific disease areas, target identification, and discovery of tool compounds.
Wendy Cornell (former head of the Merck Proprietary Information and Knowledge Management Group), like Julia Heinrich, flew over from the US for the meeting. Wendy presented on the automated extraction of conclusions from internal preclinical safety reports using I2E. These internal safety assessment reports contain a wealth of historical data around safety and toxicity of developmental compounds, and many pharma organizations have sought ways to gain benefits from these valuable legacy documents. Wendy’s group developed a strategy to access Documentum-based safety assessment reports, and were able to pull out histopath findings, organ toxicities, haematological and blood biochemistry results, even pulling out toxicokinetic parameters from tabular sections. Three use cases were presented, showing significant business impact within the Safety Assessment organization.
Linguamatics’ speakers gave an update on future innovations in the I2E roadmap, the new features in I2E 4.3, and the software’s applications in the life sciences and healthcare. Guy Singh showed how I2E 4.3’s Connected Data Technology allows users to exploit big data better no matter where the data are located (on premise, on the cloud), whatever structure they have, and doing this at speed, with digestible results. Phil Hastings gave a brief overview of Linguamatics I2E in Healthcare; and NLP Specialist James Cormack took us through Linguamatics’ approach and results for our submission to the i2b2 2014 Cardiac Risk Factors challenge. You can find out more about what we’re doing in healthcare via this short video.
We heard from a few of our partners in 5-minute lightning round presentations: IFI Claims Patent Services, ChemAxon, Copyright Clearance Center, Thomson Reuters and KNIME discussed their solutions and how they integrate with Linguamatics I2E.
In addition to the presentations, the Linguamatics Spring Users Conference provided opportunities for hands-on training, with workshops aimed at different levels of text mining experience. And of course, there was plenty of time for networking and idea sharing. Our evening events were hosted in the Old Combination Room at Corpus Christi College and the Pembroke College Old Library. We enjoyed beautiful, warm spring evenings at two of Cambridge University’s oldest colleges. One delegate remarked ‘It’s so nice to be shown hidden Cambridge treasures like these, which we would never know about if it wasn’t for the events at the Linguamatics conference.’
The whole event was a great success that brought together the text mining community from across Europe (and across the pond!).
Presentations which have been approved to share are available on I2Edia and by email request.
Thanks to everyone who attended and contributed to the Linguamatics Spring Users Conference 2015, we look forward to seeing you in October at the Text Mining Summit in Newport, RI or in Cambridge, UK next year.