DASH

News and Events

Louhi 2015 in Lisbon

Health Informatics, SYSLAB

DSV was well represented at the 6th edition of Louhi, the International Workshop on Health Text Mining and Information Analysis, which was co-located with the EMNLP conference (Empirical Methods in Natural Language Processing) in Lisbon, Portugal. Four papers by DSV were presented at the workshop: “Adverse Drug Event classification of health records using dictionary based pre-processing and machine learning” by Stefanie Friedrich and Hercules Dalianis; “Creating a rule based system for text mining of Norwegian breast cancer pathology reports” by Rebecka Weegar and Hercules Dalianis; “Representing Clinical Notes for Adverse Drug Event Detection” by Aron Henriksson; and “Expanding a dictionary of marker words for uncertainty and negation using distributional semantics” by Alyaa Alfalahi, Maria Skeppstedt, Rickard Ahlbom, Roza Baskalayci, Aron Henriksson, Lars Asker, Carita Paradis and Andreas Kerren.

Some photos from the conference are shown below.

 

IEEE Conference on Bioinformatics and Biomedicine (BIBM) in Belfast

Health Informatics, Photo, Presentation, SYSLAB

Jing Zhao and Aron Henriksson attended the IEEE International Conference on Bioinformatics and Biomedicine (BIBM) in Belfast, UK on November 2-5, 2014. The scientific program this year consisted of 111 papers, including two regular papers from DSV. Jing presented the paper “Detecting Adverse Drug Events with Multiple Representations of Clinical Measurements”, co-authored with Aron Henriksson, Lars Asker and Henrik Boström. Aron presented the paper “Generating Features for Named Entity Recognition by Learning Prototypes in Semantic Space: The Case of De-Identifying Health Records”, co-authored with Hercules Dalianis and Stewart Kowalski.

Visiting UCSD and University of Utah

Health Informatics, Photo, Presentation, SYSLAB, Visit

Aron Henriksson spent two weeks in April at the University of California, San Diego (UCSD), continuing his ongoing collaboration with Dr. Mike Conway, who is currently at the Department of Family and Preventive Medicine. A small project was initiated that aims to develop techniques for detecting signals of adverse drug events (ADEs) in Twitter streams, in particular ADEs that are perceived to have been caused by tobacco cessation products. This project will continue over the coming months.

Sumithra Velupillai, who is currently doing a postdoc in San Diego, and Aron also visited Salt Lake City, where they gave a joint talk at the Department of Biomedical Informatics at the University of Utah, headed by Professor Wendy Chapman. The talk described their respective research efforts and that of the Clinical Text Mining group as a whole.

Both visits were very inspiring, and on our way from San Diego to Salt Lake City, we took the opportunity to take in the spectacular Grand Canyon!

Research Visit at CIMeC, University of Trento

Health Informatics, SYSLAB, Visit

Aron Henriksson visited the Center for Mind/Brain Sciences (CIMeC) at the University of Trento in Rovereto, Italy for two weeks in the first half of February. The Language, Interaction and Computation Labs (CLIC) has an ongoing 5-year ERC project to study Compositional Operations in Semantic Space (COMPOSES), led by associate professor Marco Baroni. Meetings were held with COMPOSES members to exchange ideas and identify opportunities for collaboration. A joint study was initiated, where various strategies for creating ensembles of semantic spaces are explored; their DISSECT toolkit is used to conduct the experiments. Aron also gave a talk at CLIC entitled Semantic Spaces of Clinical Text — Distributional Semantics Applied to Health Care Data. It was a very inspiring visit!

Annual Symposium of the American Medical Informatics Association (AMIA 2013) in Washington DC

Health Informatics, Presentation

Sumithra Velupillai and Aron Henriksson attended the Annual Symposium of the American Medical Informatics Association (AMIA) in Washington DC (Nov. 17-20). The AMIA symposium is the main medical informatics event in the US, with participants from all states, and around the world, with participants from all continents – there were well over 2,000 participants in total. The symposium thus provided a good overview of current challenges and ongoing research in the world of medical informatics, as well as a great opportunity for networking. Aron presented a paper, Identifying Synonymy between SNOMED Clinical Terms of Varying Length Using Distributional Analysis of Electronic Health Records, co-authored with Mike Conway (UCSD), Martin Duneld (DSV, SU) and Wendy Chapman (University of Utah). We also had a nice reunion dinner with participants in the Interlock project, a two-year (2011-2013) collaboration between Stockholm University and University of California, San Diego.

Louhi 2013: The 4th International Workshop on Health Document Text Mining and Information Analysis

Health Informatics, SYSLAB

Hercules Dalianis, Sumithra Velupillai and Aron Henriksson participated in the Louhi 2013 Workshop on Health Document Text Mining and Information Analysis. This year it was hosted by NICTA in Sydney, Australia on February 11-12. Sumithra presented the paper “Porting a Rule-based Assertion Classifier for Clinical Text from English to Swedish” by Sumithra Velupillai, Maria Skeppstedt, Maria Kvist, Danielle Mowery (University of Pittsburgh), Brian Chapman (UCSD), Hercules Dalianis and Wendy Chapman (UCSD). Aron presented the paper “Optimizing the Dimensionality of Clinical Term Spaces for Improved Diagnosis Coding Support”, co-authored with Martin Hassel. Hercules was invited to give a keynote, where he talked about detecting hospital-acquired infections from patient records and the results from the Detect-HAI project. A workshop report (in Swedish) is available on request from the author.

Research Visit at UC San Diego

Health Informatics, Photo, Presentation, SYSLAB, Visit

Maria Skeppstedt and Aron Henriksson are currently visiting the Division of Biomedical Informatics (DBMI) at the University of California, San Diego, which is located in the San Diego Supercomputer Center (SDSC). The visit will last for a month and is one of several exchanges between the Natural Language Processing (NLP) group at DBMI and the Clinical Text Mining Group at DSV that have been made possible through the Interlock project.

Last week we were invited together with Wendy Chapman to give a talk at a PACE (Predictive Analytics Center of Excellence) seminar at SDSC. Wendy first talked about NLP in general, while Maria and I talked about our respective research on NLP in the clinical domain.

ICML 2012 in Edinburgh

Health Informatics, Information Systems, SYSLAB

Last week (June 26-July 1) I attended the International Conference on Machine Learning (ICML) in Edinburgh. At the ICML workshop on Machine Learning for Clinical Data Analysis, I presented the following two papers:

  • Exploration of Adverse Drug Reactions in Semantic Vector Space Models of Clinical Text by Aron Henriksson, Maria Kvist, Martin Hassel and Hercules Dalianis
  • De-identifying health records by means of active learning by Henrik Boström and Hercules Dalianis

The workshop brought together researchers in machine learning, natural language processing and medical informatics who are trying to exploit the availability of large amounts of clinical data in order to improve disease detection, chronic disease management, design of clinical trials and other aspects of health care.