I am pleased to announce the publication of the 4th edition of my book, Information Retrieval: A Biomedical & Health Perspective. Published by Springer, the book updates the content, methods, results, and research in the use of search systems for knowledge-based biomedical and health information.
I am gratified to be active in a number of areas of research in biomedical and health informatics, but my original and still most active area is information retrieval (IR), also sometimes called search. The appeal of getting information from a computer by entering a query or question held appeal to me from early times, including when I was dabbling with computers in medical school and residency in the mid-1980s. Upon entering formal training in the field in my postdoctoral fellowship in 1987, this appeal persisted, even as the thrust of research in the field was still focused on the first era of artificial intelligence.
My introduction to the field came through a monograph by Prof. Bruce Croft, which then led me to discover the work of Prof. Gerard Salton. I had the opportunity to meet Salton when he visited Harvard University during my postdoctoral informatics fellowship there. Salton literally invented the IR field and it is unfortunate that he passed away in 1995 before he could see the true impact of his work on IR systems in the world. The approach of Salton and his legions of graduate students he trained in “automated” IR was quite different than the main biomedical focus in the 1980s and 1990s, which was the set-based Boolean retrieval approaches used to search MEDLINE. My earliest work attempted to marry the automated approaches to the controlled vocabularies being developed and collated in the National Library of Medicine (NLM) Unified Medical Language System (UMLS) Metathesaurus.
Another early interest of mine in IR concerned evaluation of systems and users. A perspective of evaluation has guided a great deal of my informatics research, based on the premise that what we do, whether building systems or advocating their use by people, should be studied for its value to human health and healthcare. My foray into IR research led me to recognize the importance of the relevance-based metrics of recall, precision, and their aggregated combinations, but I also felt dissatisfied that they did not evaluate the entire IR experience, especially for users. I was fortunate to be able to attend the very first Text Retrieval Conference (TREC), and then become involved in organizing a number of its tracks in subsequent years.
I never would have imagined in my early days that we would be able to carry around the Internet - and access to the world’s knowledge - in our pockets via mobile devices. I could not fathom that essentially all scientific publishing would become electronic, and that it would include not only articles but the underlying data. I also would never have imagined that searching would be so ubiquitous by all Internet users, or that the name of a search engine would become a verb (Googling).
The world of IR has certainly changed. The basic task of “ad hoc” searching is pretty well a solved problem. There are still, however, challenging problems in IR to solve, such as some of those on which I currently work.
Both the eBook and hardcover editions of the new edition are now available through Springer and Amazon. If you or your institution have access to SpringerLink, the eBook version can be accessed there.