The goals of my research and the Biological Knowledge Laboratory (BKL) which I head, are little changed since the lab's founding in 1989. In fact, the directions were set in the summer of 1981 (see the BKL History page). At the highest level, my goal is to understand how science works and to build systems that can extract the knowledge about research goals, methods, and results, from the scientific literature.

We have pursued these goals helped by a variety of concepts, novel ideas, tools, data collections, and most importantly, people - the people who have worked in the lab, and our many colleagues around the world we have met at meetings and corresponded with. As with all science, we build on the past; both the results of the past and the approaches to research that have been pursued in the past. Our research breaks down, approximately, across text and diagrams. Well before 2010, I have done extensive work in natural language processing (NLP). Our current NLP research revolves around building the foundations for a new generation of NLP systems. Too many NLP systems and for too long, have taken an approach we call the "standard paradigm" which started with Chomsky's work in 1957, work which continues to underlie much of research today. Our new approach breaks away and moves into new territory, both in terms of the computational infrastructure we are developing and the linguistic approach underlying it, Construction Grammar.

We can argue quite soundly that diagrams, and figures in general, can only be fully understood in conjunction with their explanatory text; the text that describes them can only be fully understood in conjunction with the figures. Our focus is on diagrams, rather than figures in general. Diagrams encode and embody knowledge using concise discrete elements. Figures are far more varied and often don't project crisp information, e.g., a photo of an animal in the complex surrounds of its habitat, or a cloud formation.

Though quantitative diagrams are the most common ones in scientific publications, many other types of diagrams are used, e.g., a phylogenetic tree, a schematic of rock strata, etc. A variety of diagrams can be seen on our modest website,

Further details can be found on our diagrams page and our natural language page.

Return to Futrelle's home page

October, 2009