David Haussler

Digital Library

ACM AAAI Allen Newell Award

USA - 2003

citation

"For contributions bridging computer science and biology through research in computational learning theory, computational biology, and bioinformatics leading to major influences on the understanding of biological macromolecules and the investigation of the human genome."

By staying focused on the big scientific questions and remaining unencumbered by the norms of any well-established and narrow method of inquiry, David Haussler has forged deep and fruitful scientific interactions between computer scientists and molecular biologists and played a leading role in developing the new field of computational biology. His contributions include:

  • Pioneering the use of hidden Markov Models, Stochastic Context-Free Grammars, and discriminative kernel methods for analyzing DNA, RNA, and protein sequences, with specific applications to finding protein-coding and RNA genes in the human and other genomes and predicting three-dimensional protein folding based on similarities to other proteins with known structures;
  • Building an algorithm to assemble the first public working draft of the human genome (from more than 600,000 pieces) for the International Human Genome Sequencing Consortium and posting it on the World Wide Web, helping to keep it publicly available for free and unrestricted scientific research;
  • Developing interactive web-based browsers that display analyzed and annotated genome sequences of human beings and other organisms; this "microscope" that allows researchers to view the genome and its bioinformatics analysis at any scale from a full chromosome down to an individual nucleotide is used extensively in biomedical research (currently logging ~140,000 page requests each day from researchers worldwide)

Haussler has done significant research in machine learning, statistical decision theory, pattern recognition, and neural networks. He devised new types of learning algorithms and helped put the fields of machine learning and neural networks on a more solid theoretical foundation by proving upper and lower bounds on achievable learning rates. His current work on stochastic models of molecular evolution has introduced new mathematical models for natural adaptation.