The WS02 Guest Lectures are open to the public, and a video recording of the lecture will be archived in the CLSP library. Talks are held every Wednesday and on selected other dates throughout the course of the workshop. All lectures will be given on the Homewood Campus of Johns Hopkins University.

Click here for directions to JHU.

• Location: Shaffer Hall, Room 100
  
  
• Time: 10:30 am - 12:00 noon
  
  
• Abstract:

  A few years ago we started investigating alternatives to HMMs for shallow parsing and information extraction in text processing. I will motivate and describe the main contribution of this research, a class of models we call "conditional random fields" (CRFs). Conditional random fields offer several advantages over generative models like hidden Markov models and stochastic grammars, including natural discriminative training and the ability to relax independence assumptions in generative models. Conditional random fields also avoid a fundamental limitation of methods based on combining per-position classifiers, including our earlier effort with maximum entropy Markov models. I will discuss training methods for CRFs, and present experimental results on synthetic and natural-language data that show their advantages over HMMs and classifier-based sequence analysis methods. I will conclude with proposed extensions to parsing and comments on connections with related efforts in undirected graphical models and generalized perceptron algorithms.

  Joint work with John Lafferty, Andrew McCallum, and Fei Sha.

   

• Biography:

  Fernando Pereira is the Andrew and Debra Rachleff Professor and chair of the department of Computer and Information Science, University of Pennsylvania. He received a Ph.D. in Artificial Intelligence from the University of Edinburgh in 1982. Before joining Penn, he held industrial research and management positions at SRI International, at AT&T Labs, where he led the machine learning and information retrieval research department from September 1995 to April 2000, and most recently at WhizBang Labs, a Web information extraction company. His main research areas are computational linguistics and machine learning, and he is a main contributor to several advances in finite-state models for speech and text processing in everyday industrial use. He has 73 research publications on computational linguistics, speech recognition, machine learning and logic programming, and several issued and pending patents on speech and language processing, and on human-computer interfaces. He was elected Fellow of the American Association for Artificial Intelligence in 1991 for his contributions to computational linguistics and logic programming, and he is a past president of the Association for Computational Linguistics.