Compilation of Email Discussion on the topic of Natural Language Processing CSCE976 Spring'2002 ========================== Date: Mon, 25 Feb 2002 11:04:18 -0600 (CST) From: Lin Xu To: cse976-ml@cse.unl.edu comments about NLP: 1. I think since there are too much different style of writing. It's really hard to find a common NLP to solve all problems. It may be better to build to different NLP for different area. For example, some special NLP for math only, another for cse only. 2. Also I think since different person has different writing style. So, We can build some program which can adapt whose style. With more interactive of person and program, those program can works better and better. It will automatically recognize what is your really meaning is. Lin .'* :`...' `.,' ' `. ' .**. ; ; ': ` ``:`****,' .' : ..::. ``**":.'' `. .: `: ; `,' : For everyone you love and love you! `: ` : ; : : : ; : : : .: : : :..,' ``::. Lin Xu `....:..' ..:;'' .: . ...:::: ,'''''``::::::: `:::: `::. `:: ======================================================================= Date: Wed, 27 Feb 2002 20:50:10 -0600 (CST) From: "Berthe Y. Choueiry" To: cse976-ml@cse.unl.edu CC: choueiry@lucca.unl.edu Subject: NLP discussion Xu Lin wrote: > It may be better to build to different NLP for different area. For > example, some special NLP for math only, another for cse only. yes, I think this is the way the activities in the area are going. Progress in focused, specialized domains. Note, that this is a natural (and healthy) approach to science and problem solving in general: when we cannot solve the general problem because it is too difficult, we focus on more specialized areas as a simplification strategy to intractability. > 2. Also I think since different person has different writing > style. So, We can build some program which can adapt whose > style. With more interactive of person and program, those program > can works better and better. It will automatically recognize what is > your really meaning is. Agree. User modeling is a (relatively new) area where an automates agent learns patterns of the behavior of the users and builds profiles for them. It is common in E-commerce and Internet shopping (e.g., Amazon remembers what you ordered and advises you which books you may want to check because they match your interests). I have no idea how the NLP people include user models, speech acts, belief modeling, etc. in their research. You raise here a good question.. Anyone has any idea?? Thanks, -Berthe Y. Choueiry choueiry@cse.unl.edu =======================================================================