A Paper Abstract by Benjamin Grosof


Relationships Between Non-Monotonic Reasoning and Incremental Learning: Preliminary Outline of Invited Talk (Mar. 23 1993)

by Benjamin N. Grosof

Abstract: We observe that non-monotonic reasoning arises in many kinds and aspects of incremental learning, including: induction, preference-type bias, and shift-able bias; learning by being told; knowledge assimilation / integration; knowledge-base refinement / theory revision; explanation-based learning; and sensory perception. We review what the field of formal non-monotonic logical knowledge representation has to offer the subject of incremental learning. This includes: algorithmic methods for inference and updating / revision; formal properties and guarantees; and concepts such as default, model-preference, and prioritization.

In our attached paper "Representing and Reasoning with Defaults For Learning Agents", we discuss in more detail a number of knowledge representation issues revolving around the use of defaults for learning agents.


Last update: 1-8-98
Up to Benjamin Grosof's Papers page
Up to Benjamin Grosof home page
[ IBM Research home page ]

[ IBM home page | Order | Search | Contact IBM | Help | (C) | (TM) ]