Copy Link
Add to Bookmark
Report
AIList Digest Volume 5 Issue 116
AIList Digest Monday, 11 May 1987 Volume 5 : Issue 116
Today's Topics:
Seminars - Automatic Equation Derivation (SU) &
Managing Uncertainties: Prospective Reasoning (CMU) &
A Shell for Intelligent Help Systems (UPenn) &
A Computational Model of Creative Writing (UPenn) &
Speaking to a Computer (CMU) &
BB* Layered Environment for AI Systems (HP)
----------------------------------------------------------------------
Date: 27 Apr 87 1318 PDT
From: Vladimir Lifschitz <VAL@SAIL.STANFORD.EDU>
Subject: Seminar - Automatic Equation Derivation (SU)
AUTOMATIC DERIVATION OF THE
EQUATION OF MOTION OF A PENDULUM
Thursday, April 30, 4:15pm
Bldg. 160, Room 161K
Michael Beeson
(beeson@csli.stanford.edu)
San Jose State University
Some knowledge of elementary physics has been formalized in first-order
logic. The domain of discourse includes physical objects and their
relations, mathematical formulas, and the semantic relation between
formulas and objects. The knowledge in question has been written in
Prolog and is sufficient to support an automatic derivation of the
differential equation of motion of a pendulum. The inference engine
makes use of the Knuth-Bendix method and also of a symbolic computation
system for algebra and calculus. Perhaps this is the first program to
use both knowledge representation in logic and symbolic computation.
------------------------------
Date: 30 Apr 87 14:50:08 EDT
From: Patricia.Mackiewicz@isl1.ri.cmu.edu
Subject: Seminar - Managing Uncertainties: Prospective Reasoning (CMU)
AI SEMINAR
TOPIC: Managing Uncertainties: The MU System For
Prospective Reasoning
SPEAKER: Paul Cohen, University of Massachusetts, Amherst
WHEN: Tuesday, May 5, 1987, 3:30 p.m.
WHERE: Wean Hall 5409
ABSTRACT:
I will describe a style of problem solving, prospective reasoning, and
a development environment, MU, for building prospective reasoning
systems. Prospective reasoning is a form of planning in which
knowledge of the state of the world and the effects of actions is
incomplete. I will illustrate one implementation of prospective
reasoning in MU with examples from medical diagnosis.
------------------------------
Date: Mon, 4 May 87 11:50:36 EDT
From: tim@linc.cis.upenn.edu (Tim Finin)
Subject: Seminar - A Shell for Intelligent Help Systems (UPenn)
COLLOQUIUM
Computer and Information Science
University of Pennsylvania
A Shell for Intelligent Help Systems
Joost Breuker, Radboud Winkels, Jacobijn Sandberg
University of Amsterdam
Department of Social Science Informatics
The research reported here is part of a project
aimed at the construction of an environment for
building intelligent help systems. A help system
supports the user in handling and mastering an
information processing system. Core of this
environment is a shell that contains all domain
independent procedures and knowledge. A
comprehensive help system not only answers
questions of users, but also 'looks over their
shoulders' and interrupts when appropriate. This
means that a help system is equiped with a
PERFORMANCE INTERPRETER, consisting of a PLAN
RECOGNISER, a DIAGNOSER, and a QUESTION
INTERPRETER. Part of this shell and focus of this
paper is a generic COACH. In a help system a COACH
has two functions: to assist the user with a
current problem and to teach the user about the
IPS. The proposed COACH consists of three layers:
1) A DIDACTIC GOAL GENERATOR which genrates an
overlay of domain concepts that may be taught, 2)
STRATEGY PLANNER which constructs coaching
strategies, and 3) TACTICS which are the terminal
elements of strategies. They are the speech acts
finally "uttered" by the COACH. In this paper
these three layers are discussed in greater detail
and are related to empirical research.
Tuesday, May 12, 1987
Room 216
3:00 to 4:30
Refreshements Available
The Faculty Lounge
2:30 to 3:00
------------------------------
Date: Tue, 5 May 87 10:49:58 EDT
From: tim@linc.cis.upenn.edu (Tim Finin)
Subject: Seminar - A Computational Model of Creative Writing (UPenn)
CIS Colloquium
Computer and Information Science
University of Pennsylvania
A COMPUTATIONAL MODEL OF CREATIVE WRITING
Masoud Yazdani
Dept. of Computer Science
University of Exeter, UK
The overal aim of the project is to examine a computational model
model of creativity based on the process of meta-level inspection and
control of loosely controlled simulations. The test bed for this
study is the act of creative writing. Various proposals for
computational story writing are considered and one of them, TALE-SPIN,
is critically evaluated. A more comprehensive model for storywriting
is then presented to account for the shortcomings pointed out. The
model presented consists of five distinct processes of plot-making,
world-making, simulation, narration and text generation. These
processes are further expanded within a computational framework. A
computer program, ROALD, is described which attempts to produce
stories within this general framework. ROALD, although basically the
simulation part of the model, acts as a test bed for the more general
idea of controlled simulation. we also look at other areas (picture
making and machine learning) where related work is being carried out.
Our argument can be stated at three levels of generality:
1. That the core of the act of creative writing is simulation of life
2. That this simulation needs to be part of a model which provides
situations within which the simulations occur as well as providing
sources of constraints so that the results are consistant and
interesting.
3. That not only creative writing but other creative acts can be
be viewed as the process of a loosely controlled simulation with
metal-level validation and revision of the results.
Wednesday, May 13, 1987
Room 216
3:00 to 4:30
Refreshments Available
2:30 to 3:00
The Faculty Lounge
------------------------------
Date: 5 May 1987 1002-EDT
From: Elaine Atkinson <EDA@C.CS.CMU.EDU>
Subject: Seminar - Speaking to a Computer (CMU)
SPEAKER: Alexander Hauptmann
TITLE: "Speaking to a Computer"
DATE: Tuesday, May 5
TIME: 12:00 - 1:20 p.m.
PLACE: Adamson Wing, Baker Hall
ABSTRACT: This talk describes an empirical study of man-computer speech
interaction. I will describe the experiment, its goals and outline the
experimental design and the many results. The experiment shows that
speech to a computer is not as ill-formed as one would expect. People
speaking to a computer are more disciplined than when speaking to
each other. There are large differences in the usage of spoken language
compared to typed language, and several phenomena which are unique to
spoken or typed input respectively. Usefulness for work in speech
understanding systems for the future is considered.
------------------------------
Date: Thu 7 May 87 19:08:52-PDT
From: Ted Kamins
Reply-to: KAMINS@Sierra.Stanford.EDU
Subject: Seminar - BB* Layered Environment for AI Systems (HP)
HEWLETT-PACKARD LABORATORIES
COMPUTER COLLOQUIUM
Speaker: Barbara Hayes-Roth
Senior Research Associate
Stanford Knowledge Systems Lab
Subject: BB*: A modular and layered environment for AI systems
Time: Thursday, May 14, 1987, 4 pm
Place: Hewlett-Packard
5M Auditorium
1501 Page Mill Road
Palo Alto
Non-HP Employees: Welcome! Please come to the lobby shortly before 4 pm
so that you can be escorted to the auditorium.
Refreshments will be served following the talk.
Host: Barry Bronson (857-3033)
Stanford contact: Ted Kamins (kamins@sierra)
Abstract:
An intelligent system reasons about--controls, explains,
learns about--its actions, thereby improving its efforts to achieve
goals and function in its environment. In order to perform
effectively, a system must have knowledge of the actions it can
perform, the events and states that can occur, and the relationships
among instances of those actions, events, and states. The BB*
environment represents this knowledge in an abstraction hierarchy and
defines uniform standards of knowledge content and representation for
modules within each of three hierarchical levels: architecture,
framework, and application.
The speaker will illustrate BB* with some of its current modules: (a)
the BB1 blackboard control architecture; (b) the ACCORD framework for
arrangement-assembly tasks; and (c) several domain-specific
applications of BB1-ACCORD. BB* advantages for system representation
and performance, system design and implementation, reusable knowledge
modules, and open systems integration will be discussed.
------------------------------
End of AIList Digest
********************
