September 28th
3:00pm
ACES 6.304
|
DISCOURSE MODE:
a linguistically interesting level of text structure
PROF. CARLOTA SMITH [web][email]
CENTENIAL PROFESSOR OF LINGUISTICS
UT DEPARTMENT OF LINGUISTICS
I'll discuss a level of discourse structure which people intuitively
recognize, and which has linguistic correlates that correspond to intuition.
Recent work on discourse has emphasized the importance of genre and
other pragmatic factors in understanding and structuring discourse.
But genre is not the right category for linguistic analysis, because
it is pragmatically based. I propose a local level at which particular
linguistic forms make a difference. Within a text one recognizes passages
of different types, e.g. Narrative, Description, Report, Informative,
and Argument- Commentary. These passages represent different discourse
modes: they have a particular force and a characteristic cluster of
linguistic features and interpretations.
The linguistic properties that characterize the modes are (a) the type
of entities that they introduce into the universe of discourse and (b)
their principle of advancement, temporal or atemporal. The notion of
mode explains variety in texts. Texts of a given genre are not monolithic,
but rather have passages of different modes.
|
October 26th
3:00pm
TAY 3.128
|
PANDORA REVISITED:
a plea for circumspection in the quest for artificial intelligence
DANIEL S. HAWORTH, M.A.
DEPARTMENT OF HISTORY
The quest for artificial intelligence (AI) is not unlike the tale
of Pandora. Curiosity untempered by a clear objective led her to unwittingly
unleash a plague of tribulation upon mankind. Pushing the technological
boundaries of computational science to make the machine more human-like
in the absence of a well defined sense of purpose implies a host of
equally unsettling consequences. Using the story of Pandora as a metaphor
for what seems to be a case of highly trained minds accepting the means
in the headlong rush to develop AI as the justification for a disturbingly
ambiguous end, this presentation explores the unsettling implications
of intelligence as artifice. It concludes with a call for the members
of the computer science community to reach consensus on just what they
hope to achieve before they proceed any further.
|
November 16th
3:00pm
PAI 3.14
|
SYMBOLIC MODEL CHECKING FOR CAUSAL DOMAINS
ANUJ GOEL [email]
UT DEPARTMENT OF ELECTRICAL & COMPUTER ENGINEERING
Effective plan specification, solution extraction and plan execution
are critical to the ability of single- and multi-agent systems to operate
given the dynamism and uncertainties that exist in real-world domains.
While much of the traditional planning research has focused on increasing
the tractable search space size for fast and efficient planning, the
real-world application of domain-independent planners has suffered due
to a lack of corresponding advances with respect to new domain representations,
metric planning and replanning in dynamic and uncertain environments.
This research seeks to extend and enhance planning capabilities by applying
symbolic model checking techniques to plan for domains specified using
causal language representations.
|
November 30th
3:00pm
PAI 3.14
|
AI AND RELIGION
PROF. RAYMOND MOONEY [web]
[email]
UT DEPARTMENT OF COMPUTER SCIENCES
PROF. BENJAMIN KUIPERS [web]
[email]
UT DEPARTMENT OF COMPUTER SCIENCES
Two scientists who agree on many issues in AI and science generally,
disagree fundamentally on the topic of religion.
Is religious faith compatible with good science?
|
