Tutorial Report for ACL-2004
Inderjeet Mani

In response to the Call for Tutorial Proposals, I received 8 tutorial
proposals, broken down as follows: 3 on speech, 1 on machine lea
rning, and 4 on particular NLP problems.

These proposals were reviewed by me with the expert assistance of two
other colleagues: Marc Vilain of MITRE and Grace Chung of CNRI.  The
proposals were judged based on a subjective assessment of quality,
representativeness, expertise, attractiveness, and feasibility.

While the proposals were all well-thought-out and of high quality,
only 3 could be selected that met all of the above criteria. In ad
dition, I solicited one additional tutorial on a machine
learning-related topic (T3) which had been given at ACL-2002 (with
the foc us broadened somewhat, based on discussion with the

Past tutorial statistics from the ACL Business Manager suggested that
four half-day tutorials all held on a single day seemed to work
best. The final list of four half-day tutorials, with one on
annotation, one on speech, and two on machine learning-related topics,
is as below:

T1. Beyond Syntax: Predicates, Arguments, Valency Frames and
Linguistic Annotation Collin Baker, International Computer Science

Jan Hajic, Charles University
Martha Palmer, University of Pennsylvania
Manfred Pinkal, Saarlandes University

T2. Building linguistically motivated speech recognisers with Regulus

Manny Rayner and Beth Ann Hockey, NASA Ames Research Center
Pierrette Bouillon, ISSCO/University of Geneva

T3. Kernel Methods in Natural Language Processing

Jean-Michel Renders, Xerox Research Center Europe

T4. Adaptive Learning: From Supervised to Active Learning of
statistical models for Natural Language and Speech Processing

Giuseppe Riccardi, Dilek Hakkani-Tur and Gokhan Tur, AT&T Labs-Research

The preparation of the web material, tutorial notes and cd-rom
materials all went reasonably well, with a glitch or two that was
promptly addressed by ACL2004 staff.

As of June 28, 2004, the following pre-registration statistics were
available: T1: 44; T2: 10; T3: 68; T4: 59. Regarding T2, the
presenters informed me that at least 4 others had confirmed they would
attend. These numbers are higher than the results from ACL2002 *final*
registrations, where we had 13, 43, 46, and 54.

I look forward to a stimulating set of tutorials on July 21, 2004.