Publication View

Transformation-based learning in the fast lane (2001)

Abstract

Transformation-based learning has been successfully employed to solve many natural language processing problems. It achieves state-of-the-art performance on many natural language processing tasks and does not overtrain easily. However, it does have a serious drawback: the training time is often intorelably long, especially on the large corpora which are often used in NLP. In this paper, we present a novel and realistic method for speeding up the training time of a transformation-based learner without sacricing performance. The paper compares and contrasts the training time needed and performance achieved by our modied learner with two other systems: a standard transformation-based learner, and the ICA system (Hepple, 2000). The results of these experiments show that our system is able to achieve a signicant improvement in training time while still achieving the same performance as a standard transformation-based learner. This is a valuable contribution to systems and algorithms which utilize transformation-based learning at any part of the execution.

Publication details

Download	http://citeseerx.ist.psu.edu/viewdoc/summary?doi=?doi=10.1.1.23.26
Source	http://nlp.cs.jhu.edu/~rflorian/papers/naacl01.ps
Contributors	CiteSeerX
Repository	CiteSeerX - Scientific Literature Digital Library and Search Engine (United States)
Type	text
Language	English
Relation	10.1.1.2.6514, 10.1.1.74.5174, 10.1.1.18.6783, 10.1.1.15.9965, 10.1.1.10.572, 10.1.1.10.7114, 10.1.1.103.8498, 10.1.1.18.5155, 10.1.1.10.4172, 10.1.1.115.2724, 10.1.1.65.1346, 10.1.1.103.4895, 10.1.1.58.2675, 10.1.1.58.4802, 10.1.1.114.7026, 10.1.1.2.6251, 10.1.1.78.4885, 10.1.1.59.8187, 10.1.1.62.8875, 10.1.1.63.1050, 10.1.1.75.6336, 10.1.1.10.6952, 10.1.1.106.5052, 10.1.1.63.8300, 10.1.1.65.9841, 10.1.1.78.4905, 10.1.1.98.3299, 10.1.1.136.3402, 10.1.1.10.2616, 10.1.1.10.5138, 10.1.1.10.7926, 10.1.1.104.924, 10.1.1.105.2426, 10.1.1.105.9724, 10.1.1.12.6834, 10.1.1.18.2225, 10.1.1.18.9364, 10.1.1.2.5733, 10.1.1.21.3594, 10.1.1.57.8821, 10.1.1.59.228, 10.1.1.59.6174, 10.1.1.60.9437, 10.1.1.61.3238, 10.1.1.61.3417, 10.1.1.62.6468, 10.1.1.7.5087, 10.1.1.7.660, 10.1.1.70.1434, 10.1.1.70.2661, 10.1.1.72.6500, 10.1.1.74.7300, 10.1.1.77.6895, 10.1.1.79.7653, 10.1.1.88.615, 10.1.1.9.6165, 10.1.1.111.7867, 10.1.1.90.8534, 10.1.1.92.730, 10.1.1.97.7679, 10.1.1.115.6898, 10.1.1.116.9726, 10.1.1.117.2068, 10.1.1.118.5116, 10.1.1.120.4561, 10.1.1.124.6433, 10.1.1.129.9165, 10.1.1.136.8591