tailieunhanh - Báo cáo khoa học: "Fast Online Training with Frequency-Adaptive Learning Rates for Chinese Word Segmentation and New Word Detection"

We present a joint model for Chinese word segmentation and new word detection. We present high dimensional new features, including word-based features and enriched edge (label-transition) features, for the joint modeling. As we know, training a word segmentation system on large-scale datasets is already costly. | Fast Online Training with Frequency-Adaptive Learning Rates for Chinese Word Segmentation and New Word Detection Xu Sun Houfeng Wang Wenjie Li Department of Computing The Hong Kong Polytechnic University Key Laboratory of Computational Linguistics Peking University Ministry of Education China csxsun cswjli @ wanghf@ Abstract We present a joint model for Chinese word segmentation and new word detection. We present high dimensional new features including word-based features and enriched edge label-transition features for the joint modeling. As we know training a word segmentation system on large-scale datasets is already costly. In our case adding high dimensional new features will further slow down the training speed. To solve this problem we propose a new training method adaptive online gradient descent based on feature frequency information for very fast online training of the parameters even given large-scale datasets with high dimensional features. Compared with existing training methods our training method is an order magnitude faster in terms of training time and can achieve equal or even higher accuracies. The proposed fast training method is a general purpose optimization method and it is not limited in the specific task discussed in this paper. 1 Introduction Since Chinese sentences are written as continuous sequences of characters segmenting a character sequence into words is normally the first step in the pipeline of Chinese text processing. The major problem of Chinese word segmentation is the ambiguity. Chinese character sequences are normally ambiguous and new words out-of-vocabulary words are a major source of the ambiguity. A typical category of new words is named entities including organization names person names location names and so on. 253 In this paper we present high dimensional new features including word-based features and enriched edge label-transition features for the joint modeling of Chinese word segmentation .