TAILIEUCHUNG - Báo cáo khoa học: "Simple semi-supervised training of part-of-speech taggers"

Most attempts to train part-of-speech taggers on a mixture of labeled and unlabeled data have failed. In this work stacked learning is used to reduce tagging to a classification task. This simplifies semisupervised training considerably. Our prefered semi-supervised method combines tri-training (Li and Zhou, 2005) and disagreement-based co-training. On the Wall Street Journal, we obtain an error reduction of with SVMTool (Gimenez and Marquez, 2004). | Simple semi-supervised training of part-of-speech taggers Anders S0gaard Center for Language Technology University of Copenhagen soegaard@ Abstract Most attempts to train part-of-speech taggers on a mixture of labeled and unlabeled data have failed. In this work stacked learning is used to reduce tagging to a classification task. This simplifies semisupervised training considerably. Our prefered semi-supervised method combines tri-training Li and Zhou 2005 and disagreement-based co-training. On the Wall Street Journal we obtain an error reduction of with SVMTool Gimenez and Marquez 2004 . 1 Introduction Semi-supervised part-of-speech POS tagging is relatively rare and the main reason seems to be that results have mostly been negative. Meri-aldo 1994 in a now famous negative result attempted to improve HMM POS tagging by expectation maximization with unlabeled data. Clark et al. 2003 reported positive results with little labeled training data but negative results when the amount of labeled training data increased the same seems to be the case in Wang et al. 2007 who use co-training of two diverse POS taggers. Huang et al. 2009 present positive results for self-training a simple bigram POS tagger but results are considerably below state-of-the-art. Recently researchers have explored alternative methods. Suzuki and Isozaki 2008 introduce a semi-supervised extension of conditional random fields that combines supervised and unsupervised probability models by so-called MDF parameter estimation which reduces error on Wall Street Journal WSJ standard splits by about 7 relative to their supervised baseline. Spoustova et al. 2009 use a new pool of unlabeled data tagged by an ensemble of state-of-the-art taggers in every training step of an averaged perceptron POS tagger with 4-5 error reduction. Finally S0gaard 2009 stacks a POS tagger on an unsupervised clustering algorithm trained on large amounts of unlabeled data with mixed results. This work combines a new

TỪ KHÓA LIÊN QUAN
TAILIEUCHUNG - Chia sẻ tài liệu không giới hạn
Địa chỉ : 444 Hoang Hoa Tham, Hanoi, Viet Nam
Website : tailieuchung.com
Email : tailieuchung20@gmail.com
Tailieuchung.com là thư viện tài liệu trực tuyến, nơi chia sẽ trao đổi hàng triệu tài liệu như luận văn đồ án, sách, giáo trình, đề thi.
Chúng tôi không chịu trách nhiệm liên quan đến các vấn đề bản quyền nội dung tài liệu được thành viên tự nguyện đăng tải lên, nếu phát hiện thấy tài liệu xấu hoặc tài liệu có bản quyền xin hãy email cho chúng tôi.
Đã phát hiện trình chặn quảng cáo AdBlock
Trang web này phụ thuộc vào doanh thu từ số lần hiển thị quảng cáo để tồn tại. Vui lòng tắt trình chặn quảng cáo của bạn hoặc tạm dừng tính năng chặn quảng cáo cho trang web này.