Name
Abstract | ||
|---|---|---|
Typical information extraction (IE) systems can be seen as tasks assigning labels to words in a natural language sequence. The performance is restricted by the availability of labeled words. To tackle this issue, we propose a semi-supervised approach to improve the sequence labeling procedure in IE through a class of algorithms with {\em self-learned features} (SLF). A supervised classifier can be trained with annotated text sequences and used to classify each word in a large set of unannotated sentences. By averaging predicted labels over all cases in the unlabeled corpus, SLF training builds class label distribution patterns for each word (or word attribute) in the dictionary and re-trains the current model iteratively adding these distributions as extra word {\em features}. Basic SLF models how likely a word could be assigned to target class types. Several extensions are proposed, such as learning words' class boundary distributions. SLF exhibits robust and scalable behaviour and is easy to tune. We applied this approach on four classical IE tasks: named entity recognition (German and English), part-of-speech tagging (English) and one gene name recognition corpus. Experimental results show effective improvements over the supervised baselines on all tasks. In addition, when compared with the closely related self-training idea, this approach shows favorable advantages. |
Year | DOI | Venue |
|---|---|---|
2009 | 10.1109/ICDM.2009.40 | ICDM |
Keywords | Field | DocType |
extra word,basic slf model,word attribute,class type,class label distribution pattern,class boundary distribution,semi-supervised sequence,annotated text sequence,classical ie task,slf training,semi-supervised approach,self-learned features,hidden markov models,labeling,natural language,artificial neural networks,information extraction,feature extraction,semi supervised learning,data mining,natural language processing,learning artificial intelligence,sequence labeling | Data mining,Semi-supervised learning,Sequence labeling,Computer science,Artificial intelligence,Natural language processing,Classifier (linguistics),Pattern recognition,Feature extraction,Information extraction,Natural language,Hidden Markov model,Named-entity recognition,Machine learning | Conference |
ISSN | Citations | PageRank |
1550-4786 | 7 | 0.46 |
References | Authors | |
24 | 6 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Qi, Yanjun | 1 | 684 | 45.77 |
| Pavel Kuksa | 2 | 399 | 24.10 |
| Ronan Collobert | 3 | 4002 | 308.61 |
| Sadamasa, Kunihiko | 4 | 91 | 4.63 |
| Koray Kavukcuoglu | 5 | 10189 | 504.11 |
| Jason Weston | 6 | 13068 | 805.30 |