Abstract: The typical paradigm of the machine learning task is to obtain a large amount of labeled data and then train a classifier for further predictions/classifications. However, this paradigm may suffer from the high cost and low efficiency in manual annotation. In this talk, we use Name Entity Recognition (NER) task as an example to show how distant supervision and weak supervision can be applied to reduce human annotation. We introduce the pattern-enhanced NER methods, which automatically mines the entity naming principles to enhance the weak supervision. We demonstrate the power of these methods in intensive experiments on real-world datasets in biomedical domain.
Bio: Qi Li is an Assistant Professor in Computer Science Department at Iowa State University. She was a Postdoc in the Department of Computer Science, University of Illinois at Urbana-Champaign under the supervision of Prof Jiawei Han upon her Ph.D. graduation from SUNY Buffalo Department of Computer Science and Engineering under the direction of Prof Jing Gao. Her research interests lie in the area of data mining with a focus on the extraction and aggregation of information from multiple data sources. She is nominated for Microsoft Faculty Fellowship by ISU.