Application of topic models for conceptualization and sentiment analysis
Dongwoo Kim and Alice Oh (Alice Oh is an Assistant Professor of Korea Advanced Institute of Science and Technology. )
NICTA SML SEMINAR Machine Learning Research Group SeminarDATE: 2013-04-11
TIME: 11:15:00 - 12:00:00
LOCATION: NICTA - 7 London Circuit
CONTACT: JavaScript must be enabled to display this email address.
ABSTRACT:
In this talk, we will describe two papers in which probabilistic topic models were used to assist in understanding text.
The first paper looks at the problem of conceptualization which is mapping words and phrases to concepts. Prior research in conceptualization uses human-crafted knowledge bases that map instances to concepts, but such approaches have the limitation that the mappings are not context sensitive. To overcome this limitation, we harness the power of a probabilistic topic model which inherently captures the semantic relations between words. By combining latent Dirichlet allocation with a large-scale probabilistic knowledge base, we develop a corpus-based framework for context-dependent conceptualization. Through this simple but powerful framework, we improve conceptualization and enable a wide range of applications that rely on semantic understanding of short texts, including frame element prediction, word similarity in context, ad-query similarity, and query similarity.
In the second part of the talk, we will describe a hierarchical topic model that discovers a tree structure of aspects and sentiments from online product reviews. In this tree, each node is again a two-level tree, whose root represents an aspect and the children represent the sentiment polarities associated with it. Each aspect or sentiment polarity is modeled as a distribution of words. To automatically extract both the structure and parameters of the tree, we use recursive Chinese Restaurant Process (rCRP) as the prior and jointly infer the aspect-sentiment tree from the reviews. We experiment with two datasets and show that our model successfully discovers the aspect-sentiment tree and achieves better sentence-level classification accuracy than previously proposed aspect-sentiment joint models.
BIO:
Alice Oh is an Assistant Professor of Computer Science at Korea Advanced Institute of Science and Technology and leads the Users and Information Lab. Dongwoo Kim is a PhD student.
