AsianScientist (Jan. 2, 2018) – A group of researchers at Osaka University, Japan, has developed a dialogue system in which their machine learns the meanings of words through conversations. They published their findings at the 18th Annual SIGdial Meeting on Discourse and Dialogue.
Many conversation robots, chatbots and voice assistant apps have appeared in recent years. However, in these systems, computers basically answer questions based on what has been preprogrammed. There is another method in which a computer learns from humans by asking simple repetitive questions. Yet, if the computer asks only questions such as “What is xyz?” to acquire knowledge, users will lose interest in conversing with the computer.
In this study, a team of scientists led by Professor Kazunori Komatani developed an implicit confirmation method by which a computer acquires the meaning of an unknown word through conversations with humans. This method allowed their system to predict the category of an unknown word during dialogues.
To decide whether the prediction of word meanings was correct, the system relied on machine learning techniques, even refining its decisions about word meanings by taking into account the classification results gained from dialogues with multiple users. The researchers suggest that their work is a step towards the realization of dialogue systems with the ability to customize responses to different users.
The article can be found at: Ono et al. (2017) Lexical Acquisition through Implicit Confirmations over Multiple Dialogues.
———
Source: Osaka University; Photo: Shutterstock.
Disclaimer: This article does not necessarily reflect the views of AsianScientist or its staff.
