Machine learning helps computers predict near-synonyms
19 June 2015
Last update: 19/06/15 10:34
Choosing the best word or phrase for a given context from among candidate near-synonyms, such as “slim” and “skinny”, is something that human writers, given some experience, do naturally; but for choices with this level of granularity, it can be a difficult selection problem for computers.
Researchers from Macquarie University in Australia have published an article in the journal Natural Language Engineering, investigating whether they could use machine learning to re-predict a particular choice among near-synonyms made by a human author – a task known as the lexical gap problem.
They used a supervised machine learning approach to this problem in which the weights of different features of a document are learned computationally. Through using this approach, the computers were able to predict synonyms with greater accuracy and reduce errors.
The initial approach solidly outperformed some standard baselines, and predictions of synonyms made using a small window around the word outperformed those made using a wider context (such as the whole document).
However, they found that this was not the case uniformly across all types of near-synonyms. Those that embodied connotational or affective differences — such as “slim” versus “skinny”, with differences in how positively the meaning is presented — behaved quite differently, in a way that suggested that broader features related to the ‘tone’ of the document could be useful, including document sentiment, document author, and a distance metric for weighting the wider lexical context of the gap itself (For instance, if the chosen near-synonym was negative in sentiment, this might be linked to other expressions of negative sentiment in the document).
The distance weighting was particularly effective, resulting in a 38% decrease in errors, and these models turned out to improve accuracy not just on affective word choice, but on non-affective word choice also.
Read the full article ‘Predicting word choice in affective text’ online in the journal Natural Language Engineering
Related Posts
Australia’s Response to the Coronavirus Pandemic – Now updated
From time to time, until the crisis has passed, the HEPL blog series authors will be given the opportunity to provide short updates on their country/region’s continuing response to this worldwide catastrophe and their further reflections on those responses. Each update will be labelled accordingly with the original response at the bottom of each post. […]
Conversations with Authors: “Se Habla Español: Spanish-Language Appeals and Candidate Evaluations in the United States”
In this “Conversation with Authors,” we spoke with APSR authors Marques G. Zárate, Enrique Quezada-Llanes and Angel D. Armenta about their open access article “Se Habla Español: Spanish-Language Appeals and Candidate Evaluations in the United States.” APSR: The first question is two-fold. Where did you get the inspiration for this paper? How did this co-authorship […]
Influential women in computer science
In support of International Women’s Day, we celebrate the contributions of 15 of the most influential women throughout the history of computer science. Their biographies and accomplishments in the field are celebrated here. If there is someone you would like to see us highlight next year, leave your message in the comment section. Read also: […]
International Women’s Day 2020: Influential women in STEM
International Women’s Day 2020 falls on Sunday, 8th March this year. In the run up to this date, each week day we’ll be highlighting one woman whose accomplishments in science, technology, engineering and/or mathematics not only elevated their fields but also took us one step closer to a gender-equal world. We hope you’ll join us […]
;
