Cross-lingual transfer learning for intent classification
(This project will have an external supervisor from Kindly, in addition to an internal supervisor from LTG.)
Conversational agents (aka chatbots) are computer programs that engage in conversations with humans using natural language, be it text or speech. Text-based chatbots in specific have become a ubiquitous technology in customer support, enabling companies to provide round-the-clock automated assistance to their customers. Some companies, however, operate in different locations and serve different customer bases, and hence developing cross-lingual chatbots is of high relevance from research and practical perspectives. Indeed, the interest in cross-lingual learning in NLP has surged over the past few years. This project will explore one of the tasks that facilitate building such cross-lingual chatbots.
Intent classification and response selection are two of the core tasks in almost all conversational agents, in addition to many other NLP tasks such as speech recognition, language detection, named entity recognition etc. This master’s project will focus on the task of cross-lingual intent classification which, simply, amounts to recognizing the user’s intent given some input (i.e. the user’s message). Some of the concrete research questions this project may focus on are:
- Given an intent classification model for language L1 (for example, English), can we use this model to bootstrap another intent classification model for language L2 (for example, Norwegian)? This question will rely on cross-lingual transfer learning methods
- Can we build one cross-lingual intent classification model that recognizes the user’s intent in two or more languages? How does such model perform compared to monolingual models trained separately for the corresponding languages?
Kindly is a language technology company based in Oslo. Kindly was founded in 2016 and has since grown to a team of 40+ employees, several of whom have graduated from the Language Technology Group (LTG) at UiO with master’s and doctoral degrees. Natural language processing and machine learning are at the core of the Kindly chatbot platform, which powers the conversational agents for some of the leading enterprises in the Nordics, such as Norwegian Air Shuttle, Elkjøp, Kahoot!, Thon Hotels, and Finn. This master’s project will be carried out in close cooperation with Kindly.