Deakin - Coventry Cotutelle - Speech-based Conversational Question-Answering System

A doctoral cotutelle project in 'Speech-based Conversational Question-Answering System for Health Care Information Retrieval between Deakin University (Australia) and Coventry University (United Kingdom).'

Project Supervisor

Additional Supervision Team

Location

Geelong Waurn Ponds Campus (Australia) and Coventry University (United Kingdom)

Research topic

This is a doctoral cotutelle project  between Deakin University (Australia) and Coventry University (United Kingdom).

The successful PhD Student will be awarded a scholarship from Deakin University with the supervision team being drawn from Deakin University and Coventry University. The PhD Student will graduate with two testamurs, one from Deakin University and one from Coventry University, each of which recognises that the program was carried out as part of a jointly supervised doctoral program. The program is for a duration of 4 years.

The PhD Student is anticipated to spend up to 12 months of the total period of the program at Coventry University, with the remainder of the program based at Deakin University.

Searching digital information is traditionally thought as a task that consists of using keywords for finding relevant documents or resources on a search engine to answer an information need. Nonetheless, search engine users are now keener to express their information need using questions in a natural language (e.g., “When did Beyonce start becoming popular?”) rather than keywords, and they tend to expect obtaining concise answers (e.g., “in the late 90s”), rather than an entire set of documents to answer their questions. Search engines that enable this feature that consists of asking and answering questions in a natural language are known as Question-Answering (QS) systems. QA systems have a long and rich history of research and development dating back to the 60s. While early systems focused on narrow and closed domains where they were only able to answer questions without considering the underlying context, the focus today is more on the conversational aspect that consists of incorporating the history of the conversation in the whole process. This gave the rise to speech-based conversational search systems, which are iterative and interactive search systems for which the ultimate goal is to address user information needs through multi-turn speech-based natural language conversations. Furthermore, this evolution is supported and encouraged by the emergence and the proliferation of intelligent assistants such as Siri, Alexa, and Google Assistant, where information-seeking activities are carried out over voice or text-based conversational interfaces. However, we note that conversational query answering systems are not yet well explored for retrieving medical / health care information.

The goal of this project is to explore, research, develop, and implement such medical conversational QA systems. Specifically, the aims and objectives of this research are:

  • Research and develop efficient algorithms to (i) retrieve documents from the collection that presumably contain the right answers, and (ii) to extract the answers from these documents in the form of passages, which are then formatted and presented to the user.
  • Research on designing efficient speech-based user interfaces given the constraints associated with conversational QA systems. For instance, the input interface can’t effectively use techniques like speech-based autocomplete to guide the user. On the other hand, the output interface is even more constrained as the output is a single answer (sentence) to the searcher that does not leave room for error.
  • Implement a conversational medical QA system that considers all the associated constraints.

Project aim

The project aims will be met by relying on the use of the current cutting-edge technologies in Natural Language Processing, e.g., Transformer-based models or pretrained models such as BERT, XLNet, and GPT-3. Also, in order to ensure that the project’s aims will be met on time, a specification and work plan will be drawn up, detailing how the research will be conducted.

This project will bring together domestic and international researchers from Deakin University, Coventry University, and the University of New South Wales with multidisciplinary skills of computer science, health informatics, and artificial intelligence.

Important dates

Applications close 5pm, Thursday 15 December 2022

Benefits

This scholarship is supported by Deakin University, is available over 3 years and includes:

  • Stipend of $28,900 per annum tax exempt (2021 rate)
  • Full Tuition Fee Waiver for up to 4 years
  • Funding to support travel of PhD Student between Deakin University and Coventry University
  • Relocation allowance of $500 - $1500 (for single to family) for students moving from interstate
  • International students only:  Single Overseas Student Health Cover policy for the duration of the student visa.

Eligibility criteria

To be eligible you must:

  • be either a domestic or international candidate currently residing in Australia. Domestic includes candidates with Australian Citizenship, Australian Permanent Residency or New Zealand Citizenship.
  • *meet the PhD entry requirements of both Deakin University and Coventry university, including English language proficiency requirements (IELTS overall minimum score of 7.0 with a minimum of 6.5 in each component)
  • a masters degree in a relevant subject area, with overall mark at minimum 80%
  • be able to physically locate to both Coventry University (UK) and Deakin University (Australia)

Please refer to the research degree entry pathways page and Coventry’s research entry criteria page for further information.

How to apply

Applicants should firstly contact Dr Mohamed Reda Bouadjenek to discuss the project. After discussing your application with the Deakin Supervisor, you will be required to submit an Expression of interest directly with the relevant Faculty. The successful candidate will be invited to lodge a formal HDR application to Deakin.

The successful candidate will also be required to lodge a separate PhD application to Coventry University via the Coventry ‘Making an Application’ page.

Please be aware that screening for this advert will commence immediately and the scholarship may be awarded prior to the closing date.

Contact us

For more information about this scholarship, please contact:

Dr Mohamed Reda Bouadjene
Email reda.bouadjenek@deakin.edu.au
+61 3 522 78380