You are here
» Research » Student research projects » Interpretable Machine Intelligence in Early Detection of Ovarian CancerInterpretable Machine Intelligence in Early Detection of Ovarian Cancer
People
Principal investigator
External Member
Description
Ovarian Cancer (OC) is a significant health issue with a lasting impact on the whole community. Over the last 20 years, improvements in survival rates have been insufficient. The major contributing factor to the high mortality rate is diagnosis at advanced stage and the continuing lack of a clinically useful early detection strategy.
This MPhil or PhD study will be part of a multidisciplinary initiative in computer science and medicine by will utilise existing evidence to develop and improve approaches for the early diagnosis of OC. Recently, our national and international partners in medicine have successfully delivered a grant-funded discovery project. Their main highlight was discovering through retrospective case-control analyses an informative biomarker, also known as disease signature, of early-stage serous OC, which is responsible for most of the disease mortality. Together we then gave evidence of machine learning being powerful and accurate in scaling this classification capability to screening clinical data of asymptomatic individuals for this biomarker and thereby support both early intervention of OC and renewal related knowledge in computer and health sciences.
These discoveries suggest the aforementioned OC data science and analytics strategies it could be ideal as a first line test for population screening, and consequently, we are seeking for Expressions of Interests (EoI) from potential Higher Degree Research (HDR) candidates in computing. EoIs are due by 19 September 2021. An Australian Government Domestic Research Training Program Scholarship with a generous grant-funded top-up will be available for one successful candidate for a January 2022 start at the Australian National University (ANU) School of Computing.
Goals
The HDR project will construct interpretable methods, based on machine intelligence, for early detection of OC and computational abilities to understand its disease mechanism in order to support scientific discoveries in both health and computer sciences. More precisely, it will
- design, create, and validate though science of annotation an expert-labelled reference standard for machine learning that ideally compromises resource frugality and credibility,
- investigate minimalistic high-recall high-precision biomarkers across clinical tests, measurements, and other data related to OC that could be suitable for population screening, and
- devise, develop, and implement new transparent, reproducible, and reliable methods for interpreting the outcomes of data analytics and machine learning that can operate at scale.
Requirements
To be successful, the student will need to have strong programming skills, an interest in (and preferably some experience with) physiological data, signal processing, and medical applications, and a strong ability to work in teams, including good communication skills. We are planning to implement the project in Python, R, and/or Weka, and hence, programming skills in more than one of these languages are desirable. Data analysis and experimental design skills are also required. A track record of app development and/or technical skills in working with integrated devices would be an advantage. There is the potential for commercial applications arising from this project, so Intellectual Property (IP) agreements will be essential.
To express your interest in the scholarship and top up, please email the following documents to Associate Professor Hanna Suominen (hanna.suominen@anu.edu.au) by 19 September 2021:
- A current CV/resume,
- A research proposal (max 2 pages, PDF or DOCX),
- A flowchart of relevant data science and analytics methods (e.g., a PDF or PPTX slide) to address one or more of the following project goals:
- design, create, and validate though science of annotation an expert-labelled reference standard for machine learning that ideally compromises resource frugality and credibility,
- investigate minimalistic high-recall high-precision biomarkers across clinical tests, measurements, and other data related to OC that could be suitable for population screening, or
- devise, develop, and implement new transparent, reproducible, and reliable methods for interpreting the outcomes of data analytics and machine learning that can operate at scale.
- An annotated bibliography (see https://www.anu.edu.au/students/academic-skills/writing-assessment/other-assessments/annotated-bibliography, max 5 pages, PDF or DOCX) of up to 10 key papers relevant to the proposed flowchart,
- A showcase of a previous data science and analytics experiment (preferably in Python, R, and/or Weka.), including, but not limited to, screen captures of running experiments and their results (max 5 pages, PDF or DOCX),
- Colour copies of all transcripts and completion certificates of prior study in English (preferably original language copies and their official English translations, if available), and
- Contact details for 3 referees.
It is expected that suitable candidates will be able to qualify for an Australian Government Domestic Research Training Program Scholarship in the ANU School of Computing round whose applications are due by 31 October 2021.
Background Literature
Further information about the generous stipend (approx. 28,600 AUD per annum in 2021) and fee-waiver in the ANU is available at http://www.anu.edu.au/students/scholarships-fees/scholarships/anu-phd-scholarships. Our successful candidate will score a stipend top-up as well.
See https://cecs.anu.edu.au/study/phd-mphil for general information about PhD Scholarships in the ANU in computer science. Please do not hesitate to contact Associate Professor Hanna Suominen or Professor Amanda Barnard for further information.
Gain
Study in a truly global university that attracts the best and brightest students, academics staff, and researchers from around the world to Australia to transform research through interdisciplinary collaboration on the world's most intractable problems to impact in people’s health and healthcare. The ANU is a global university that consistently ranks among the world’s finest. Its distinct intellectual capacity is reflected in 95% of its research output being ranked above world standard. This research excellence contributes to the social, economic, and human capital of the nation.
The ANU is located in Canberra, Australia’s capital and is a partner to the Australian government and a resource for its people. Canberra was evaluated in the Organisation for Economic Co-operation and Development (OECD) Regional Well-Being Report 2014 as the most liveable city in the world.
One of the reasons ANU stands apart from other universities is its ability to look beyond the day to day and address some of the biggest problems facing the world. Research leaders of the ANU, such as Professor Amanda Barnard and Associate Professor Hanna Suominen of the ANU School of Computing, bring people together from all across the university and beyond, to bring new perspectives to major challenges confronting society – like OC. This is unique, and with them, you would have a possibility to contribute to and excel in a study that is of both world impact and national relevance.
This PhD project in computer science is part will be part of a multidisciplinary initiative with our partners in medicine at the University of Queensland (UQ) Centre for Clinical Research (e.g., Professor Greg Rice and Associate Professor Carlos Salomon Gallo), Medical Research Council Clinical Trials Unit of the University College London, UK, and Department of Obstetrics and Gynaecology of the National University of Singapore. Consequently, cross-institutional collaboration and pandemic-permitting research traveling opportunities will be available. They will help the student to expand their knowledge bases, abilities, and skills from computer to health sciences, contribute to their demonstrable evidence of successful teamwork with others nationally and internally across organisations, professions, and disciplines.
Keywords
Artificial Intelligence, Data Mining, Interpretability, Machine Learning, Medical Informatics, Ovarian Cancer, Signal Processing
