iotLab

The IoT Health Program

[PC: Alberto Frigo - month of activities where each line is a day]

The Internet of Things (IoT) is a network of physical devices (e.g., smartphones, smart watches etc.), having the capacity to sense, collect and communicate data seamlessly. The IoT-Health research program aims to capitalize on advancements in technology along with sophisticated information and data processing to better understand disease progression in chronic health conditions such as mental illness and cancer. In this research program we couple IoT processes with Machine Learning (Deep Learning Neural Networks), to gather passive data about human health and then use this information to develop predictive algorithms for chronic diseases.

Project 1: Mood Inference Engine.

This project aims to infer mood from "phone conversation" without any human intervention. Deep Learning Neural Networks would be the core of this development, with a major focus on "privacy" and "robust" computing. This project has great commercial value and is central to many projects being developed in this lab. A completely novel method is being developed, which will be first of its kind.

Collaborators: Dr Nic Lane (Bell Labs), Dr Julien Epps (UNSW), Professor Margot Brereton (QUT), Dr Raja Jurdak (CSIRO), Professor Roland Goecke (UC), Professor Michael Breakspear (QIMR)

Industry Partners: Queensland Health, Cogninet, Nexus e-care.

PhD Student sought: Yes.

Project 2: Relapse Prediction in Mental Illness.

This project aims to develop a relapse prediction system on a smartphone for mental health patients. It will "passively" infer information about four domains (1) mood, (2) sleep, (3) physical activity and (4) social connectedness and will profile a person based on these. A suit of algorithms will be developed which will continuously run in the background and detect any "persistent change" in those domains and will attempt to predict a relapse based on those changes.

Collaborators: Dr Nic Lane (Bell Labs), Dr Julien Epps (UNSW), Professor Margot Brereton (QUT), Dr Raja Jurdak (CSIRO), Professor Roland Goecke (UC), Professor Michael Breakspear (QIMR)

Industry Partners: Queensland Health, Nexus e-care.

PhD Student sought: Yes.

Project 3: Doctor's Resilience.

This project is led by Dr Michael Ireland at the Innovative Mental Health Service group at USQ. Doctors especially emergency doctors have long hours of stressful job, which often causes burnout leading to an error in treatment, which can cost even lives. This project aims at predict burnouts so that interventions can be offered to doctors to prevent the burnout to occur. The technical side of this project is of interest to me - i.e., burnout prediction. Various physiological sensor data from wearable sensors will be used to model the behavior using Deep Learning models and any persistent deviation will be identified and mapped to burnout for developing a prediction model.

This project has serious merit as this will not be limited to doctors but will span to other emergency professionals - firefighters, police etc. In particular, in Australia even in general employment sectors, there is a push for maintaining employee health and wellbeing. So this technology will be widespread in near future.

PhD Student sought: Yes.

Project 4: Adherence prediction for chronic disease patients.

This project is led by Dr Sonja March, the director of the Innovative Mental Health Service Program at USQ. Chronic disease patients often fail to adhere to medication, which causes further complications and increased disease burdens. This project aims to predict the non-adherence behavior so that interventions can be brought into place early enough. My focus is on using wearable sensor data both physiological and behavioral and using advanced deep learning algorithms to predict the non-adherence behavior.

PhD Studnet sought: Yes.

Project 5: Mood incorporated recommendation systems.

Recommendation Systems are quite popular nowadays. Highly successful companies like Netflix earn their 75% of revenue from recommendations. A mood-aware recommendations can be more appropriate and could achieve higher accuracy. This project aims to develop a mood-aware recommendation system and validate its impact on a recommendation.

PhD Student sought: Yes.

Project 6: Marker bias from mood inference.

Recent studies validate that Mood has a profound impact on marking. Poor marking can cause dropouts at the college and universities and is a threat to the Australian economy. The project aims to quantify the impact of mood on marking, which will eventually contribute towards developing a marking bias cancellation system.

PhD Student sought: Yes.

Project 7: Distress Detection in Cancer Patient

The aim of the research is to develop an automated system to detect “distress” in cancer patients enabling early referral to interventions, to target anxiety and depression, to mitigate suicidal ideation and to improve adherence to treatment, which would significantly improve quality of life for individuals and reduce the costs on the health care system.

Collaborators: Professor Jeff Dunn, Professor Suzanne K Chambers, Professor Joanne Aitken.

Industry Partners: Cancer Council Queensland.

PhD Student sought: Yes.

People