Helping participants of online communities thrive, support their pro-social behaviours and duty-of-care are challenging tasks. This is particularly difficult in the online peer support groups that are becoming increasingly popular on social networks like Facebook or organizations like ReachOut.com in Australia. In one of our current projects we are developing Moderator Assistant, that uses natural language processing techniques to provide automated support for multiple online support groups. The system generates templated interventions that the moderator can use, based on key-terms and concepts extracted from the text posted by participants. In a second project, cybermate, we use NLP and behaviour analytics to build interventions targeted directly to final users. Both systems implement behaviour analysis features to measure the impact of the interventions.
Engineers and psychologists can collaborate to build mental health support systems that are more efficient and have a more positive impact on those who need it. One way to do this is to combine Natural Language Processing, machine learning and software engineering to build better online mental health support systems. Large online health support groups provide an increasingly important type of support to people with mental health problems (Christensen, et al., 2009)(Webb, Burns, & Collin, 2008). Thousands of people go to public social networking websites such as Facebook and LiveJournal (http:// livejournal.com/) seeking help, but generally find very few trained people providing professional feedback. Peer-to-peer communities and self-support groups are amongst the most promising forms of e-health (Eysenbach, et al., 2004). Peer-to-peer support (Davidson, et al., 2006) is based on the assumption that people who have overcome difficulty can provide valuable support, guidance and hope to other people facing similar problems. They also help to improve the ‘helper’s’ self-esteem and reduce self-stigma (Corrigan, 2006). Peer-to-peer support can be considered a form of mental health intervention used independently or bundled with other forms of intervention but is often not moderated. Other online communities provide more structured support via organizations such as the Inspire Foundation Australia (http://inspire.org.au/). These organizations provide services through websites such as ReachOut.com where there is help amongst peers yet the community is supported by professionals. In these websites and online communities, young people can seek and receive help from trained staff, and use professionally developed resources.In both scenarios, particularly the latter, moderators must spend a significant amount of time providing written feedback. Maintaining the quality of feedback and complying with duty of care is challenging even within small communities, but when the community grows their support might become unsustainable. Recent human-computer interaction (HCI) research on mental health has explored online interventions such as internet-based cognitive behavior therapy systems (Christensen, Griffiths and Farrer, 2009), relational agent (Bickmore and Gruber, 2010), virtual reality (Coylea, et al., 2007) and game-based Internet interventions (Coyle, et al., 2011). Researchers (Doherty, et al., 2012) have focused on defining guidelines and strategies for such systems in order to improve usability and user engagement. The system presented in this paper has implemented some of these guidelines.The aim of this project is to develop a system framework, called Moderator Assistant, to help moderators to easily monitor one or more online support communities, quickly produce interventions automatically generated by the system, and analyze individual behaviors in that online group.
The projects are highly multidisciplinary and we are looking for students with a computer science/software engineering background and strong interest in psychology, or viceversa candidates with a background in psychology but with a professional interest in technology.The candidate would work in close collaboration with professionals at the Inspire Foundation and the Young and Well CRC. Furthermore this collaboration means the work will have a significant impact on the way mental health support is delivered in Australia.
The opportunity ID for this research opportunity is 1825