Index
Research
Elsewhere

About the Rowan Martin-Hughes in question

Meta-optimization
Making tools that optimize the optimization of limited resource allocation.

Also hosted is my father's newly renovated homepage displaying some of the pottery which I have had the pleasure of growing up with: here.

Research interests

See my bio at the Burnet Institute for a current list of publications relating to the optimization of resources dedicated to eliminating or controlling communicable diseases, using mathematical models for:
  • HIV
  • Tuberculosis (TB)
  • Malaria
  • Maternal, newborn, and child health (MNCH)
    • with additional models under development.

    Past research interests

    Having been awarded a PhD in 2010 and moved into the development sphere and out of academia (**or so I thought), the below is no longer current but nonetheless relevant:

    I'm interested in combining the trust benefits of small communities with the convenience benefits of the global community in an online setting. I'm using various aspects of game theory and bits and pieces from all over the place in an attempt to create a better reputation management system which can convince people to Do The Socially Acceptable Thing online.

    One way I would describe this is that I have a background in mathematics, work in a computer science department, attend the odd philosophy or political theory seminar, read economics papers, learn aspects of psychology, and think deeply about sociology. In case I've missed anything I then socialise with chemists, engineers and biologists to name a few. After that I wrap it into a bundle that looks like game theory and call it a day.

    What do I hope to achieve from the thinly spread attention to each of this assortment of fields? As the demographics of the world are changing I think the social complexity of the world is quickly reaching the limit of where humans can not make decisions which look sensible to even themselves because of the many layers of obfuscation involved in the unconscious calculation of payoffs from different possible actions.

    The changing structure of networks also means that a number of fitness strategies are no longer as beneficial as they once were, so defection in games becomes more commonplace. From here I feel the approach of single-mindedly labelling people as "good" or "bad" or on any such scale is largely insufficient (or inefficient at least) because it neglects the fact that people who make such "bad" choices have good individual reasons to do so, meaning that finding more efficient ways to punish defectors will only have a very limited success, particularly in the online and pseudonymous domain where punishments can only feasibly be very low anyway.

    Submitted PhD Thesis Title/Abstract

    Agent Strategies in Online Communities

    Economic models and mechanism design have given many techniques for inducing cooperative behaviour in agents (either humans or software designed to interact with humans), under a wide variety of circumstances. However, evolutionary theory and psychological studies give very different motivations for human behaviour.

    This thesis works towards integrating these perspectives, allowing for greater understanding of what makes online communities such as eBay function successfully. Models of reputation systems and other mechanisms are studied from the perspectives of future discounting, evolutionary pressure, agent learning, and community design.

    Through simulation, analysis, and discussion, the roles of agents in allowing a community to function are identified. This allows for suggestions for how agents can best be persuaded to form cooperative social norms. Conversely, the different roles which a community plays in providing benefits to agents are identified. For each of these, a different set of mechanisms is appropriate as the community grows.

    Elsewhere online

    Given a sufficiently unique hyphenated surname, I haven't come across any imposters yet, but just in case of any confusion: I don't have an account with any social networks.

    Burnet Institute: Employer since 2017 (although association with the research group started earlier than that), continuing the themes of computer science with developing/maintaining disease models and scripts to use those models, development work optimizing spending of national disease budgets in coordination with national country teams, and a great deal of analysis of results and data to ensure everything makes sense.

    CCC: The Cooperation Committee for Cambodia, an umbrella organisation for NGOs in Cambodia with the goal of "a strong and capable civil society, cooperating and responsive to Cambodia's development challenges", where I worked as a Knowledge Management Advisor from 2011 to 2014.

    Last.fm: A fairly complete musical profile since 2005 although getting less complete as less listening is done via locally saved music that gets recorded!

    Livejournal: It doesn't say anything about the present, but there are remnants of a PhD student of a great many years ago appears to be still partially available (sans image hosting) as of last review.