Friday, 13 November 2015

Michael Kasumovic & Daniel Falster - A fighting framework fight: game theory vs. machine learning for strategic behaviour during animal combat

The final reading group will meet this Friday (20th November) at 1pm, We will be reading two papers on animal and plant competition, one by Michael Kasumovic and colleagues and one by Daniel Falster and colleagues.

Michael Kasumovic is an evolutionary biologist who studies the innate differences between males and females and how the environment, both social and ecological, modifies these differences. His study species include insects, spiders and humans. Daniel Falster is an evolutionary biologist and ecologist with a particular interest in using mathematical models to test fundamental ideas about the processes shaping biological communities.

Michael's paper is in Animal Behavior titled Assessment during aggressive contests between male jumping spiders, and Daniel paper can be found in bioRxiv titled Multi-trait eco-evolutionary dynamics explain niche diversity and evolved neutrality in forests

1 comment:

  1. We are starting with the spider paper - and wondering if the key thing is trying to resolve self-assessment vs mutual-assessment, wouldn't you get useful information on this by putting each spider into fight scenarios with multiple other spiders, instead of just doing one-off contests. i.e. have a round-robin tournament. Also, there are references to game theory models - could you fit these models directly to the data to see which one fitted the data better?

    So the plants paper - this was really cool and showing how diversity can evolve in a (laterally) uniform environment based on just a couple of simple rules for how plants work and how they vary (in plant height and leaf construction cost). It must have taken a lot of work to put this all together! It made some assumptions and there is the statistician's question (similar to last paper) of how well do they represent what we see and can you use data to inform the model?

    So what do these two papers have in common - a link seems to be game theoretical models, can you take ideas Falster et al used in modelling diverisification in forests and apply that sort of stuff to the combat setting. I guess we will find out at the conference!