The SwarmLab at ABS2014
Just after ISBE2014, the SwarmLab will head to Princeton, NJ for the 51st Annual Conference of the Animal Behavior Society. The conference will be held at Princeton University from Saturday, August 9th to Thursday, August 14th, 2014.
On the evening of the opening day (Sunday, August 10th), PI Simon Garnier will co-organize a workshop “Quantitative analysis of collective behavior – from the lab to the wild” with Prof. Iain Couzin, Dr. Andrew King, Dr. Nicolas Perony and Dr. Damien Farine. This workshop will conclude a day started with the keynote address of Prof. Iain Couzin on “Collective sensing and decision making in animal groups”.
The same day, Princeton PhD candidate Matthew Lutz will present the work he has performed on the living architectures of New World army ants, in collaboration with PI Simon Garnier and post-doc Chris Reid from the SwarmLab. His presentation will be at 10.45 am in the Taplin Auditorium (Fine Hall, Princeton University). In the afternoon, post-doc Chris Reid will present his work on the problem-solving abilities of the slime mold Physarum polycephallum at 5.45pm in Jadwin Hall (room 10).
Hereafter is the abstract for Chris’ presentation:
How slime mould cracks the Two-Armed-Bandit problem; insights into unicellular decision making
Should I exploit well-known options, or do I risk further exploration for potentially higher reward? This is the exploration-exploitation tradeoff, and while it faces casino gamblers and foraging organisms alike, the optimal solution is unknown. The tradeoff has been studied using the 2-Armed Bandit problem, where a player aims to maximize their gain when faced with two slot machines, each with a distinct but unknown reward rate. Studies thus far have only used organisms with brains. We tested the slime mold Physarum polycephalum with the 2-Armed Bandit problem by assessing the effect of sampling on foraging patch choice in a T-maze.We generate insight into the basic processes of decision making in a cell, including the use of relative vs absolute reward criteria (in both the frequency of reward, and the combination of frequency and magnitude), and the effect of static vs dynamic exploration environments. We propose several biologically plausible decision criteria the organism may be using and, through Bayesian inference, determine which of these models best explains the empirical data. We challenge the common view that neurological hardware is required to solve complex problems.
