What do hyenas have to do with software engineering? Computer science has a long tradition of attempting to emulate the behaviors and intellectual capacities of complex biological organisms. But, if we want to create intelligent, complex, and interactive computational systems, we should look to natural systems not just for inspiration, but for genuine understanding of how nature has addressed these challenges already. Striped hyenas are a good model species to look to as they are a member of the (informal) class of incipiently social and socially flexible carnivores. Animals in this group exhibit a unique range of behavioral flexibility (they work well in groups or alone), they exhibit high levels of intelligence (they can store, recall, and use incredibly complex memory streams), and they operate within complex networks of interacting competitive species. So, expanding our capacities for engineering truly intelligent systems (e.g., ‘robot hyenas’) by capitalizing on our understanding of complex natural systems, processes, and dynamics may not be such a crazy goal to have.
In the other direction, what does computer science have to do with hyenas? New computational tools and research approaches can help biological researchers test hypotheses about natural systems that they could not address via experiments in nature, including conservation scenarios. Additionally, exploring the intersection of biology and computer science (again, robot hyenas) creates unique opportunities to inspire and excite young students about science and engineering.
We are developing the digital evolution platform Avida as a tool for furthering our understanding of the evolution of complex behaviors, such as group formation, cooperation, and predator-prey dynamics. In this work, we are expanding the ecological ‘sophistication’ of Avida, evaluating the potential of new virtual hardware architectures for permitting the evolution of truly dynamic, intelligent behavioral control algorithms in autonomous agents, and testing hypotheses regarding the role of environmental conditions and competition in promoting the evolution of complexity and intelligence. This work carries important implications for artificial intelligence, in addition to biology. For example, if we are to develop truly intelligent systems (e.g. systems that can predict and respond to conditions beyond those it is explicitly engineered to deal with), what are the most effective means for promoting the development of intelligent traits? And what are the most effective architectures for such systems? Can we evolve such systems, avoiding the assumptions and inherent restraints of engineering them solely from the bottom up? If so, what environmental factors promote their evolution?
A sample of an evolved predator ‘brain’. This execution network illustrates the actions taken and routes traversed through it’s genome over its lifetime. The new genetic architectures we’ve developed for Avida produce non-linear behavioral, contingently responsive execution networks. Nodes represent instructions. Edges connect sequentially executed instructions. Node and edge weights indicate frequencies of instruction execution and behavioral pathway traversal, respectively.
In a previous life, I studied the incipiently social striped hyena Hyaena hyaena as a model species for evolutionary ecology research focused into the origins of sociality. Striped hyenas are valuable for addressing questions about the evolution of group formation, a prerequisite for the evolution of true social and cooperative behaviors. My research focused on identifying the set(s) of environmental conditions (e.g. resource distributions and abundances across time and space) that determine whether the sharing of resources is an evolutionary ‘permitted’ strategy and, once permitted, what additional external pressure promote the further development of cooperative social strategies.
Grandma greeting her grandson at her new den. Striped hyenas exhibit a surprising level of flexibility in social interactions and systems, from strictly solitary to cooperative raising of young, depending on food resource conditions.