Heeding Good Advice: Scaling Down and Specializing in the Age of Big AI

In a position paper published last year, Togelius and Yannakakis describe a depressing aspect of the current artificial intelligenceartificial intelligence(AI) landscapelandscape, particularly as it relates to academic research. They outline some of the challenges faced by researchers working in traditional academic institutions, that are finding it increasingly difficult to compete with large transnational companies with comparatively unlimited resources. Togelius and Yannakakis take a proactive stance, providing a series of thoughtful insights and survival strategies for depressed AI academics. In this chapter, we discuss some of the advice offered by Togelius and Yannakakis, from the perspective of research work in genetic programming (GP). Moreover, the chapter describes two research projects we are currently developing that align with two of the survival strategies they put forth: reducing the scale and focusing on specialized domains. We conclude that in the current era of AI research, it is important to heed good advice so that traditional academic spaces, particularly for those working with limited resources, continue to thrive and contribute toward shaping future research in AI, ML, and GP.