Observer-dependent Models @ the Philosopher's Web Cafe

 

I gave a talk called "Observer-dependent Models" to the Philosopher's Web Cafe on December 11. I have made the slides available here, and the recording is here. Thanks to Jesse Parent (Orthogonal Lab Manager) and Charlotte Guo (series host) for hosting. It will be almost like being there (almost). 

The talk involved reviewing and redefining the role of observers in empirical and simulated systems. "Observer" refers mostly to computational agents (agent-based simulation and AI), although many of the ideas introduced here may apply to the analysis of empirical observations (experiments). Here is the abstract:

In many areas of science and philosophy, observers are seen as an integral part of understanding the natural world. Aside from a pedagogical role, observers are seen as less important in computational forms of inquiry. In this talk, I will reconsider a role for the observer in computational models as fully integrated with the agent. perhaps more fundamentally, causal outcomes and system dynamics are seen to be contingent on observers, while empirical observations themselves are dependent upon the actions of observers. As this is an article of faith in some interpretations of quantum mechanics, we extend this to algorithmic systems with a combinatorial solution space. The role for observers in computational and empirical investigations is established superficially using a number of concepts, including cybernetics, embodiment, and perceptual information processing. Then we will be introduced to more concrete examples of observer-oriented computational agents, such as observer-emitter systems and viewpoint networks. Finally, we will discuss how this approach goes beyond constructivism to consider multiple observers, multiple perspectives (relativism), and how they affect the interpretation of results.

There is a lot to follow up on from this talk, including a number of themes to explore within the topic of agent-based observers, with more to come in the new year.