Priors and Probabilistic Influences on Perception



Poppy Crum is Chief Scientist at Dolby Laboratories. She also holds an appointment as Adjunct Professor at Stanford University in the Center for Computer Research in Music and Acoustics and the Program in Symbolic Systems. At Dolby, Poppy directs the growth of internal science. She is responsible for integrating neuroscience and sensory data science into algorithm design, technological development, and technology strategy. At Stanford, her work focuses on the impact and feedback potential of new technologies including gaming and immersive environments on neuroplasticity and learning. In her talk, Poppy looks at these points...

1. Our perceived experience of sensory information in the world is malleable, contextually and experientially dependent, probabilistic in approximate description, and predictable with some uncertainty.
2. Our priors differentiate each of our perceptual experiences across a constant input to our sensory systems. In this way, bias from our priors is one component influencing how we collectively most often end up with a heterogeneous experience of homogeneous sensory input.
3. For technology to optimally engage and enhance each of our individual human systems it must be: personalized to our sensory/biological baselines, learning and integrating semantic and contextual information of our surroundings and internal state, and interfacing effective dimensionally reduced data to the user.
4. Closing the technology/physiology loop and allowing always on sensors to objectify our internal states has great possibilities, but it means the age of the poker face may come to an end. How will this change our social behaviors? How can we use this to enhance the qualities that make us most human?