J. Branson Byers

J. Branson Byers

Graduate Student

Hello! I’m Branson. I’m interested in reward. In particular, I think about how much choice people have in the rewards they go after and experience. Sometimes I think people can shape what is rewarding to them. Other times it might be very difficult. I think the amount of control people have over what is rewarding to them ultimately effects what they do. It might effect the goals we set for ourselves or how we go about pursuing those goals as well. What is rewarding to us also might play a role in how we see ourselves and the stories we tell about who we think we (and others) are. This all goes to say that I think reward seems to be related to our identities. It’s worth mentioning that maybe our identities shape what we find rewarding as well.

There are many related questions to this line of thinking. How does our control over reward help or hurt our mental health? When do narratives help or hurt us? Might this be related to mindfulness?

I approach this line of thinking using computational models. I think computational models are great ways to quantify how people act. I also think they are a great tool that can help scientists be specific about what they think is going on inside peoples’ heads when we make choices in our lives.

I am most familliar with methods like Bayesian Inference and Reinforcement Learning. Bayesian methods provide a very intuitive way of testing your ideas as a scientist, since you can see how much evidence supports your idea (hypothesis) directly, and compare that idea directly to other ideas you might have about what is going on. Other methods are better at ruling out out ideas, rather than directly testing them, but not everyone agrees about this.

Reinforcement Learning describes a way that machines can learn. Some people think it is similar to how people learn, but only in certain circumstances. I like Reinforcement Learning because these models (if designed well) are often interpretable, meaning people can look under the hood and understand what they are doing.

If you think these things are interesting too, totally disagree with me, or just want to say hello, send me an email. I’d love to talk.