Microsoft Researchers Introduce a Theoretical Framework Using Variational Bayesian Theory Incorporating a Bayesian Intention Variable
What is the purpose of Microsoft's Bayesian behavior framework?
Microsoft's Bayesian behavior framework aims to synergize habitual and goal-directed behaviors in biological and artificial agents using variational Bayesian methods3. It introduces a Bayesian latent variable called "intention" to represent dynamic intentions that can adjust based on sensory cues (habitual) and specific goals (goal-directed), allowing a seamless transition and interaction between the two behavior types6.
How does goal-directed behavior differ from habitual behavior?
Goal-directed behavior is a deliberate, conscious action taken to achieve a specific outcome, and it can adapt to changes in context or value of the outcome. Habitual behavior, on the other hand, is an automatic, subconscious response to a situation based on prior experiences, and it is inflexible and less sensitive to the context or outcome value.
What is habitual behavior in decision-making?
Habitual behavior in decision-making refers to automatic responses that are deeply ingrained through experience and require little to no conscious thought. These behaviors are fast, automatic, and model-free, often driven by sensory cues and past rewards. Unlike goal-directed behavior, which requires deliberate planning and action to achieve specific outcomes, habitual behavior relies on established patterns and is less influenced by the current value of the outcome.