Proxy Modeling

A proxy is defined as something that approximates, i.e. acts as a proxy for, something else.

When we are modeling in NLP we can pay attention to the 'Behavioral Manifestations of Internal Representations' or BMIR's of the model. We can think of the model’s BMIRs as “proxies” for the underlying state of the model.

By adopting one or more of the BMIRs of the model, we can attempt to replicate a portion of their internal state, and experience in whole or part, the behavior or skill with this proxy state.

Implicit Modeling is an example of using physiology as a proxy for the model’s internal state.

NLP Training in New York