Knowledge Adaptation as Posterior Correction
This work addresses the challenge of machine adaptivity, which is incremental as it reframes existing methods under a unified Bayesian framework.
The paper tackles the problem of enabling AI models to adapt quickly like humans by showing that various adaptation methods can be viewed as correcting approximate posteriors, with more accurate posteriors leading to smaller corrections and faster adaptation.
Adaptation is the holy grail of intelligence, but even the best AI models (like GPT) lack the adaptivity of toddlers. So the question remains: how can machines adapt quickly? Despite a lot of progress on model adaptation to facilitate continual and federated learning, as well as model merging, editing, unlearning, etc., little is known about the mechanisms by which machines can naturally learn to adapt in a similar way as humans and animals. Here, we show that all such adaptation methods can be seen as different ways of `correcting' the approximate posteriors. More accurate posteriors lead to smaller corrections, which in turn imply quicker adaptation. The result is obtained by using a dual-perspective of the Bayesian Learning Rule of Khan and Rue (2023) where interference created during adaptation is characterized by the natural-gradient mismatch over the past data. We present many examples to demonstrate the use of posterior-correction as a natural mechanism for the machines to learn to adapt quickly.