Stochastic approximation in non-markovian environments revisited
This work addresses foundational challenges in machine learning for researchers and practitioners dealing with complex, non-stationary data in transformers and continual learning.
The paper tackles the problem of stochastic approximation in non-Markovian and non-ergodic environments, proposing an analytic framework to understand transformer-based learning, specifically the attention mechanism, and continual learning, which rely on the entire past.
Based on some recent work of the author on stochastic approximation in non-markovian environments, the situation when the driving random process is non-ergodic in addition to being non-markovian is considered. Using this, we propose an analytic framework for understanding transformer based learning, specifically, the `attention' mechanism, and continual learning, both of which depend on the entire past in principle.