Provably and Practically Efficient Neural Contextual Bandits
This work addresses a theoretical gap in bandit algorithms for neural networks, offering improved generality and practical performance for machine learning applications.
The paper tackles the neural contextual bandit problem by extending analysis to smooth activation functions beyond ReLU, deriving non-asymptotic error bounds and proposing an algorithm with provable sublinear regret and empirical efficiency.
We consider the neural contextual bandit problem. In contrast to the existing work which primarily focuses on ReLU neural nets, we consider a general set of smooth activation functions. Under this more general setting, (i) we derive non-asymptotic error bounds on the difference between an overparameterized neural net and its corresponding neural tangent kernel, (ii) we propose an algorithm with a provably sublinear regret bound that is also efficient in the finite regime as demonstrated by empirical studies. The non-asymptotic error bounds may be of broader interest as a tool to establish the relation between the smoothness of the activation functions in neural contextual bandits and the smoothness of the kernels in kernel bandits.