Bandit Multiclass Linear Classification: Efficient Algorithms for the Separable Case
International Conference on Machine Learning, pp. 624-633, 2019.
We study the problem of efficient online multiclass linear classification with bandit feedback, where all examples belong to one of $K$ classes and lie in the $d$-dimensional Euclidean space. Previous works have left open the challenge of designing efficient algorithms with finite mistake bounds when the data is linearly separable by a ma...More
PPT (Upload PPT)