In particular, they study batch-mode active learning from two aspects: (1) Suppose the batch sizes are equal. Fix the number of rounds examined, how many labels one has to request in total to achieve some certain error rate; (2) given that the cost to obtain a batch of labels is sublinear in the size of the batch ( as referred to as ``buy-in-bulk discount''), how is the total cost of the proposed batch algorithms compared with that of fully-sequential active learning methods. For the first aspect, the authors propose batch-based variants of the well-known CAL algorithm for sequential active learning, and provide upper bounds on label complexity of k-batch active learning, for both the realizable case and the non-realizable case (with Tsybakov noise). For the second aspect, they provide a cost-adaptive modification of the CAL algorithm, and find that the total cost by this algorithm may often be significantly smaller than that of the analogous methods in the fully sequential setting.

Neural Information Processing Systems http://nips.cc/

In many practical applications of active learning, it is more cost-effective to request labels in large batches, rather than one-at-a-time. This is because the cost of labeling a large batch of examples at once is often sublinear in the number of examples in the batch. In this work, we study the label complexity of active learning algorithms that request labels in a given number of batches, as well as the tradeoff between the total number of queries and the number of rounds allowed. We additionally study the total cost sufficient for learning, for an abstract notion of the cost of requesting the labels of a given number of examples at once. In particular, we find that for sublinear cost functions, it is often desirable to request labels in large batches (i.e., buying in bulk); although this may increase the total number of labels requested, it reduces the total cost required for learning.

In many practical applications of active learning, it is more cost-effective to request labels in large batches, rather than one-at-a-time. This is because the cost of labeling a large batch of examples at once is often sublinear in the number of examples in the batch. In this work, we study the label complexity of active learning algorithms that request labels in a given number of batches, as well as the tradeoff between the total number of queries and the number of rounds allowed. We additionally study the total cost sufficient for learning, for an abstract notion of the cost of requesting the labels of a given number of examples at once. In particular, we find that for sublinear cost functions, it is often desirable to request labels in large batches (i.e., buying in bulk); although this may increase the total number of labels requested, it reduces the total cost required for learning.