Learning twig and path queries

We investigate the problem of learning XML queries, path queries and twig queries, from examples given by the user. A learning algorithm takes on the input a set of XML documents with nodes annotated by the user and returns a query that selects the nodes in a manner consistent with the annotation. We study two learning settings that differ with the types of annotations. In the first setting the user may only indicate required nodes that the query must select (i.e., positive examples). In the second, more general, setting, the user may also indicate forbidden nodes that the query must not select (i.e., negative examples). The query may or may not select any node with no annotation. We formalize what it means for a class of queries to be learnable. One requirement is the existence of a learning algorithm that is sound i.e., always returning a query consistent with the examples given by the user. Furthermore, the learning algorithm should be complete i.e., able to produce every query with sufficiently rich examples. Other requirements involve tractability of the learning algorithm and its robustness to nonessential examples. We identify practical classes of Boolean and unary, path and twig queries that are learnable from positive examples. We also show that adding negative examples to the picture renders learning unfeasible.