Forschungsberichte Mathematische Logik Learning via Queries and Oracles Universitt at Heidelberg Mathematisches Institut Learning via Queries and Oracles
نویسنده
چکیده
Inductive inference considers two types of queries: Queries to a teacher about the function to be learned and queries to a non-recursive oracle. This paper combines these two types | it considers three basic models of queries to a teacher (QEXXSucc], QEXX<] and QEX+]) together with membership queries to some oracle. The results for each of these three models of query-inference are the same: If an oracle is omniscient for query-inference then it is already omniscient for EX. There is an oracle of trivial EX-degree, which allows nontrivial query-inference. Furthermore, queries to a teacher can not overcome diierences between oracles and the query-inference degrees are a proper reenement of the EX-degrees. In the case of nite learning, the query-inference degrees coincide with the Turing-degrees. Furthermore oracles can not close the gap between the diierent types of queries to a teacher.
منابع مشابه
Forschungsberichte Mathematische Logik Noisy Inference and Oracles Universitt at Heidelberg Mathematisches Institut Noisy Inference and Oracles
A learner noisily infers a function or set, if every correct item is presented innnitely often while in addition some incorrect data ("noise") is presented a nite number of times. It is shown that learning from a noisy informant is equal to nite learning with K-oracle from a usual informant. This result has several variants for learning from text and using diierent oracles. Furthermore, partial...
متن کاملActive Learning with Oracle Epiphany
We present a theoretical analysis of active learning with more realistic interactions with human oracles. Previous empirical studies have shown oracles abstaining on difficult queries until accumulating enough information to make label decisions. We formalize this phenomenon with an “oracle epiphany model” and analyze active learning query complexity under such oracles for both the realizable a...
متن کاملLearning from Instance-level Relationships
A large proportion of research work on supervised learning has been focused on learning from positive and negative examples. In some situations, for example in some bioinformatics applications, there exists some information about the relationships between different objects (data instances) that can be used in learning, but has been under-utilized. In this project we mainly study two such kinds ...
متن کاملA New Algorithm for Learning Range Restricted Horn Expressions
A learning algorithm for the class of range restricted Horn expressions is presented and proved correct. The algorithm works within the framework of learning from entailment, where the goal is to exactly identify some pre-fixed and unknown expression by making questions to membership and equivalence oracles. This class has been shown to be learnable in previous work. The main contribution of th...
متن کاملNoise-Tolerant Interactive Learning Using Pairwise Comparisons
We study the problem of interactively learning a binary classifier using noisylabeling and pairwise comparison oracles, where the comparison oracle answerswhich one in the given two instances is more likely to be positive. Learning fromsuch oracles has multiple applications where obtaining direct labels is harder butpairwise comparisons are easier, and the algorithm can leverage...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 1996