Rajesh Parekh

ICML Conference 1999 Conference Paper

Simple DFA are Polynomially Probably Exactly Learnable from Simple Examples

• Rajesh Parekh
• Vasant G. Honavar

AAAI Conference 1996 Short Paper

An Incremental Interactive Algorithm for Regular Grammar Inference

• Rajesh Parekh

Grammar inference, a problem with many applications in pattern recognition and language learning, is defined as follows: For an unknown grammar G, given a finite set of positive examples S+ that belong to L(G), and possibly a finite set of negative examples S-, infer a grammar G* equivalent to G. Different restrictions on S+ and S- and the interaction of the learner with the teacher or the environment give rise to different variants of this task. We present, an interactive incremental algorithm for inference of a finite state automaton (FSA) corresponding to an unknown regular grammar.

AAAI Conference 1996 Short Paper

Constructive Neural Network Learning Algorithms

• Rajesh Parekh

Constructive Algorithms offer an approach for incremental construction of potentially minimal neural network architectures for pattern classification tasks. These algorithms obviate the need for an ad-hoc apriori choice of the network topology. The constructive algorithm design involves alternately augmenting the existing network topology by adding one or more threshold logic units and training the newly added threshold neuron(s) using a stable variant of the perceptron learning algorithm (e.g., pocket algorithm, thermal perceptron, and barycentric correction procedure). Several constructive algorithms including tower, pyramid, tiling, upstart, and perceptron cascade have been proposed for a-category pattern classification. These algorithms differ in terms of their topological and connectivity constraints as well as the training strategies used for individual neurons.