Pilot Ai Labs, Inc.

Salesforce Com, Inc.

Ankit Kumar

Jonathan Su

Brian Pierce

Elliot English

Caiming Xiong

Richard Socher

Romain Paulus

Stephen Merity

Ozan Irsoy

Mohit Iyyer

A novel unified neural network framework, the dynamic memory network, is disclosed. This unified framework reduces every task in natural language processing to a question answering problem over an input sequence. Inputs and questions are used to create and connect deep memory sequences. Answers are then generated based on dynamically retrieved memories.

Dynamic memory network

In general, certain embodiments of the present disclosure provide methods and systems for object detection by a neural network comprising a convolution-nonlinearity step and a recurrent step. In a training mode, a dataset is passed into the neural network, and the neural network is trained to accurately output a box size and a center location of an object of interest. The box size corresponds to the smallest possible bounding box around the object of interest and the center location corresponds to the location of the center of the bounding box. In an inference mode, an image that is not part of the dataset is passed into the neural network. The neural network automatically identifies an object of interest and draws a box around the identified object of interest. The box drawn around the identified object of interest corresponds to the smallest possible bounding box around the object of interest.

System and method for improved general object detection using neural networks

Growing community of inventors

San Diego, CA, United States of America

Ankit Kumar