Microsoft Technology Licensing, LLC

Maluuba Inc.

Zheng Ye

Adam Trischler

Xingdi Yuan

Philip Bachman

Kaheer Suleman

Alessandro Sordoni

Jing He

Jean Merheb-Harb

Examples of the present disclosure provide systems and methods relating to a machine comprehension test with a learning-based approach, harnessing neural networks arranged in a parallel hierarchy. This parallel hierarchy enables the model to compare the passage, question, and answer from a variety of perspectives, as opposed to using a manually designed set of features. Perspectives may range from the word level to sentence fragments to sequences of sentences, and networks operate on word-embedding representations of text. A training methodology for small data is also provided.

Parallel-hierarchical model for machine comprehension on small data

Described herein are systems and methods for providing a natural language comprehension system that employs a two-stage process for machine comprehension of text. The first stage indicates words in one or more text passages that potentially answer a question. The first stage outputs a set of candidate answers for the question, along with a first probability of correctness for each candidate answer. The second stage forms one or more hypotheses by inserting each candidate answer into the question and determines whether a sematic relationship exists between each hypothesis and each sentence in the text. The second processing circuitry generates a second probability of correctness for each candidate answer and combines the first probability with the second probability to produce a score that is used to rank the candidate answers. The candidate answer with the highest score is selected as a predicted answer.

Machine comprehension of unstructured text

Disclosed embodiments include systems and methods relevant to improvements to natural language processing used to determine an intent and one or more associated parameters from 5 a given input string. In an example, an input string is received and first and second different n-grams are applied to the input string. Recurrent neural network models are then used to generate output data based in part on the first and second different n-grams. In particular embodiments a recurrent neural network in both forward and backward directions specific to unigrams is applied. Intent detection and semantic labeling are applied to the output of the recurrent neural network models.

Method and system for processing an input query using a forward and a backward neural network specific to unigrams

Growing community of inventors

Wuhan, China

Zheng Ye