Skip to content

BillPei/question_answering_system

 
 

Repository files navigation

Question Answering System

Summary

The goal of this project is to build a question answering system that performs textual reasoning to answer and pass a 4th grade science exam.

System Architecture

Process Frame

We define process frame as a structure that captures complete information of a scientific process in a structured format. The frame element that we are using currently contains following frame elements.

  • Undergoer
  • Cause
  • Manner
  • Result
System Overview

The system takes descriptive question and four answer choices as input. This is then processed by subsystems shown in the following figure to produce an answer and an associated confidence score.

System

Example

The following image gives some intuition on how and why structural alignment helps question answering.

Frame Alignment

Code

The following modules are required to run the system:

  • Python 2.7
  • NumPy
  • Pandas
  • Scipy
  • scikit-learn
  • nltk
  • Matplotlib
  • Requests
  • JSON

and textual alignment service by AI2 (Allen Institute for Artificial Intelligence).

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%