Data

Tools

Indexes

Applications

Experiments

Open Source Science

learning-of-symbolic-representations-for-rescue-scenarios-in-disaster-zones

This repository contains the code and resources for my Bachelor's Thesis

https://github.com/diabahmed/learning-of-symbolic-representations-for-rescue-scenarios-in-disaster-zones

Science Score: 44.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ✓
    CITATION.cff file
    Found CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ✓
    .zenodo.json file
    Found .zenodo.json file
  • ○
    DOI references
  • ○
    Academic publication links
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (11.4%) to scientific vocabulary

Keywords

ai knowledge-ex machine rein
Last synced: 6 months ago · JSON representation ·

Repository

This repository contains the code and resources for my Bachelor's Thesis

Basic Info
  • Host: GitHub
  • Owner: diabahmed
  • License: mit
  • Language: Jupyter Notebook
  • Default Branch: main
  • Homepage:
  • Size: 2.61 MB
Statistics
  • Stars: 0
  • Watchers: 1
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Topics
ai knowledge-ex machine rein
Created almost 3 years ago · Last pushed over 2 years ago
Metadata Files
Readme License Citation

README.md

Learning of Symbolic Representations for Rescue Scenarios in Disaster Zones

This repository contains the code and resources for my Bachelor's Thesis titled "Learning of Symbolic Representations for Rescue Scenarios in Disaster Zones". The project focuses on the application of Neuro-Symbolic AI to extract symbolic interpretable knowledge in disaster response scenarios.

Project Overview

The project aims to train an AI agent to navigate complex disaster scenarios, such as the aftermath of an earthquake or a flood, and extract valuable symbolic knowledge from its experiences. The methodology involves two main steps:

  1. Q-Learning: A type of reinforcement learning used to train the agent to navigate a grid-based environment representing a disaster zone.
  2. Symbolic Knowledge Extraction: A process of transforming the learned knowledge into a human-readable format, providing interpretable insights into the agent's decision-making process.

Repository Structure

  • Knowledge Extraction (QLearning): Contains the source code for the project.
  • output_demo: Contains a sample demo the results of the experiments.

Getting Started

Prerequisites

  • Python 3.7 or higher
  • NumPy
  • OpenAI Gym

Installation

  1. Clone the repository: bash git clone https://github.com/ahmillect/Learning-of-Symbolic-Representations-for-Rescue-Scenarios-in-Disaster-Zones.git
  2. Navigate to the project directory: bash cd Learning-of-Symbolic-Representations-for-Rescue-Scenarios-in-Disaster-Zones
  3. Install the required packages: bash pip install gym numpy

Usage

  • To train and validate the agent, run: python Knowledge Extraction (QLearning).ipynb

Results

The agent showed significant improvement in performance over 1 million episodes of training, with a substantial increase in cumulative reward. The extracted symbolic knowledge, validated using a rule validation script, showed that a significant percentage of the extracted rules were valid.

Future Work

Potential areas for further research include exploring different reinforcement learning algorithms, refining the rule validation process, expanding the complexity of the environment, and enhancing the symbolic knowledge extraction process.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

I would like to express my deepest appreciation to my supervisor, Dr. Nourhan Ehab, for her guidance and support throughout the course of this thesis.

Owner

  • Name: Ahmed Diab
  • Login: diabahmed
  • Kind: user
  • Location: Berlin, Germany
  • Company: German International University

Software Engineer

Citation (CITATION.cff) 

# This CITATION.cff file was generated with cffinit.
# Visit https://bit.ly/cffinit to generate yours today!

cff-version: 1.2.0
title: >-
  Learning of Symbolic Representations for Rescue Scenarios
  in Disaster Zones
message: >-
  If you use this software, please cite it using the
  metadata from this file.
type: software
authors:
  - given-names: Ahmed Khaled
    email: ahmillect@outlook.com
    affiliation: The German University in Cairo
repository-code: >-
  https://github.com/ahmillect/Learning-of-Symbolic-Representations-for-Rescue-Scenarios-in-Disaster-Zones
abstract: >-
  The development of intelligent search and rescue systems
  is of paramount importance in disaster response scenarios,
  as rapid and effective assistance can save lives and
  mitigate damage. In this thesis, we present a novel
  approach that combines reinforcement learning with
  symbolic knowledge extraction to create an autonomous
  agent capable of navigating complex environments and
  making informed decisions in a disaster zone, specifically
  in our custom-designed DisasterZone environment. By
  leveraging the strengths of Q-learning and Ron Sun's
  \cite{sun1999knowledge} approach to extracting IF-THEN
  rules from the learned policy, we aim to develop a more
  interpretable and explainable model for disaster response
  tasks, enabling better collaboration between humans and
  artificial intelligence systems. Our findings reveal that
  the training time of the agent is proportional to the
  complexity of the environment, showcasing the efficiency
  of our approach. However, limitations and challenges arise
  from the lack of computational power, as training is
  conducted using a CPU-based approach rather than a GPU
  one.
keywords:
  - Neuro-Symbolic AI
  - Q-Learning
  - Reinforcement Learning
license: MIT

GitHub Events

Total
Last Year