Master thesis structure

This is a structure tree of the master thesis. The topics are selected from master-thesis-topic-list. The length of the master thesis should be ~60 pages

Alternative table of contents

abstract

• Introduction

  • (???) The importance of knowledge representation for scientific progress in physics

  • what-is-a-semantic-network

  • What is an ontology

  • Types of semantic networks

  • How is physics knowledge structured

  • Goal of the master thesis

  • Overview

• Structuring the different subject areas of physics

  • Collecting physics terms (4 Pages)

    • most-important-terms-of-topic-question

    • Using the Wikipedia page titles

    • Manual selection

    • Physics term generation using LLMs

  • Creating a subtopic tree (11 Pages)

    • What is a subtopic tree

    • [subtopic-tree-algorithms-description

    • visualization-of-the-subtopic-trees

    • Subtopic tree evaluation methods](../approaches/subtopic-tree-algorithms-description.md)

    • physics-term-adapted-subtopic-tree-evaluation

    • physics-term-adapted-subtopic-tree-evaluation-free-association-algorithm

    • subtopic-tree-comparison + comparison-of-frAss-and-subd-for-different-parameters

    • subtopic-network-search-model-comparison

• Representing physical knowledge as semantic triples

  • Generating semantic triples of natural language (16 Pages)

    • What are the criteria for well formed triples (see evaluation process validness)

    • physics-term-correlation-network-evaluation

    • Triple generation with LLMs

      • Correlated triple generation correlated-triples-evaluation

      • Freely associated triples free-associated-triples-evaluation

      • comparison-of-frAss-and-subd-for-different-parameters

    • Evaluation metric for the correlated triples

    • Manual triple evaluation

    • Displaying the triples in a 2D plane

• Answering physical questions with semantic networks

  • Navigating the knowledge base

    • Navigating the triples

  • Creating a semantic network of questions and answers

    • generating-networks-of-questions-and-answers

    • checking-the-knowledge-consistency-of-large-language-models

• Extracting semantic data from scientific texts

  • Building a citation graph using the semantic-scholar-api

  • (3 Pages)

    • hepml-citation-graph

  • Triple generation using Stanford open IE

    • Paper content extraction

    • Quality evaluation

  • converting-sentences-into-semantic-triples

  • Creating a semantic network of sentences (5 Pages)

    • Manual creation of a semantic network of sentences

    • Automated creation of a semantic network of sentences

• Creating a physics ontology

  • Creating a specified semantic network (4 Pages)

    • Introduction to the topic area of top tagging

    • Creating a physics ontology

    • Extracting the R30 values

    • Inserting them into the ontology

  • Creating a general physics ontology with the help of LLMs

    • generating-a-physics-ontology

    • generate-knowledge-graph-based-on-an-ontology

  • (???) Combining different types of semantic networks (???)

• Handling equations with semantic networks

  • Creating a semantic network of equations (5 Pages)

    • Technical realization

    • Generating the network of equation with LLMs

    • Combining multiple equations into one network

  • Representing a derivation as knowledge graph

• Summary and outlook

  • Summary (1 Page)

  • Possible applications (4 Pages)

  • Conclusion (1 Page)

The old table of contents