# Ideas on how to continue the master thesis

* ==Create a free associated semantic network==.
	* Evaluate the quality and correctness of the generated semantic triples manually
		* classify them into categories (true, true in specific contexts, true but very general, false)
	* Implement a efficient [navigation mechanism](navigate-infinitely-large-semantic-network-by-choices.md) for free associated, infinitely dens semantic networks that is based on finite choices.
		* Measure how good a large language model is in finding a specific semantic triple within the free associated, infinitely dens semantic network
	* Test how good a large language model can navigate the semantic network only by looking at the predicates
	* Measure how good a large language model is in reconstructing concepts based on their semantic connections
* Convert semantic triples into sentences
* Extract semantic triples from sentences
* Explore ways to visualize the semantic network
* Explore how strong different areas of physics are connected to each other by counting the number of semantic connections that link concepts of different areas.
* Create a semantic network of questions and answers about physics topics
	* [test-ability-of-telling-apart-lies-from-truth-in-different-areas](test-ability-of-telling-apart-lies-from-truth-in-different-areas.md)
* Create a collection of the most versatil and meaningful predicates
	* Generate examples of semantic triples that are using a specific predicate
* Create a formula network
* Experimenting with merging mechanisms of concepts
* Physics ontologie creation with LLMs