- Seminar: BLIND AND OFF-GRID ACOUSTIC ECHOES RETRIEVAL USING MULTICHANNEL ANNIHILATING FILTERS
The aim of this master thesis is to provide an easy to use interface to convert course notes written in e.g. LaTeX into a graph representation and back. This will require the following steps.
- Determine the graph structure to represent knowledge and test these graph structures on various use-cases. This results in the definition of a generic graph structure to represent the knowledge to be learned.
- Determine how it is possible to use the capabilities of LaTeX (e.g. using labels and macros) to get a (preferably one-to-one) mapping between the LaTeX document and the graphs structure. Specific LaTeX macros can be added to reinforce the structure of the text, but the their number should be limited to a minimum in order to reduce the learning curve of a author.
- Write a LaTeX parser in Python to import the LaTeX file and to store the graph representation in the neo4j graph database.
- Write a LaTeX and pdf/html document generator starting from a sub-graph extracted from the knowledge graph database.
The aim is to make a system that does all the above steps: parsing a LaTeX input file which is (as closely as possible compatible with) the LaTeX files of the course, generating a knowledge graph stored in neo4j, and be able to generate subparts of the knowledge graph is a user readable format (e.g. LaTeX, pdf, html…).
Knowledge graph extracted from LaTeX input and stored in Neo4j