Bachelor Thesis Compression-Based Authorship Verification with Large Language Models

Fraunhofer-Gesellschaft

ThesisGerman needed: not said

Still open on 1 Nov

You apply on the company's own website.

What you'll do

Goal
The goal of the thesis is to investigate for several classical AV methods to what extent central components can be replaced or supplemented by LLMs. A particular focus is on compression-based methods. The next-token probabilities provided by LLMs are to be used instead of conventional compression models and integrated into existing AV methods. Furthermore, the relationship between the modeling or compression quality of an LLM and the resulting performance of the AV method is to be investigated. Results from related tasks suggest that this relationship is not necessarily monotonic (4).
Expected Results
In the course of the work, several classical AV methods are to be conceptually transferred to the use of Large Language Models and prototypically implemented. The developed variants are then to be evaluated on several suitable AV corpora and compared with the respective classical methods. In particular, the effects of different LLMs and their modeling quality on the verification performance are to be analyzed.

The company did not list what you need. Open the job on the company site to read it.

Goal
The goal of the thesis is to investigate for several classical AV methods to what extent central components can be replaced or supplemented by LLMs. A particular focus is on compression-based methods. The next-token probabilities provided by LLMs are to be used instead of conventional compression models and integrated into existing AV methods. Furthermore, the relationship between the modeling or compression quality of an LLM and the resulting performance of the AV method is to be investigated. Results from related tasks suggest that this relationship is not necessarily monotonic (4).
Expected Results
In the course of the work, several classical AV methods are to be conceptually transferred to the use of Large Language Models and prototypically implemented. The developed variants are then to be evaluated on several suitable AV corpora and compared with the respective classical methods. In particular, the effects of different LLMs and their modeling quality on the verification performance are to be analyzed.

The company did not list what you need. Open the job on the company site to read it.