Adaptive Text Summarisation for Broadening Access to Technical Texts

Technical and specialist texts play a crucial role in documenting and advancing various fields, institutions, and academic disciplines. However, these texts often assume a certain level of prior knowledge, relying on domain-specific concepts and language that can be difficult for non-experts to understand. Text Summarisation techniques aim to retrieve and express the key points of a text within a coherent, informative summary. When trained or prompted appropriately, these techniques can also make textual information more accessible by adapting content to suit different audiences. Therefore, when applied to technical documents, summarisation has the potential to help bridge the gap between experts and non-experts, expanding access to specialised knowledge. However, the complexity and unique characteristics of such texts present significant challenges for automatic processing, limiting the effectiveness of modern summarisation systems.

This thesis explores strategies to enhance the application of summarisation models to technical texts.
Furthermore, significant emphases are placed on broadening accessibility to non-technical audiences, improving content understanding through detailed analyses, and exploring how relevant external knowledge can be leveraged to improve model outputs.