Navigation auf


Institut für Computerlinguistik

Tannon Kew

Tannon Kew, M.A.

  • PhD candidate

I am a PhD student working on natural language generation. My main focus is on controlled conditional text generation and investigating interesting approaches to tasks such as automatic text simplification, conversation dialogue modelling and review response generation.

In my current role, I have worked on a range of projects related to natural language generation tasks. These include:

  • The Innosuisse-funded project ReAdvisor, in which we developed review response generation models to support authors in writing high-quality responses to online customer reviews in the hospitality industry.
  • Developing the 20Minuten dataset for German news summarization and text simplification.
  • Target-level text simplification, which aims to automatically rewrite complex language into something that is easier to read and understand for specific audiences. In this area, I am particularly interested in leveraging different control methods that allow for guiding the generated simplfications towards a particular target audience.
  • Investigating the role of sequence-to-sequence pre-training objectives on zero-shot control methods for conversational dialogue models.
  • Evaluating LLMs on the task of text simplification.
  • Developing enterprise LLMs for German language settings (research internship atTextshuttle).


ZORA Publication List

Download Options