The State of Natural Language Feedback Generation

The State of Natural Language Feedback Generation

Thursday 16/05 10:30-12:00h

Workshop Space B

Needs Analysis

2023 was the year of generative AI. The possibilities that arose through ChatGPT, Llama and the likes produced a host of new ideas for how to integrate them into Technology Enhanced Learning (TEL). For example, they can be used to generate individualized learner feedback, in combination with other Natural Laguage Processing (NLP) based feedback generation techniques. In addition, the area is developing rapidly. PhD candidates and others alike need to steadily advance their research by syncing up on current developments, possibilities, and ideas for NLP technology application in TEL. We want to make that possible through our proposed workshop.

Learning Objectives

By the end of this workshop, participants should:

• have a basic understanding of challenges and methods in applying NLP technology in TEL for feedback generation.
• know about the current developments in the application of NLP-based technology for feedback generation, including, but not limited to the methods that the workshop team develops themselves.
• have developed an understanding on how to generate feedback using language technology, employing a human-centered approach.”

Pre-activities

The workshop is designed for those with a basic understanding of Learning Analytics and wish to deepen their knowledge about the application of NLP-based technology in TEL. Basic knowledge of AI and its sub-fields, especially NLP, will be beneficial but not mandatory. Participants are expected to have an interest in the application of new technologies in enhancing the learning process, and a willingness to engage in discussions and group activities.

Session Description

In our workshop, we want to provide insights into current developments of the use of language technology in TEL. For this, we’ll start with a recap of the basics about feedback generation for natural language tasks and in natural language that we presented and discussed in last years workshop. After that, we want to give an overview of the methods that were developed since and what has improved over the last year. We’ll present some methods that we have tried ourselves and discuss methods that other workshop participants have tried out or are interested in experimenting with. In the end we want to discuss on how to improve natural language feedback generation in the future: What techniques are to be employed? How can we combine other methods of giving feedback with feedback generation? How can we enhance and combine process and product feedback generation, all while keeping the human centered aspect in mind? Find it out (with us) in our workshop!