Generating Role-Playing Game Quests With GPT Language Models

Vartinen, Susanna; Hämäläinen, Perttu; Guckelsberger, Christian

doi:10.1109/tg.2022.3228480

Cited by 30 publications

(8 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Previous studies analyzed the relationship between LLMs and games. Värtinen et al [10] explored the use of GPT language models (GPT-2 and GPT-3) to procedurally generate quests for role-playing games (RPGs), as an alternative to hand-authored quests. Overall, the study highlights the promise of AI in generating game content but also points to the need for further improvements in the GPT language models.…”

Section: Large Language Models and Gamesmentioning

confidence: 99%

How ChatGPT can inspire and improve serious board game design

Gatti Junior,

Marasco,

Kim

et al. 2023

IJSG

View full text Add to dashboard Cite

Designing engaging serious board games that effectively address students' diverse and complex needs presents a significant challenge for educators. As a possible solution, Large Language Models (LLMs) such as ChatGPT can assist educators in designing and evaluating game-based learning experiences. This study explores three primary ways ChatGPT can enhance educators' game design process. Firstly, ChatGPT can assist with brainstorming, suggesting game themes and mechanisms aligned with curriculum and learning goals. Secondly, it can provide templates or exemplars of game components, allowing educators to create customized games that offer what their students need. Lastly, ChatGPT can offer valuable feedback on game prototypes, identifying areas for improvement and guidance to enhance the game's efficacy as an educational tool. We attempted to advance the ongoing discourse on the roles of artificial intelligence and board games in education by providing valuable insights into the potential of these tools.

show abstract

Section: Large Language Models and Gamesmentioning

confidence: 99%

How ChatGPT can inspire and improve serious board game design

Gatti Junior,

Marasco,

Kim

et al. 2023

IJSG

View full text Add to dashboard Cite

show abstract

“…It was trained on a collection of web data and, thus, outputs text for general purposes. Previous work has fine-tuned GPT for multiple domains and tasks, such as the task of quest generation in games (Värtinen et al, 2022) or the medical domain (Schneider et al, 2021). In addition to domain adaption, GPT was tailored to specific text styles and characteristics.…”

Section: Related Workmentioning

confidence: 99%

Language Models for German Text Simplification: Overcoming Parallel Data Scarcity through Style-specific Pre-training

Anschütz,

Oehms,

Wimmer

et al. 2023

Findings of the Association for Computational Linguistics: ACL 2023

View full text Add to dashboard Cite

Automatic text simplification systems help to reduce textual information barriers on the internet. However, for languages other than English, only few parallel data to train these systems exists. We propose a two-step approach to overcome this data scarcity issue. First, we fine-tuned language models on a corpus of German Easy Language, a specific style of German. Then, we used these models as decoders in a sequence-to-sequence simplification task. We show that the language models adapt to the style characteristics of Easy Language and output more accessible texts. Moreover, with the style-specific pre-training, we reduced the number of trainable parameters in text simplification models. Hence, less parallel data is sufficient for training. Our results indicate that pre-training on unaligned data can reduce the required parallel data while improving the performance on downstream tasks.

show abstract

“…The inputs are arrangements of 1024 sequential tokens. The bigger show was prepared on 256 cloud TPU v3 centers [23].…”

Section: Gpt-2mentioning

confidence: 99%

MARGEN: Marathi Question Answering Generative Conversation Model

Bhalshankar¹,

Deshmukh²

2023

Advances in Computer Science Research

View full text Add to dashboard Cite

The conversational system aka chatbot market capture was worth USD 526 million in 2021 around the world. The innovations created such as Machine learning, Deep learning, Natural Language Processing (NLP), and Big data analytics have given a modern speed-quickening fuel to Artificial Intelligence. Well, a generative chatbot could be an exceptionally effective and shrewd conversational system as distant as its learning component is concerned. They can be trained from scratch like a newborn child by using Deep Learning techniques. GPT-1,2,3 is well known for English Language NLP tasks whereas IndicBART is also moving ahead for 11 Indian Languages for the NLU tasks. With the DL techniques like RNN, LSTM, and SGD, we conducted an in-depth survey of recent deep learning conversational models available on open-source portals, examining over 25 deep learning models for conversational systems before proposing our model. "MARGEN" is a proposed model which means the user can get the system's reply with proper generative answers for the query in Marathi text and/or audio speech formats.

show abstract

Generating Role-Playing Game Quests With GPT Language Models

Cited by 30 publications

References 31 publications

How ChatGPT can inspire and improve serious board game design

How ChatGPT can inspire and improve serious board game design

Language Models for German Text Simplification: Overcoming Parallel Data Scarcity through Style-specific Pre-training

MARGEN: Marathi Question Answering Generative Conversation Model

Contact Info

Product

Resources

About