Summary of Leveraging Large Language Models For Active Merchant Non-player Characters, by Byungjun Kim et al.
Leveraging Large Language Models for Active Merchant Non-player Characters
by Byungjun Kim, Minju Kim, Dayeon Seo, Bugeun Kim
First submitted to arxiv on: 15 Dec 2024
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Computation and Language (cs.CL)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A novel approach to creating more engaging and dynamic non-player characters (NPCs) in game environments is presented. The issue with current merchant NPCs is their passivity, which stems from limited pricing and communication capabilities. To overcome these limitations, a large language model-based framework called MART is proposed. This framework consists of an appraiser module and a negotiator module that can be trained using various methods such as supervised finetuning (SFT) and knowledge distillation (KD). The effectiveness of different training methods and LLM sizes was evaluated through two experiments, revealing that SFT and KD are suitable for implementing active merchant NPCs. However, the study also identified three irregular cases arising from the responses of LLMs. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary In this paper, researchers tackle a problem in game development: making non-player characters (NPCs) more interactive. Right now, NPC merchants can’t adjust prices or communicate with players very well. To fix this, they created a special framework called MART that uses big language models. This framework has two parts: one that evaluates items and another that negotiates with players. The scientists tested different ways to train these models and found that some methods work better than others. They also discovered some weird cases where the models didn’t respond as expected. Overall, this study aims to help game developers make more realistic and engaging NPCs. |
Keywords
» Artificial intelligence » Knowledge distillation » Large language model » Supervised
