Authors: (1) Yugen Sato, Meiji University; (2) Taisei Nakajima, Meiji University; (3) Tatsuki Kawamoto, Meiji University; (4) Tomohiro Takagi, Meiji University. Table of Links Abstract & Introduction Related works Method Experiment Conclusion & References Abstract Large-scale language models (LLMs), such as ChatGPT, are becoming increasingly sophisticated and exhibit human-like capabilities, playing an essential role in assisting humans in a variety of everyday tasks. An important application of AI is interactive recommendation systems that respond to human inquiries and make recommendations tailored to the user. In most conventional interactive recommendation systems, the language model is used only as a dialogue model, and there is a separate recommendation system. This is due to the fact that the language model used as a dialogue system does not have the capability to serve as a recommendation system. Therefore, we will realize the construction of a dialogue system with recommendation capability by using OpenAI’s Chat-GPT, which has a very high inference capability as a dialogue system and the ability to generate high-quality sentences and verify the effectiveness of the system. Conversational recommendation, Dialogue management, Large language models Keywords 1 Introduction Large-scale language models (LLMs) such as ChatGPT [3] have shown remarkable performance in a variety of natural language processing (NLP) tasks. By leveraging large-scale pre-training on large text corpora and reinforcement learning from human feedback (RLHF), LLMs not only have extensive knowledge, but also perform well in language understanding, generation, interaction, and inference. GPT-4 [5] even exceeds human performance. Prompt engineering techniques (chain-of-thought [6] and in-context learning) allow LLMs to unlock their unlimited potential to perform the complex tasks of everyday life, making LLMs the subject of significant attention from both academia and industry. ChatGPT is a successful example of such an application, where the AI model is equipped with the ability to analyze context and respond to user queries based on knowledge gained from vast amounts of training data. It has the ability to respond to user queries. In traditional interactive recommendation systems, the language model is used only as a dialogue model in many cases, and the recommendation system exists separately from it. The system provided recommendations to the user only when the dialogue model and the recommendation system interacted with each other. The reason for this is that the dialogue system does not have the capability to act as a recommendation system, and OpenAI’s Chat-GPT, with its very high inference capability and ability to generate high quality sentences, can make recommendations in a dialogue. Therefore, we will experiment with OpenAI’s Chat-GPT as a dialogue system with recommendation capability. In a typical interactive recommendation system, the user asks the system a question such as "I am looking for an item like . and the system responds with some kind of recommendation. However, since our goal is to extract the user’s needs, we assume a situation where the system asks the user questions to analyze the user’s preferences and thoughts. We will test the effectiveness and ability of Chat-GPT to dynamically manage the flow of dialogue in such a situation. This paper is under CC 4.0 license. available on arxiv

Extracting User Needs With Chat-GPT for Dialogue Recommendation

Too Long; Didn't Read

Writings, Papers and Blogs on Text Models

About Author

TOPICS

THIS ARTICLE WAS FEATURED IN...