From Commands to Prompts: LLM-based Semantic File System for AIOS
Zeru Shi, Kai Mei, Mingyu Jin, Yongye Su, Chaoji Zuo, Wenyue Hua, Wujiang Xu, Yujie Ren, Zirui Liu, Mengnan Du, Dong Deng, Yongfeng Zhang
2024-10-17
Summary
This paper introduces a new system called LSFS (LLM-based Semantic File System) that allows users to manage files using natural language prompts instead of traditional commands.
What's the problem?
Traditional file systems require users to navigate complex folders and remember specific commands, which can be confusing and inconvenient. This makes it hard for people to find and manage their files effectively, especially if they have many documents or if the file names are not easy to remember.
What's the solution?
To solve this problem, the authors developed LSFS, which uses large language models (LLMs) to let users interact with their files using simple, everyday language. Instead of typing commands like 'open' or 'delete,' users can say things like 'show me my notes on biology.' LSFS creates a semantic index for files, allowing it to understand the meaning behind user requests and perform actions like retrieving files, summarizing content, or rolling back to previous versions. The system also includes a set of APIs that support various file operations and improves overall efficiency.
Why it matters?
This research is important because it makes file management much easier and more intuitive for users. By allowing natural language interactions, LSFS can help people work more efficiently and reduce frustration when handling files. This advancement could be particularly valuable in environments where quick access to information is crucial, such as in schools, offices, or during research.
Abstract
Large language models (LLMs) have demonstrated significant potential in the development of intelligent applications and systems such as LLM-based agents and agent operating systems (AIOS). However, when these applications and systems interact with the underlying file system, the file system still remains the traditional paradigm: reliant on manual navigation through precise commands. This paradigm poses a bottleneck to the usability of these systems as users are required to navigate complex folder hierarchies and remember cryptic file names. To address this limitation, we propose an LLM-based semantic file system ( LSFS ) for prompt-driven file management. Unlike conventional approaches, LSFS incorporates LLMs to enable users or agents to interact with files through natural language prompts, facilitating semantic file management. At the macro-level, we develop a comprehensive API set to achieve semantic file management functionalities, such as semantic file retrieval, file update monitoring and summarization, and semantic file rollback). At the micro-level, we store files by constructing semantic indexes for them, design and implement syscalls of different semantic operations (e.g., CRUD, group by, join) powered by vector database. Our experiments show that LSFS offers significant improvements over traditional file systems in terms of user convenience, the diversity of supported functions, and the accuracy and efficiency of file operations. Additionally, with the integration of LLM, our system enables more intelligent file management tasks, such as content summarization and version comparison, further enhancing its capabilities.