Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Investigating the Potential of Using Large Language Models for Scheduling

Deddy Jobson, Yilin Li

TL;DR

This study reveals that LLMs, even under zero-shot settings, create reasonably good first drafts of conference schedules, and when clustering papers, using only titles as LLM inputs produces results closer to human categorization than using titles and abstracts with TFIDF.

Abstract

The inaugural ACM International Conference on AI-powered Software introduced the AIware Challenge, prompting researchers to explore AI-driven tools for optimizing conference programs through constrained optimization. We investigate the use of Large Language Models (LLMs) for program scheduling, focusing on zero-shot learning and integer programming to measure paper similarity. Our study reveals that LLMs, even under zero-shot settings, create reasonably good first drafts of conference schedules. When clustering papers, using only titles as LLM inputs produces results closer to human categorization than using titles and abstracts with TFIDF. The code has been made publicly available.

Investigating the Potential of Using Large Language Models for Scheduling

TL;DR

This study reveals that LLMs, even under zero-shot settings, create reasonably good first drafts of conference schedules, and when clustering papers, using only titles as LLM inputs produces results closer to human categorization than using titles and abstracts with TFIDF.

Abstract

The inaugural ACM International Conference on AI-powered Software introduced the AIware Challenge, prompting researchers to explore AI-driven tools for optimizing conference programs through constrained optimization. We investigate the use of Large Language Models (LLMs) for program scheduling, focusing on zero-shot learning and integer programming to measure paper similarity. Our study reveals that LLMs, even under zero-shot settings, create reasonably good first drafts of conference schedules. When clustering papers, using only titles as LLM inputs produces results closer to human categorization than using titles and abstracts with TFIDF. The code has been made publicly available.
Paper Structure (14 sections, 8 equations, 1 figure, 2 tables)

This paper contains 14 sections, 8 equations, 1 figure, 2 tables.

Table of Contents

  1. Introduction
  2. Problem Statement
  3. Experiments
  4. Zero-Shot Scheduling by LLMs
  5. Integer Program using LLM to measure similarity
  6. Conclusion
  7. Acknowledgements
  8. Prompt Design
  9. Integer Programming Formulation
  10. Objective Function
  11. Constraints
  12. Link $z$ and $x$ variables
  13. Each paper scheduled exactly once
  14. Session length constraints

Figures (1)

  • Figure 1: Comparing the homogeneity and completeness score for different number of papers per session. Homogeneity deteriorates when the scale of the data is larger, but not completeness.