View Proposal

---

Proposer

Alessandro Suglia

Title

ActionLLM: Developing Large Language Models that Learn To Use Tools

Goal

Developing LLMs that can access external tools via API calls

Description

Current large language models like ChatGPT and Bard have remarkable abilities to generate very fluent natural language and they have achieved impressive performance on several benchmarks. However, for many relevant tasks in the real world, we have to make sure that these models can also interact with external sources (e.g., Wikipedia or search engines) as well as other external tools such as calculators, databases and so on. In this project, we will explore the field of large language models that can use external tools to achieve very specific goals and tasks.

Resources

Please find below some important papers that you can use to read on this topic: GorillaLM: https://arxiv.org/abs/2305.15334 ViperGPT: https://arxiv.org/abs/2303.08128

Background

The candidate student should be familiar with Python programming as well as Machine Learning/Deep Learning with a specific focus on Natural Language Processing

Url

Difficulty Level

Moderate

Ethical Approval

None

Number Of Students

1

Supervisor

Alessandro Suglia

Keywords

deep learning, neural networks, language models

Degrees

Master of Science in Artificial Intelligence
Master of Science in Data Science
Master of Science in Robotics