Social Learning: Towards Collaborative Learning with Large Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Mohtashami, Amirkeivan, Hartmann, Florian, Gooding, Sian, Zilka, Lukas, Sharifi, Matt, Arcas, Blaise Aguera y
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916118108569600
author Mohtashami, Amirkeivan
Hartmann, Florian
Gooding, Sian
Zilka, Lukas
Sharifi, Matt
Arcas, Blaise Aguera y
author_facet Mohtashami, Amirkeivan
Hartmann, Florian
Gooding, Sian
Zilka, Lukas
Sharifi, Matt
Arcas, Blaise Aguera y
contents We introduce the framework of "social learning" in the context of large language models (LLMs), whereby models share knowledge with each other in a privacy-aware manner using natural language. We present and evaluate two approaches for knowledge transfer between LLMs. In the first scenario, we allow the model to generate abstract prompts aiming to teach the task. In our second approach, models transfer knowledge by generating synthetic examples. We evaluate these methods across diverse datasets and quantify memorization as a proxy for privacy loss. These techniques inspired by social learning yield promising results with low memorization of the original data. In particular, we show that performance using these methods is comparable to results with the use of original labels and prompts. Our work demonstrates the viability of social learning for LLMs, establishes baseline approaches and highlights several unexplored areas for future work.
format Preprint
id arxiv_https___arxiv_org_abs_2312_11441
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Social Learning: Towards Collaborative Learning with Large Language Models
Mohtashami, Amirkeivan
Hartmann, Florian
Gooding, Sian
Zilka, Lukas
Sharifi, Matt
Arcas, Blaise Aguera y
Machine Learning
Computation and Language
We introduce the framework of "social learning" in the context of large language models (LLMs), whereby models share knowledge with each other in a privacy-aware manner using natural language. We present and evaluate two approaches for knowledge transfer between LLMs. In the first scenario, we allow the model to generate abstract prompts aiming to teach the task. In our second approach, models transfer knowledge by generating synthetic examples. We evaluate these methods across diverse datasets and quantify memorization as a proxy for privacy loss. These techniques inspired by social learning yield promising results with low memorization of the original data. In particular, we show that performance using these methods is comparable to results with the use of original labels and prompts. Our work demonstrates the viability of social learning for LLMs, establishes baseline approaches and highlights several unexplored areas for future work.
title Social Learning: Towards Collaborative Learning with Large Language Models
topic Machine Learning
Computation and Language
url https://arxiv.org/abs/2312.11441