Prompts as Auto-Optimized Training Hyperparameters: Training Best-in-Class IR Models from Scratch with 10 Gold Labels

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xian, Jasper, Samuel, Saron, Khoubsirat, Faraz, Pradeep, Ronak, Sultan, Md Arafat, Florian, Radu, Roukos, Salim, Sil, Avirup, Potts, Christopher, Khattab, Omar
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929388204851200
author Xian, Jasper
Samuel, Saron
Khoubsirat, Faraz
Pradeep, Ronak
Sultan, Md Arafat
Florian, Radu
Roukos, Salim
Sil, Avirup
Potts, Christopher
Khattab, Omar
author_facet Xian, Jasper
Samuel, Saron
Khoubsirat, Faraz
Pradeep, Ronak
Sultan, Md Arafat
Florian, Radu
Roukos, Salim
Sil, Avirup
Potts, Christopher
Khattab, Omar
contents We develop a method for training small-scale (under 100M parameter) neural information retrieval models with as few as 10 gold relevance labels. The method depends on generating synthetic queries for documents using a language model (LM), and the key step is that we automatically optimize the LM prompt that is used to generate these queries based on training quality. In experiments with the BIRCO benchmark, we find that models trained with our method outperform RankZephyr and are competitive with RankLLama, both of which are 7B parameter models trained on over 100K labels. These findings point to the power of automatic prompt optimization for synthetic dataset generation.
format Preprint
id arxiv_https___arxiv_org_abs_2406_11706
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Prompts as Auto-Optimized Training Hyperparameters: Training Best-in-Class IR Models from Scratch with 10 Gold Labels
Xian, Jasper
Samuel, Saron
Khoubsirat, Faraz
Pradeep, Ronak
Sultan, Md Arafat
Florian, Radu
Roukos, Salim
Sil, Avirup
Potts, Christopher
Khattab, Omar
Information Retrieval
Computation and Language
Machine Learning
We develop a method for training small-scale (under 100M parameter) neural information retrieval models with as few as 10 gold relevance labels. The method depends on generating synthetic queries for documents using a language model (LM), and the key step is that we automatically optimize the LM prompt that is used to generate these queries based on training quality. In experiments with the BIRCO benchmark, we find that models trained with our method outperform RankZephyr and are competitive with RankLLama, both of which are 7B parameter models trained on over 100K labels. These findings point to the power of automatic prompt optimization for synthetic dataset generation.
title Prompts as Auto-Optimized Training Hyperparameters: Training Best-in-Class IR Models from Scratch with 10 Gold Labels
topic Information Retrieval
Computation and Language
Machine Learning
url https://arxiv.org/abs/2406.11706