Saved in:
Bibliographic Details
Main Authors: Jain, Eeshaan, Nandy, Tushar, Aggarwal, Gaurav, Tendulkar, Ashish, Iyer, Rishabh, De, Abir
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2409.12255
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909319672365056
author Jain, Eeshaan
Nandy, Tushar
Aggarwal, Gaurav
Tendulkar, Ashish
Iyer, Rishabh
De, Abir
author_facet Jain, Eeshaan
Nandy, Tushar
Aggarwal, Gaurav
Tendulkar, Ashish
Iyer, Rishabh
De, Abir
contents Existing subset selection methods for efficient learning predominantly employ discrete combinatorial and model-specific approaches which lack generalizability. For an unseen architecture, one cannot use the subset chosen for a different model. To tackle this problem, we propose $\texttt{SubSelNet}$, a trainable subset selection framework, that generalizes across architectures. Here, we first introduce an attention-based neural gadget that leverages the graph structure of architectures and acts as a surrogate to trained deep neural networks for quick model prediction. Then, we use these predictions to build subset samplers. This naturally provides us two variants of $\texttt{SubSelNet}$. The first variant is transductive (called as Transductive-$\texttt{SubSelNet}$) which computes the subset separately for each model by solving a small optimization problem. Such an optimization is still super fast, thanks to the replacement of explicit model training by the model approximator. The second variant is inductive (called as Inductive-$\texttt{SubSelNet}$) which computes the subset using a trained subset selector, without any optimization. Our experiments show that our model outperforms several methods across several real datasets
format Preprint
id arxiv_https___arxiv_org_abs_2409_12255
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Efficient Data Subset Selection to Generalize Training Across Models: Transductive and Inductive Networks
Jain, Eeshaan
Nandy, Tushar
Aggarwal, Gaurav
Tendulkar, Ashish
Iyer, Rishabh
De, Abir
Machine Learning
Existing subset selection methods for efficient learning predominantly employ discrete combinatorial and model-specific approaches which lack generalizability. For an unseen architecture, one cannot use the subset chosen for a different model. To tackle this problem, we propose $\texttt{SubSelNet}$, a trainable subset selection framework, that generalizes across architectures. Here, we first introduce an attention-based neural gadget that leverages the graph structure of architectures and acts as a surrogate to trained deep neural networks for quick model prediction. Then, we use these predictions to build subset samplers. This naturally provides us two variants of $\texttt{SubSelNet}$. The first variant is transductive (called as Transductive-$\texttt{SubSelNet}$) which computes the subset separately for each model by solving a small optimization problem. Such an optimization is still super fast, thanks to the replacement of explicit model training by the model approximator. The second variant is inductive (called as Inductive-$\texttt{SubSelNet}$) which computes the subset using a trained subset selector, without any optimization. Our experiments show that our model outperforms several methods across several real datasets
title Efficient Data Subset Selection to Generalize Training Across Models: Transductive and Inductive Networks
topic Machine Learning
url https://arxiv.org/abs/2409.12255