Improving Neural Topic Models with Wasserstein Knowledge Distillation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Adhya, Suman, Sanyal, Debarshi Kumar
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929391378890752
author Adhya, Suman
Sanyal, Debarshi Kumar
author_facet Adhya, Suman
Sanyal, Debarshi Kumar
contents Topic modeling is a dominant method for exploring document collections on the web and in digital libraries. Recent approaches to topic modeling use pretrained contextualized language models and variational autoencoders. However, large neural topic models have a considerable memory footprint. In this paper, we propose a knowledge distillation framework to compress a contextualized topic model without loss in topic quality. In particular, the proposed distillation objective is to minimize the cross-entropy of the soft labels produced by the teacher and the student models, as well as to minimize the squared 2-Wasserstein distance between the latent distributions learned by the two models. Experiments on two publicly available datasets show that the student trained with knowledge distillation achieves topic coherence much higher than that of the original student model, and even surpasses the teacher while containing far fewer parameters than the teacher's. The distilled model also outperforms several other competitive topic models on topic coherence.
format Preprint
id arxiv_https___arxiv_org_abs_2303_15350
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Improving Neural Topic Models with Wasserstein Knowledge Distillation
Adhya, Suman
Sanyal, Debarshi Kumar
Computation and Language
Information Retrieval
Machine Learning
Topic modeling is a dominant method for exploring document collections on the web and in digital libraries. Recent approaches to topic modeling use pretrained contextualized language models and variational autoencoders. However, large neural topic models have a considerable memory footprint. In this paper, we propose a knowledge distillation framework to compress a contextualized topic model without loss in topic quality. In particular, the proposed distillation objective is to minimize the cross-entropy of the soft labels produced by the teacher and the student models, as well as to minimize the squared 2-Wasserstein distance between the latent distributions learned by the two models. Experiments on two publicly available datasets show that the student trained with knowledge distillation achieves topic coherence much higher than that of the original student model, and even surpasses the teacher while containing far fewer parameters than the teacher's. The distilled model also outperforms several other competitive topic models on topic coherence.
title Improving Neural Topic Models with Wasserstein Knowledge Distillation
topic Computation and Language
Information Retrieval
Machine Learning
url https://arxiv.org/abs/2303.15350