Guardado en:
Detalles Bibliográficos
Autores principales: Chen, Dong, Zhang, Shuo, Zhuang, Yueting, Tang, Siliang, Liu, Qidong, Wang, Hua, Xu, Mingliang
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:https://arxiv.org/abs/2406.15471
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866914844657057792
author Chen, Dong
Zhang, Shuo
Zhuang, Yueting
Tang, Siliang
Liu, Qidong
Wang, Hua
Xu, Mingliang
author_facet Chen, Dong
Zhang, Shuo
Zhuang, Yueting
Tang, Siliang
Liu, Qidong
Wang, Hua
Xu, Mingliang
contents Pretrained large models (PLMs), such as ChatGPT, have demonstrated remarkable performance across diverse tasks. However, the significant computational requirements of PLMs have discouraged most product teams from running or fine-tuning them. In such cases, to harness the exceptional performance of PLMs, one must rely on expensive APIs, thereby exacerbating the economic burden. Despite the overall inferior performance of small models, in specific distributions, they can achieve comparable or even superior results. Consequently, some input can be processed exclusively by small models. On the other hand, certain tasks can be broken down into multiple subtasks, some of which can be completed without powerful capabilities. Under these circumstances, small models can handle the simple subtasks, allowing large models to focus on challenging subtasks, thus improving the performance. We propose Data Shunt$^+$ (DS$^+$), a general paradigm for collaboration of small and large models. DS$^+$ not only substantially reduces the cost associated with querying large models but also effectively improves large models' performance. For instance, ChatGPT achieves an accuracy of $94.43\%$ on Amazon Product sentiment analysis, and DS$^+$ achieves an accuracy of $95.64\%$, while the cost has been reduced to only $31.18\%$. Besides, experiments also prove that the proposed collaborative-based paradigm can better inject specific task knowledge into PLMs compared to fine-tuning.
format Preprint
id arxiv_https___arxiv_org_abs_2406_15471
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Improving Large Models with Small models: Lower Costs and Better Performance
Chen, Dong
Zhang, Shuo
Zhuang, Yueting
Tang, Siliang
Liu, Qidong
Wang, Hua
Xu, Mingliang
Computation and Language
Artificial Intelligence
Machine Learning
Pretrained large models (PLMs), such as ChatGPT, have demonstrated remarkable performance across diverse tasks. However, the significant computational requirements of PLMs have discouraged most product teams from running or fine-tuning them. In such cases, to harness the exceptional performance of PLMs, one must rely on expensive APIs, thereby exacerbating the economic burden. Despite the overall inferior performance of small models, in specific distributions, they can achieve comparable or even superior results. Consequently, some input can be processed exclusively by small models. On the other hand, certain tasks can be broken down into multiple subtasks, some of which can be completed without powerful capabilities. Under these circumstances, small models can handle the simple subtasks, allowing large models to focus on challenging subtasks, thus improving the performance. We propose Data Shunt$^+$ (DS$^+$), a general paradigm for collaboration of small and large models. DS$^+$ not only substantially reduces the cost associated with querying large models but also effectively improves large models' performance. For instance, ChatGPT achieves an accuracy of $94.43\%$ on Amazon Product sentiment analysis, and DS$^+$ achieves an accuracy of $95.64\%$, while the cost has been reduced to only $31.18\%$. Besides, experiments also prove that the proposed collaborative-based paradigm can better inject specific task knowledge into PLMs compared to fine-tuning.
title Improving Large Models with Small models: Lower Costs and Better Performance
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2406.15471