CorrSynth - A Correlated Sampling Method for Diverse Dataset Generation from LLMs

Suhas Kowshik, Abhishek Divekar, Vijit Malik


Abstract
Large language models (LLMs) have demonstrated remarkable performance in diverse tasks using zero-shot and few-shot prompting. Even though their capabilities of data synthesis have been studied well in recent years, the generated data suffers from a lack of diversity, less adherence to the prompt, and potential biases that creep into the data from the generator model. In this work, we tackle the challenge of generating datasets with high diversity, upon which a student model is trained for downstream tasks. Taking the route of decoding-time guidance-based approaches, we propose CorrSynth, which generates data that is more diverse and faithful to the input prompt using a correlated sampling strategy. Further, our method overcomes the complexity drawbacks of some other guidance-based techniques like classifier-based guidance. With extensive experiments, we show the effectiveness of our approach and substantiate our claims. In particular, we perform intrinsic evaluation to show the improvements in diversity. Our experiments show that CorrSynth improves both student metrics and intrinsic metrics upon competitive baselines across four datasets, showing the innate advantage of our method.
Anthology ID:
2024.emnlp-main.899
Volume:
Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing
Month:
November
Year:
2024
Address:
Miami, Florida, USA
Editors:
Yaser Al-Onaizan, Mohit Bansal, Yun-Nung Chen
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
16076–16095
Language:
URL:
https://aclanthology.org/2024.emnlp-main.899
DOI:
Bibkey:
Cite (ACL):
Suhas Kowshik, Abhishek Divekar, and Vijit Malik. 2024. CorrSynth - A Correlated Sampling Method for Diverse Dataset Generation from LLMs. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 16076–16095, Miami, Florida, USA. Association for Computational Linguistics.
Cite (Informal):
CorrSynth - A Correlated Sampling Method for Diverse Dataset Generation from LLMs (Kowshik et al., EMNLP 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.emnlp-main.899.pdf