Exploiting Asymmetry for Synthetic Training Data Generation: SynthIE and the Case of Information Extraction
Submission Type: Regular Long Paper
Submission Track: Information Extraction
Submission Track 2: Information Extraction
Keywords: LLM, Synthetic Data Generation, Information Extraction, Closed Information Extraction, Parsing, Structured Output
TL;DR: Using cIE as a representative task, we show that exploiting asymmetries in synthetic data creation is a highly effective and versatile method for generating high-quality data, even for tasks beyond the direct problem-solving capabilities of the LLM.
Abstract: Large language models (LLMs) have great potential for synthetic data generation. This work shows that useful data can be synthetically generated even for tasks that cannot be solved directly by LLMs: for problems with structured outputs, it is possible to prompt an LLM to perform the task in the reverse direction, by generating plausible input text for a target output structure. Leveraging this asymmetry in task difficulty makes it possible to produce large-scale, high-quality data for complex tasks. We demonstrate the effectiveness of this approach on closed information extraction, where collecting ground-truth data is challenging, and no satisfactory dataset exists to date. We synthetically generate a dataset of 1.8M data points, establish its superior quality compared to existing datasets in a human evaluation, and use it to finetune small models (220M and 770M parameters), termed SynthIE, that outperform the prior state of the art (with equal model size) by a substantial margin of 57 absolute points in micro-F1 and 79 points in macro-F1. Code, data, and models are available at anonymous.
Submission Number: 4142
Loading