Papers › ProGen: Progressive Zero-shot Dataset Generation via In-context Feedback

ProGen: Progressive Zero-shot Dataset Generation via In-context Feedback

22 Oct 2022arXiv:2210.12329archive 2025-07-28

Jiacheng Ye, Jiahui Gao, Jiangtao Feng, Zhiyong Wu, Tao Yu, Lingpeng Kong

Recently, dataset-generation-based zero-shot learning has shown promising results by training a task-specific model with a dataset synthesized from large pre-trained language models (PLMs). The final task-specific model often achieves compatible or even better performance than PLMs under the zero-shot setting, with orders of magnitude fewer parameters. However, synthetic datasets have their drawbacks. They have long been suffering from low-quality issues (e.g., low informativeness and redundancy). This explains why the massive synthetic data does not lead to better performance -- a scenario we would expect in the human-labeled data. To improve the quality of dataset synthesis, we propose a progressive zero-shot dataset generation framework, ProGen, which leverages the feedback from the task-specific model to guide the generation of new training data via in-context examples. Extensive experiments on five text classification datasets demonstrate the effectiveness of the proposed approach. We also show ProGen achieves on-par or superior performance with only 1\% synthetic dataset size compared to baseline methods without in-context feedback.

PaperPDFCode

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

Code

hkunlp/progen officialmentioned in papermentioned on GitHubpytorch report
hkunlp/symgen mentioned on GitHubApache-2.0 report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

Not run by Syntology. Nothing on this page verifies that the listed code works.

Tasks

Data-free Knowledge DistillationDataset GenerationInformativenessText ClassificationZero-Shot Learningtext-classification

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Data-free Knowledge Distillation QNLI ProGen (T5-base) Accuracy 85.9 #3 of 4 Archive leaderboard report
Data-free Knowledge Distillation SQuAD ProGen (T5-base) Exact Match 68.1 #4 of 4 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections