黑箱教师模型可能泄露测试数据，导致知识蒸馏失效。
Training on the Test Model: Contamination in Ranking Distillation
- 用模拟污染数据测试蒸馏效果，发现即使少量污染也影响结果
- 实验证明：测试集仅占训练集极小比例时仍存在显著污染
- 提醒使用者警惕无法溯源的商业大模型蒸馏风险
基于预训练语言模型的神经排序方法在即兴搜索中表现优异，但计算开销大。为此常采用知识蒸馏，让小型高效模型学习大型昂贵模型的能力。典型场景是将基于API的商业大模型（如LLM）的知识迁移到生产级小模型。然而，由于多数商业模型的训练数据和流程不透明，无法确认所选测试集合是否曾被教师模型见过，存在无意的数据泄露风险。本文研究了教师模型污染对蒸馏的影响，通过构建‘最坏情况’模拟环境评估多种蒸馏技术。结果表明，即便测试数据仅占教师训练样本的一小部分，污染仍会显著发生。因此建议在使用数据来源不明的黑箱教师模型时保持谨慎。
原文摘要 · Abstract (English)
Neural approaches to ranking based on pre-trained language models are highly effective in ad-hoc search. However, the computational expense of these models can limit their application. As such, a process known as knowledge distillation is frequently applied to allow a smaller, efficient model to learn from an effective but expensive model. A key example of this is the distillation of expensive API-based commercial Large Language Models into smaller production-ready models. However, due to the opacity of training data and processes of most commercial models, one cannot ensure that a chosen test collection has not been observed previously, creating the potential for inadvertent data contamination. We, therefore, investigate the effect of a contaminated teacher model in a distillation setting. We evaluate several distillation techniques to assess the degree to which contamination occurs during distillation. By simulating a ``worst-case'' setting where the degree of contamination is known, we find that contamination occurs even when the test data represents a small fraction of the teacher's training samples. We, therefore, encourage caution when training using black-box teacher models where data provenance is ambiguous.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。