Differentially Private Knowledge Distillation via Synthetic Text Generation

• URL: http://arxiv.org/abs/2403.00932v2
• Date: Wed, 5 Jun 2024 03:07:44 GMT
• Title: Differentially Private Knowledge Distillation via Synthetic Text Generation
• Authors: James Flemings, Murali Annavaram,
• Abstract summary: We propose DistilDP: a novel differentially private knowledge distillation algorithm. DistilDP exploits synthetic data generated by a differentially private teacher LLM. Our experimental results demonstrate that DistilDP can substantially improve the utility over existing baselines.
• Score: 5.201318326501886
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large Language models (LLMs) are achieving state-of-the-art performance in many different downstream tasks. However, the increasing urgency of data privacy puts pressure on practitioners to train LLMs with Differential Privacy (DP) on private data. Concurrently, the exponential growth in parameter size of LLMs necessitates model compression before deployment of LLMs on resource-constrained devices or latency-sensitive applications. Differential privacy and model compression generally must trade off utility loss to achieve their objectives. Moreover, simultaneously applying both schemes can compound the utility degradation. To this end, we propose DistilDP: a novel differentially private knowledge distillation algorithm that exploits synthetic data generated by a differentially private teacher LLM. The knowledge of a teacher LLM is transferred onto the student in two ways: one way from the synthetic data itself -- the hard labels, and the other way by the output distribution of the teacher evaluated on the synthetic data -- the soft labels. Furthermore, if the teacher and student share a similar architectural structure, we can further distill knowledge by aligning the hidden representations between both. Our experimental results demonstrate that DistilDP can substantially improve the utility over existing baselines, at least $9.0$ PPL on the Big Patent dataset, with strong privacy parameters, $\epsilon=2$. These promising results progress privacy-preserving compression of autoregressive LLMs. Our code can be accessed here: https://github.com/james-flemings/dp_compress.

Related papers

• Learning with Less: Knowledge Distillation from Large Language Models via Unlabeled Data [54.934578742209716]
  In real-world NLP applications, Large Language Models (LLMs) offer promising solutions due to their extensive training on vast datasets. LLKD is an adaptive sample selection method that incorporates signals from both the teacher and student. Our comprehensive experiments show that LLKD achieves superior performance across various datasets with higher data efficiency.
  arXiv  Detail & Related papers  (2024-11-12T18:57:59Z)
• Learning Privacy-Preserving Student Networks via Discriminative-Generative Distillation [24.868697898254368]
  Deep models may pose a privacy leakage risk in practical deployment. We propose a discriminative-generative distillation approach to learn privacy-preserving deep models. Our approach can control query cost over private data and accuracy degradation in a unified manner.
  arXiv  Detail & Related papers  (2024-09-04T03:06:13Z)
• Entropy Law: The Story Behind Data Compression and LLM Performance [115.70395740286422]
  We find that model performance is negatively correlated to the compression ratio of training data, which usually yields a lower training loss. Based on the findings of the entropy law, we propose a quite efficient and universal data selection method. We also present an interesting application of entropy law that can detect potential performance risks at the beginning of model training.
  arXiv  Detail & Related papers  (2024-07-09T08:14:29Z)
• CodecLM: Aligning Language Models with Tailored Synthetic Data [51.59223474427153]
  We introduce CodecLM, a framework for adaptively generating high-quality synthetic data for instruction-following abilities. We first encode seed instructions into metadata, which are concise keywords generated on-the-fly to capture the target instruction distribution. We also introduce Self-Rubrics and Contrastive Filtering during decoding to tailor data-efficient samples.
  arXiv  Detail & Related papers  (2024-04-08T21:15:36Z)
• Differentially Private Synthetic Data via Foundation Model APIs 2: Text [56.13240830670327]
  A lot of high-quality text data generated in the real world is private and cannot be shared or used freely due to privacy concerns. We propose an augmented PE algorithm, named Aug-PE, that applies to the complex setting of text. Our results demonstrate that Aug-PE produces DP synthetic text that yields competitive utility with the SOTA DP finetuning baselines.
  arXiv  Detail & Related papers  (2024-03-04T05:57:50Z)
• Under the Surface: Tracking the Artifactuality of LLM-Generated Data [21.002983022237604]
  This work delves into the expanding role of large language models (LLMs) in generating artificial data. To the best of our knowledge, this is the first study to aggregate various types of LLM-generated text data. Despite artificial data's capability to match human performance, this paper reveals significant hidden disparities.
  arXiv  Detail & Related papers  (2024-01-26T07:53:27Z)
• Self-Play Fine-Tuning Converts Weak Language Models to Strong Language Models [52.98743860365194]
  We propose a new fine-tuning method called Self-Play fIne-tuNing (SPIN) At the heart of SPIN lies a self-play mechanism, where the LLM refines its capability by playing against instances of itself. This sheds light on the promise of self-play, enabling the achievement of human-level performance in LLMs without the need for expert opponents.
  arXiv  Detail & Related papers  (2024-01-02T18:53:13Z)
• From Quantity to Quality: Boosting LLM Performance with Self-Guided Data Selection for Instruction Tuning [52.257422715393574]
  We introduce a self-guided methodology for Large Language Models (LLMs) to autonomously discern and select cherry samples from open-source datasets. Our key innovation, the Instruction-Following Difficulty (IFD) metric, emerges as a pivotal metric to identify discrepancies between a model's expected responses and its intrinsic generation capability.
  arXiv  Detail & Related papers  (2023-08-23T09:45:29Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.