mlsanityData to 14 Oct 2026

DPZero: Private Fine-Tuning of Language Models without Backpropagation

arXiv 2310.09639 ↗First posted 14 Oct 2023cs.LG cs.CR math.OCcited by 26

Not on the map. Why?

Abstract (arXiv)

The widespread practice of fine-tuning large language models (LLMs) on domain-specific data faces two major challenges in memory and privacy. First, as the size of LLMs continues to grow, the memory demands of gradient-based training methods via backpropagation become prohibitively high. Second, given the tendency of LLMs to memorize training data, it is important to protect potentially sensitive information in the fine-tuning data from being regurgitated. Zeroth-order methods, which rely solely on forward passes, substantially reduce memory consumption during training. However, directly combining them with standard differentially private gradient descent suffers more as model size grows. To bridge this gap, we introduce DPZero, a novel private zeroth-order algorithm with nearly dimension-independent rates. The memory efficiency of DPZero is demonstrated in privately fine-tuning RoBERTa and OPT on several downstream tasks. Our code is available at https://github.com/Liang137/DPZero.

Key facts. Cited by 26 papers in the map (data through 14 Oct 2026). Builds on 94 of its 153 references that are in the map. First posted on arXiv 14 Oct 2023.

Builds on

94 of 153 references.

References · 94

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  2. 2020Language Models are Few-Shot Learnerscited by 32,419
  3. 2022Training language models to follow instructions with human feedbackcited by 15,971
  4. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  5. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,372
  6. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  7. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,899
  8. 2019DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lightercited by 4,307
  9. 2016SQuAD: 100,000+ Questions for Machine Comprehension of Textcited by 4,148
  10. 2021Prefix-Tuning: Optimizing Continuous Prompts for Generationcited by 3,915
  11. 2022OPT: Open Pre-trained Transformer Language Modelscited by 3,512
  12. 2016Deep Learning with Differential Privacycited by 2,728
38 more
  1. 2017A Broad-Coverage Challenge Corpus for Sentence Understanding through Inferencecited by 2,682
  2. 2016Membership Inference Attacks against Machine Learning Modelscited by 2,281
  3. 2019BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questionscited by 2,183
  4. 2015A large annotated corpus for learning natural language inferencecited by 1,847
  5. 2019ZeRO: Memory Optimizations Toward Training Trillion Parameter Modelscited by 1,301
  6. 2020Making Pre-trained Language Models Better Few-shot Learnerscited by 1,256
  7. 2018Adafactor: Adaptive Learning Rates with Sublinear Memory Costcited by 905
  8. 2019DROP: A Reading Comprehension Benchmark Requiring Discrete Reasoning Over Paragraphscited by 903
  9. 2013Stochastic First- and Zeroth-order Methods for Nonconvex Stochastic Programmingcited by 724
  10. 2020Intrinsic Dimensionality Explains the Effectiveness of Language Model Fine-Tuningcited by 671
  11. 2017ZOO: Zeroth Order Optimization based Black-box Attacks to Deep Neural Networks without Training Substitute Modelscited by 666
  12. 2017Evolution Strategies as a Scalable Alternative to Reinforcement Learningcited by 574
  13. 2016Linear Convergence of Gradient and Proximal-Gradient Methods Under the Polyak-Lojasiewicz Conditioncited by 539
  14. 2023Symbolic Discovery of Optimization Algorithmscited by 434
  15. 2021Large Language Models Can Be Strong Differentially Private Learnerscited by 412
  16. 2021Differentially Private Fine-tuning of Language Modelscited by 400
  17. 2018Measuring the Intrinsic Dimension of Objective Landscapescited by 370
  18. 2023Fine-Tuning Language Models with Just Forward Passescited by 345
  19. 2013The Composition Theorem for Differential Privacycited by 297
  20. 2019An Investigation into Neural Net Optimization via Hessian Eigenvalue Densitycited by 292
  21. 2017Empirical Analysis of the Hessian of Over-Parametrized Neural Networkscited by 290
  22. 2018SPIDER: Near-Optimal Non-Convex Optimization via Stochastic Path Integrated Differential Estimatorcited by 287
  23. 2023Analyzing Leakage of Personally Identifiable Information in Language Modelscited by 285
  24. 2022The Forward-Forward Algorithm: Some Preliminary Investigationscited by 266
  25. 2023Sophia: A Scalable Stochastic Second-order Optimizer for Language Model Pre-trainingcited by 265
  26. 2023Membership Inference Attacks against Language Models via Neighbourhood Comparisoncited by 258
  27. 2018Gradient Descent Happens in a Tiny Subspacecited by 217
  28. 2013Optimal rates for zero-order convex optimization: the power of two function evaluationscited by 198
  29. 2019Private Stochastic Convex Optimization with Optimal Ratescited by 186
  30. 2020Stability of Stochastic Gradient Descent on Nonsmooth Convex Lossescited by 173
  31. 2020Private Stochastic Convex Optimization: Optimal Rates in Linear Timecited by 153
  32. 2020Understanding Gradient Clipping in Private SGD: A Geometric Perspectivecited by 142
  33. 2018Differentially Private Empirical Risk Minimization Revisited: Faster and More Generalcited by 128
  34. 2015An Optimal Algorithm for Bandit and Zero-Order Convex Optimization with Two-Point Feedbackcited by 111
  35. 2018Zeroth-Order Stochastic Variance Reduction for Nonconvex Optimizationcited by 111
  36. 2023Privacy-Preserving In-Context Learning with Differentially Private Few-Shot Generationcited by 96
  37. 2020Bypassing the Ambient Dimension: Private SGD with Gradient Subspace Identificationcited by 90
  38. 2021Private Stochastic Convex Optimization: Optimal Rates in ℓ1 Geometrycited by 83

50 most cited of 94 All 94 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 7; 2025: 11; 2026: 8.

26 papers.

Sort:

By area:

  1. 2024Private Fine-tuning of Large Language Models with Zeroth-order Optimizationcited by 41
  2. 2024On the Crucial Role of Initialization for Matrix Factorizationcited by 15
  3. 2024Differentially Private Zeroth-Order Methods for Scalable Large Language Model Finetuningcited by 14
  4. 2025POPri: Private Federated Learning using Preference-Optimized Synthetic Datacited by 14
  5. 2024Divide, Reweight, and Conquer: A Logit Arithmetic Approach for In-Context Learningcited by 9
  6. 2024Fine-Tuning and Deploying Large Language Models Over Edges: Issues and Approachescited by 6
  7. 2025Zeroth-Order Optimization Finds Flat Minimacited by 6
  8. 2025Bilevel ZOFO: Efficient LLM Fine-Tuning and Meta-Trainingcited by 5
  9. 2024Characterizing the Training Dynamics of Private Fine-tuning with Langevin diffusioncited by 4
  10. 2026Low-Rank Adaptation Redux for Large Modelscited by 3
  11. 2025Memory-Efficient Differentially Private Training with Gradient Random Projectioncited by 2
  12. 2025Private Zeroth-Order Optimization with Public Datacited by 2
14 more
  1. 2026Test-Time Detoxification without Training or Learning Anythingcited by 2
  2. 2024Bridging Today and the Future of Humanity: AI Safety in 2024 and Beyondcited by 1
  3. 2025Communication-Efficient and Differentially Private Vertical Federated Learning with Zeroth-Order Optimizationcited by 1
  4. 2025Cocoon: A System Architecture for Differentially Private Training with Correlated Noisescited by 1
  5. 2026Position: Zeroth-Order Optimization in Deep Learning Is Underexplored, Not Underpoweredcited by 1
  6. 2025Privacy Amplification in Differentially Private Zeroth-Order Optimization with Hidden Statescited by 0
  7. 2025On the Inherent Privacy of Zeroth Order Projected Gradient Descentcited by 0
  8. 2025Zeroth-Order Sharpness-Aware Learning with Exponential Tiltingcited by 0
  9. 2025ConMeZO: Adaptive Descent-Direction Sampling for Gradient-Free Finetuning of Large Language Modelscited by 0
  10. 2026Zeroth-Order Optimization at the Edge of Stabilitycited by 0
  11. 2026Self-Improving In-Context Learningcited by 0
  12. 2026Inference-Native Zeroth-Order Optimization for LLMscited by 0
  13. 2026When Privacy Hurts Mergeability: Geometry-Aware Model Merging under Differential Privacycited by 0
  14. 2026SPADE-DFL: Communication-Efficient Decentralized Federated Learning via Derivative-Free Linearized ADMMcited by 0