mlsanityData to 7 Oct 2026

Privacy-Preserving Domain Adaptation of Semantic Parsers

arXiv 2212.10520 ↗First posted 20 Dec 2022cs.CLcited by 12

Not on the map. Why?

Abstract (arXiv)

Task-oriented dialogue systems often assist users with personal or confidential matters. For this reason, the developers of such a system are generally prohibited from observing actual usage. So how can they know where the system is failing and needs more training data or new functionality? In this work, we study ways in which realistic user utterances can be generated synthetically, to help increase the linguistic and functional coverage of the system, without compromising the privacy of actual users. To this end, we propose a two-stage Differentially Private (DP) generation method which first generates latent semantic parses, and then generates utterances based on the parses. Our proposed approach improves MAUVE by 2.5× and parse tree function type overlap by 1.3× relative to current approaches for private synthetic data generation, improving both on fluency and semantic coverage. We further validate our approach on a realistic domain adaptation task of adding new functionality from private user data to a semantic parser, and show overall gains of 8.5% points in accuracy with the new feature.

Key facts. Cited by 12 papers in the map (data through 7 Oct 2026). Builds on 41 of its 58 references that are in the map. First posted on arXiv 20 Dec 2022.

Builds on

41 of 58 references.

References · 41

  1. 2016Deep Learning with Differential Privacycited by 2,682
  2. 2016Membership Inference Attacks against Machine Learning Modelscited by 2,221
  3. 2020Extracting Training Data from Large Language Modelscited by 2,144
  4. 2018The Secret Sharer: Evaluating and Testing Unintended Memorization in Neural Networkscited by 847
  5. 2017Renyi Differential Privacycited by 643
  6. 2021MAUVE: Measuring the Gap Between Neural Text and Human Text using Divergence Frontierscited by 466
  7. 2019Does Learning Require Memorization? A Short Tale about a Long Tailcited by 452
  8. 2021Large Language Models Can Be Strong Differentially Private Learnerscited by 405
  9. 2021Differentially Private Fine-tuning of Language Modelscited by 392
  10. 2013The Composition Theorem for Differential Privacycited by 290
  11. 2022Memorization Without Overfitting: Analyzing the Training Dynamics of Large Language Modelscited by 218
  12. 2022What Does it Mean for a Language Model to Preserve Privacy?cited by 194
29 more
  1. 2022Quantifying Privacy Risks of Masked Language Models Using Membership Inference Attackscited by 168
  2. 2020Task-Oriented Dialogue as Dataflow Synthesiscited by 122
  3. 2021Large-Scale Differentially Private BERTcited by 117
  4. 2022Synthetic Text Generation with Differential Privacy: A Simple and Practical Recipecited by 108
  5. 2018Semantic Parsing for Task Oriented Dialog using Hierarchical Representationscited by 101
  6. 2020On the Generalization Benefit of Noise in Stochastic Gradient Descentcited by 81
  7. 2020Grounded Adaptation for Zero-shot Executable Semantic Parsingcited by 76
  8. 2022Differentially Private Language Models for Secure Data Sharingcited by 68
  9. 2021Selective Differential Privacy for Language Modelingcited by 61
  10. 2020Enabling Fast Differentially Private SGD via Just-in-Time Compilation and Vectorizationcited by 58
  11. 2020Differentially Private Language Models Benefit from Public Pre-trainingcited by 46
  12. 2019Model-based Interactive Semantic Parsing: A Unified Framework and A Text-to-SQL Case Studycited by 40
  13. 2020Conversational Semantic Parsing for Dialog State Trackingcited by 40
  14. 2021Transferable Dialogue Systems and User Simulatorscited by 35
  15. 2022Submix: Practical Private Prediction for Large-Scale Language Modelscited by 31
  16. 2020An Imitation Game for Learning Semantic Parsers from User Interactioncited by 23
  17. 2020Learning Adaptive Language Interfaces through Decompositioncited by 22
  18. 2019Controlled Text Generation for Data Augmentation in Intelligent Artificial Agentscited by 20
  19. 2020Unsupervised Dual Paraphrasing for Two-stage Semantic Parsingcited by 20
  20. 2017Cross-domain Semantic Parsing via Paraphrasingcited by 19
  21. 2021NeuralWOZ: Learning to Collect Task-Oriented Dialogue via Model-Based Simulationcited by 16
  22. 2022Data Augmentation with Paraphrase Generation and Entity Extraction for Multimodal Dialogue Systemcited by 16
  23. 2021On The Ingredients of an Effective Zero-shot Semantic Parsercited by 13
  24. 2022Provably Confidential Language Modellingcited by 13
  25. 2019Data Augmentation with Atomic Templates for Spoken Language Understandingcited by 10
  26. 2022A Double-Graph Based Framework for Frame Semantic Parsingcited by 10
  27. 2021Differentially Private n-gram Extractioncited by 8
  28. 2022Addressing Resource and Privacy Constraints in Semantic Parsing Through Data Augmentationcited by 7
  29. 2021A Single Example Can Improve Zero-Shot Data Generationcited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 5; 2024: 4; 2025: 2; 2026: 1.

12 papers.

  1. 2023Privacy-Preserving In-Context Learning with Differentially Private Few-Shot Generationcited by 94
  2. 2024Trust No Bot: Discovering Personal Disclosures in Human-LLM Conversations in the Wildcited by 43
  3. 2024CopyBench: Measuring Literal and Non-Literal Reproduction of Copyright-Protected Text in Language Model Generationcited by 35
  4. 2025ALFA: Aligning LLMs to Ask Good Questions A Case Study in Clinical Reasoningcited by 24
  5. 2023Differentially Private Natural Language Models: Recent Advances and Future Directionscited by 20
  6. 2025A False Sense of Privacy: Evaluating Textual Data Sanitization Beyond Surface-level Privacy Leakagecited by 15
  7. 2023Synthetic Query Generation for Privacy-Preserving Deep Retrieval Systems using Differentially Private Language Modelscited by 10
  8. 2024Data-adaptive Differentially Private Prompt Synthesis for In-Context Learningcited by 10
  9. 2023Utility Theory of Synthetic Data Generationcited by 8
  10. 2024Differentially Private Learning Needs Better Model Initialization and Self-Distillationcited by 3
  11. 2023LatticeGen: A Cooperative Framework which Hides Generated Text in a Lattice for Privacy-Aware Generation on Cloudcited by 2
  12. 2026Differentially Private Preference Data Synthesis for Large Language Model Alignmentcited by 1