Prompt Conditioned VAE: Enhancing Generative Replay for Lifelong Learning in Task-Oriented Dialogue
Not on the map. Why?
Abstract (arXiv)
Lifelong learning (LL) is vital for advanced task-oriented dialogue (ToD) systems. To address the catastrophic forgetting issue of LL, generative replay methods are widely employed to consolidate past knowledge with generated pseudo samples. However, most existing generative replay methods use only a single task-specific token to control their models. This scheme is usually not strong enough to constrain the generative model due to insufficient information involved. In this paper, we propose a novel method, prompt conditioned VAE for lifelong learning (PCLL), to enhance generative replay by incorporating tasks' statistics. PCLL captures task-specific distributions with a conditional variational autoencoder, conditioned on natural language prompts to guide the pseudo-sample generation. Moreover, it leverages a distillation process to further consolidate past knowledge by alleviating the noise in pseudo samples. Experiments on natural language understanding tasks of ToD systems demonstrate that PCLL significantly outperforms competitive baselines in building LL models.
Key facts. Cited by 8 papers in the map (data through 7 Oct 2026). Builds on 72 of its 94 references that are in the map. First posted on arXiv 14 Oct 2022.
Builds on
72 of 94 references.
References · 72
- 2017Attention Is All You Needcited by 59,932
- 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
- 2020Language Models are Few-Shot Learnerscited by 32,199
- 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
- 2015Distilling the Knowledge in a Neural Networkcited by 10,076
- 2019Parameter-Efficient Transfer Learning for NLPcited by 4,195
- 2013Auto-Encoding Variational Bayescited by 3,980
- 2019The Curious Case of Neural Text Degenerationcited by 2,881
- 2016Learning without Forgettingcited by 2,465
- 2016iCaRL: Incremental Classifier and Representation Learningcited by 2,313
- 2021Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Language Processingcited by 2,222
- 2019Language Models as Knowledge Bases?cited by 1,922
38 more
- 2017Gradient Episodic Memory for Continual Learningcited by 1,871
- 2016Progressive Neural Networkscited by 1,332
- 2018Continual Lifelong Learning with Neural Networks: A Reviewcited by 1,313
- 2020Making Pre-trained Language Models Better Few-shot Learnerscited by 1,255
- 2017Continual Learning with Deep Generative Replaycited by 1,152
- 2017Memory Aware Synapses: Learning what (not) to forgetcited by 1,094
- 2018Experience Replay for Continual Learningcited by 1,074
- 2015A Diversity-Promoting Objective Function for Neural Conversation Modelscited by 1,068
- 2018Efficient Lifelong Learning with A-GEMcited by 1,044
- 2019How Can We Know What Language Models Know?cited by 953
- 2021Learning to Prompt for Continual Learningcited by 938
- 2018Riemannian Walk for Incremental Learning: Understanding Forgetting and Intransigencecited by 734
- 2018Overcoming catastrophic forgetting with hard attention to the taskcited by 672
- 2020It's Not Just Size That Matters: Small Language Models Are Also Few-Shot Learnerscited by 668
- 2015Generating Sentences from a Continuous Spacecited by 607
- 2018Progress & Compress: A scalable framework for continual learningcited by 516
- 2019Towards Scalable Multi-domain Conversational Agents: The Schema-Guided Dialogue Datasetcited by 426
- 2020Efficient Intent Detection with Dual Sentence Encoderscited by 419
- 2019An Evaluation Dataset for Intent Classification and Out-of-Scope Predictioncited by 378
- 2018Snips Voice Platform: an embedded Spoken Language Understanding system for private-by-design voice interfacescited by 376
- 2016Expert Gate: Lifelong Learning with a Network of Expertscited by 369
- 2017PathNet: Evolution Channels Gradient Descent in Super Neural Networkscited by 362
- 2018Learning without Memorizingcited by 249
- 2021PPT: Pre-trained Prompt Tuning for Few-shot Learningcited by 232
- 2017Learning Discourse-level Diversity for Neural Dialog Models using Conditional Variational Autoencoderscited by 222
- 2019Cyclical Annealing Schedule: A Simple Approach to Mitigating KL Vanishingcited by 191
- 2017FearNet: Brain-Inspired Model for Incremental Learningcited by 177
- 2019Benchmarking Natural Language Understanding Services for building Conversational Agentscited by 166
- 2019LAMOL: LAnguage MOdeling for Lifelong Language Learningcited by 156
- 2020Optimus: Organizing Sentences via Pre-trained Modeling of a Latent Spacecited by 114
- 2021LFPT5: A Unified Framework for Lifelong Few-shot Language Learning Based on Prompt Tuning of T5cited by 103
- 2017Encoder Based Lifelong Learningcited by 102
- 2021GALAXY: A Generative Pre-trained Model for Task-Oriented Dialog with Semi-Supervised Learning and Explicit Policy Injectioncited by 102
- 2018Semantic Parsing for Task Oriented Dialog using Hierarchical Representationscited by 101
- 2019Uncertainty-guided Continual Learning with Bayesian Neural Networkscited by 100
- 2021Achieving Forgetting Prevention and Knowledge Transfer in Continual Learningcited by 91
- 2020A Wholistic View of Continual Learning with Deep Neural Networks: Forgotten Lessons and the Bridge to Active and Open World Learningcited by 86
- 2020Continual Learning in Task-Oriented Dialogue Systemscited by 82
50 most cited of 72 All 72 (JSON)
Built on by
Citing papers per year: 2022: 2; 2023: 1; 2024: 2; 2025: 2; 2026: 1.
8 papers.
- 2022Continual Learning of Natural Language Processing Tasks: A Surveycited by 83
- 2024Towards Lifelong Learning of Large Language Models: A Surveycited by 60
- 2024Recent Advances of Foundation Language Models-based Continual Learning: A Surveycited by 46
- 2023Continual Learning with Dirichlet Generative-based Rehearsalcited by 7
- 2025LSEBMCL: A Latent Space Energy-Based Model for Continual Learningcited by 2
- 2026MedCL-Bench: Benchmarking stability-efficiency trade-offs and scaling in biomedical continual learningcited by 2
- 2022Semi-Supervised Lifelong Language Learningcited by 1
- 2025Enhancing Memory Recall in LLMs with Gauss-Tin: A Hybrid Instructional and Gaussian Replay Approachcited by 0
- 2026MedCL-Bench: Benchmarking stability-efficiency trade-offs and scaling in biomedical continual learning17 Mar 2026
- 2025Enhancing Memory Recall in LLMs with Gauss-Tin: A Hybrid Instructional and Gaussian Replay Approach13 Aug 2025
- 2025LSEBMCL: A Latent Space Energy-Based Model for Continual Learning9 Jan 2025
- 2024Towards Lifelong Learning of Large Language Models: A Survey10 Jun 2024
- 2024Recent Advances of Foundation Language Models-based Continual Learning: A Survey28 May 2024
- 2023Continual Learning with Dirichlet Generative-based Rehearsal13 Sep 2023
- 2022Semi-Supervised Lifelong Language Learning23 Nov 2022
- 2022Continual Learning of Natural Language Processing Tasks: A Survey23 Nov 2022