本文质疑持续学习中默认使用prompt tuning的合理性，发现在Learning to Prompt和S-Prompts中改用LoRA能提升准确率且保持推理速度，强调需通过严格消融实验推动该领域进展。
摘要：
最近的持续学习（CL）方法将预训练的Transformer与prompt tuning（一种参数高效微调（PEFT）技术）相结合。我们认为，先前研究对prompt tuning的选择是一个未经论证且未经消融的决策，虽然被后续研究不加批判地采用，但仍需要进一步研究以理解其影响。在本文中，我们开展了这项研究，发现选择prompt tuning作为PEFT方法会损害CL系统的整体性能。为了说明这一点，我们在两种最先进的持续学习方法Learning to Prompt和S-Prompts中，用LoRA替换了prompt tuning。这些变体在广泛的领域增量和类增量基准上始终达到更高的准确率，同时在推理速度上具有竞争力。我们的工作强调了一个关键论点：未经检验的选择会阻碍该领域的进展，需要进行严格的消融实验（如PEFT方法），以推动CL技术在现实应用中的有意义采用。
英文摘要：
Recent Continual Learning (CL) methods have combined pretrained Transformers with prompt tuning, a parameter-efficient fine-tuning (PEFT) technique. We argue that the choice of prompt tuning in prior works was an undefended and unablated decision, which has been uncritically adopted by subsequent research, but warrants further research to understand its implications. In this paper, we conduct this research and find that the choice of prompt tuning as a PEFT method hurts the overall performance of the CL system. To illustrate this, we replace prompt tuning with LoRA in two state-of-the-art continual learning methods: Learning to Prompt and S-Prompts. These variants consistently achieve higher accuracy across a wide range of domain-incremental and class-incremental benchmarks, while being competitive in inference speed. Our work highlights a crucial argument: unexamined choices can hinder progress in the field, and rigorous ablations, such as the PEFT method, are required to drive meaningful adoption of CL techniques in real-world applications.
Parameter-Efficient Fine-Tuning for Continual Learning: A Neural Tangent Kernel Perspective
School of Electrical and Information Engineering, Tianjin Key Laboratory of Brain-Inspired Intelligence Technology, Tianjin University（电子信息工程学院，天津脑启发智能技术重点实验室，天津大学）；Shanghai Artificial Intelligence Laboratory（上海人工智能实验室）；College of Information Science and Electronic Engineering, Zhejiang University（信息科学与电子工程学院，浙江大学）；Department of Computer Science, the University of Sheffield（计算机科学系，谢菲尔德大学）；Institute of Artificial Intelligence (TeleAI) of China Telecom（中国电信人工智能研究院（TeleAI））
GEM-Style Constraints for PEFT with Dual Gradient Projection in LoRA
Department of Computer Science University of North Carolina Greensboro（计算机科学系北卡罗来纳州立大学）；Department of Computer Science North Carolina State University（计算机科学系北卡罗来纳州立大学）
Stability-Plasticity Balance via Singular-Vector Selection in LLM Continual Learning
School of Artificial Intelligence, Beihang University（北京航空航天大学人工智能学院）；Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing（未来区块链与隐私计算北京高精尖创新中心）；Institute of Computing Technology, Chinese Academy of Sciences（中国科学院计算技术研究所）
GRID: Scaling Task-Agnostic Inference in Continual Prompt Tuning
State University of New York at Buffalo（纽约州立大学布法罗分校）；Department of Computer Science and Engineering（计算机科学与工程系）；Institute for Artificial Intelligence and Data Science（人工智能与数据科学研究院）
MM-Prompt: Cross-Modal Prompt Tuning for Continual Visual Question Answering
Khoury College of Computer Sciences, Northeastern University（东北大学克劳尔计算机科学学院）；New Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所模式识别新实验室）
University of Toronto（多伦多大学）；Vector Institute（向量研究所）；Meta AI
We introduce Progressive Prompts - a simple and efficient approach for continual learning in language models. Our method allows forward transfer and resists catastrophic forgetting, without relying on data replay or a large number of task-specific parameters. Progressive Prompts learns a new soft prompt for each task and sequentially concatenates it with the previously learned prompts, while keeping the base model frozen. Experiments on standard continual learning benchmarks show that our approach outperforms state-of-the-art methods, with an improvement >20% in average test accuracy over the …
Faculty of Information Technology, University of Jyväskylä（约赫斯库利亚大学信息科技学院）；National University of Singapore（新加坡国立大学）；School of Computer Science and Technology, Dalian University of Technology（大连理工大学计算机科学与技术学院）
Georgia Institute of Technology（佐治亚理工学院）；MIT-IBM Watson AI Lab（MIT-IBM沃森人工智能实验室）；Rice University（莱斯大学）；IBM Research（IBM研究院）
Computer vision models suffer from a phenomenon known as catastrophic forgetting when learning novel concepts from continuously shifting training data. Typical solutions for this continual learning problem require extensive rehearsal of previously seen data, which increases memory costs and may violate data privacy. Recently, the emergence of large-scale pre-trained vision transformer models has enabled prompting approaches as an alternative to data-rehearsal. These approaches rely on a key-query mechanism to generate prompts and have been found to be highly resistant to catastrophic forgettin…
Dynamic Cross-Modal Prompt Generation for Multimodal Continual Instruction Tuning
School of Artificial Intelligence, Nanjing University（南京大学人工智能学院）；State Key Laboratory for Novel Software Technology, Nanjing University（南京大学新型软件技术国家重点实验室）
Personalized Federated Continual Learning via Multi-granularity Prompt
School of Computing and Artificial Intelligence, Southwestern University of Finance and Economics（西南财经大学计算机与人工智能学院）；Webank（微众银行）；College of Computer Science, Sichuan University（四川大学计算机学院）；JD Intelligent Cities Research（京东城市智能研究院）；School of Computing and Artificial Intelligence, Southwest Jiaotong University（西南交通大学计算机与人工智能学院）
Critical Patch-Aware Sparse Prompting with Decoupled Training for Continual Learning on the Edge
School of Computer Science and Engineering, Chung-Ang University（中央大学计算机科学与工程学院）；Department of Artificial Intelligence, Chung-Ang University（中央大学人工智能系）
Dynamically Anchored Prompting for Task-Imbalanced Continual Learning
Xiamen University（厦门大学）；Univ. Grenoble Alpes（格勒诺布尔大学）；Inria（法国国家信息与自动化研究所）；CNRS（法国国家科学研究中心）；Grenoble INP（格勒诺布尔理工学院）；LJK（让·昆廷诺实验室）；Guangdong Laboratory of Artificial Intelligence and Digital Economy (SZ)（广东省人工智能与数字经济实验室（深圳））
Bilevel ZOFO: Efficient LLM Fine-Tuning and Meta-Training
Department of Computer Science University of Maryland - College Park（计算机科学系马里兰大学- College Park）；Department of Computer Science and Engineering University of Texas at Arlington（计算机科学与工程系德克萨斯理工大学阿灵顿分校）
GAP-Prompt: Gated Adaptive Prompting for Efficient Continual Learning
Université d’Orléans（奥尔良大学）；INSA CVL（中央卢瓦尔河谷国立应用科学学院）；CY Cergy Paris University（塞尔吉-蓬图瓦兹大学）；ENSEA（法国高等电子与数字学院）；CNRS（法国国家科学研究中心）；Université de Technologie de Troyes（特鲁瓦技术大学）
PROL : Rehearsal Free Continual Learning in Streaming Data via Prompt Online Learning
STEM University of South Australia（南澳大利亚STEM大学）；Institute for Infocomm Research, A*STAR & IPAL（信息与通信研究所，A*STAR及IPAL）；Systems Research Institute, Polish Academy of Sciences（波兰科学院系统研究所）
Pengcheng laboratory（鹏城实验室）；Harbin Institute of Technology（哈尔滨工业大学）；Xi’an Jiaotong University（西安交通大学）；Shenzhen Institute of Advanced Technology, Chinese Academy of Sciences（中国科学院深圳先进技术研究院）
The mainstream paradigm behind continual learning has been to adapt the model parameters to non-stationary data distributions, where catastrophic forgetting is the central challenge. Typical methods rely on a rehearsal buffer or known task identity at test time to retrieve learned knowledge and address forgetting, while this work presents a new paradigm for continual learning that aims to train a more succinct memory system without accessing task identity at test time. Our method learns to dynamically prompt (L2P) a pre-trained model to learn tasks sequentially under different task transitions…
Consistent Prompting for Rehearsal-Free Continual Learning
School of Artificial Intelligence, Sun Yat-sen University（中山大学人工智能学院）；Cheng Kar-Shun Robotics Institute, The Hong Kong University of Science and Technology（香港科技大学郑家纯机器人研究院）；Key Laboratory of Machine Intelligence and Advanced Computing, Ministry of Education（教育部机器智能与先进计算重点实验室）
