# Overcoming the Stability Gap in Continual Learning

arXiv: https://arxiv.org/abs/2306.01904 · HTML: https://mlsanity.com/paper/2306.01904/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- Builds on 76 of its 108 references that are in the map.
- First posted on arXiv 2 Jun 2023.

## Builds on (most cited first, 30 of 76)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [ImageNet Large Scale Visual Recognition Challenge](https://mlsanity.com/paper/1409.0575/) (2014)
- [Overcoming catastrophic forgetting in neural networks](https://mlsanity.com/paper/1612.00796/) (2016)
- [Zero-Shot Text-to-Image Generation](https://mlsanity.com/paper/2102.12092/) (2021)
- [Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification](https://mlsanity.com/paper/1502.01852/) (2015)
- [A ConvNet for the 2020s](https://mlsanity.com/paper/2201.03545/) (2022)
- [Learning without Forgetting](https://mlsanity.com/paper/1606.09282/) (2016)
- [iCaRL: Incremental Classifier and Representation Learning](https://mlsanity.com/paper/1611.07725/) (2016)
- [Gradient Episodic Memory for Continual Learning](https://mlsanity.com/paper/1706.08840/) (2017)
- [Progressive Neural Networks](https://mlsanity.com/paper/1606.04671/) (2016)
- [Continual Lifelong Learning with Neural Networks: A Review](https://mlsanity.com/paper/1802.07569/) (2018)
- [Memory Aware Synapses: Learning what (not) to forget](https://mlsanity.com/paper/1711.09601/) (2017)
- [Efficient Lifelong Learning with A-GEM](https://mlsanity.com/paper/1812.00420/) (2018)
- [Learning to Prompt for Continual Learning](https://mlsanity.com/paper/2112.08654/) (2021)
- [Dark Experience for General Continual Learning: a Strong, Simple Baseline](https://mlsanity.com/paper/2004.07211/) (2020)
- [Riemannian Walk for Incremental Learning: Understanding Forgetting and Intransigence](https://mlsanity.com/paper/1801.10112/) (2018)
- [Large Scale Incremental Learning](https://mlsanity.com/paper/1905.13260/) (2019)
- [ConvNeXt V2: Co-designing and Scaling ConvNets with Masked Autoencoders](https://mlsanity.com/paper/2301.00808/) (2023)
- [Large-Scale Long-Tailed Recognition in an Open World](https://mlsanity.com/paper/1904.05160/) (2019)
- [Lifelong Learning with Dynamically Expandable Networks](https://mlsanity.com/paper/1708.01547/) (2017)
- [Overcoming catastrophic forgetting with hard attention to the task](https://mlsanity.com/paper/1801.01423/) (2018)
- [MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer](https://mlsanity.com/paper/2110.02178/) (2021)
- [DualPrompt: Complementary Prompting for Rehearsal-free Continual Learning](https://mlsanity.com/paper/2204.04799/) (2022)
- [Green AI](https://mlsanity.com/paper/1907.10597/) (2019)
- [End-to-End Incremental Learning](https://mlsanity.com/paper/1807.09536/) (2018)
- [Carbon Emissions and Large Neural Network Training](https://mlsanity.com/paper/2104.10350/) (2021)
- [Learning under Concept Drift: A Review](https://mlsanity.com/paper/2004.05785/) (2020)
- [CODA-Prompt: COntinual Decomposed Attention-based Prompting for Rehearsal-Free Continual Learning](https://mlsanity.com/paper/2211.13218/) (2022)

## Built on by (most cited first, 13 of 13)

- [CLAP4CLIP: Continual Learning with Probabilistic Finetuning for Vision-Language Models](https://mlsanity.com/paper/2403.19137/) (2024)
- [Efficient Continual Pre-training by Mitigating the Stability Gap](https://mlsanity.com/paper/2406.14833/) (2024)
- [A Practitioner's Guide to Continual Multimodal Pretraining](https://mlsanity.com/paper/2408.14471/) (2024)
- [GRASP: A Rehearsal Policy for Efficient Online Continual Learning](https://mlsanity.com/paper/2308.13646/) (2023)
- [What Variables Affect Out-of-Distribution Generalization in Pretrained Models?](https://mlsanity.com/paper/2405.15018/) (2024)
- [Controlling Neural Collapse Enhances Out-of-Distribution Detection and Transfer Learning](https://mlsanity.com/paper/2502.10691/) (2025)
- [Improving Multimodal Large Language Models Using Continual Learning](https://mlsanity.com/paper/2410.19925/) (2024)
- [Exploring the Stability Gap in Continual Learning: The Role of the Classification Head](https://mlsanity.com/paper/2411.04723/) (2024)
- [The Expanding Scope of the Stability Gap: Unveiling its Presence in Joint Incremental Learning of Homogeneous Tasks](https://mlsanity.com/paper/2406.05114/) (2024)
- [A Good Start Matters: Enhancing Continual Learning with Data-Driven Weight Initialization](https://mlsanity.com/paper/2503.06385/) (2025)
- [On Understanding of the Dynamics of Model Capacity in Continual Learning](https://mlsanity.com/paper/2508.08052/) (2025)
- [Dynamic gain neuromodulation attenuates the stability gap under joint training](https://mlsanity.com/paper/2507.14056/) (2025)
- [Loss Smoothing for Stable Adaptation Under Distribution Shift](https://mlsanity.com/paper/2607.00634/) (2026)

## Abstract (arXiv)

Pre-trained deep neural networks (DNNs) are being widely deployed by industry for making business decisions and to serve users; however, a major problem is model decay, where the DNN's predictions become more erroneous over time, resulting in revenue loss or unhappy users. To mitigate model decay, DNNs are retrained from scratch using old and new data. This is computationally expensive, so retraining happens only once performance significantly decreases. Here, we study how continual learning (CL) could potentially overcome model decay in large pre-trained DNNs and greatly reduce computational costs for keeping DNNs up-to-date. We identify the "stability gap" as a major obstacle in our setting. The stability gap refers to a phenomenon where learning new data causes large drops in performance for past tasks before CL mitigation methods eventually compensate for this drop. We test two hypotheses to investigate the factors influencing the stability gap and identify a method that vastly reduces this gap. In large-scale experiments for both easy and hard CL distributions (e.g., class incremental learning), we demonstrate that our method reduces the stability gap and greatly increases computational efficiency. Our work aligns CL with the goals of the production setting, where CL is needed for many applications.