# Alleviating Representational Shift for Continual Fine-tuning

arXiv: https://arxiv.org/abs/2204.10535 · HTML: https://mlsanity.com/paper/2204.10535/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- Builds on 32 of its 50 references that are in the map.
- First posted on arXiv 22 Apr 2022.

## Builds on (most cited first, 30 of 32)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift](https://mlsanity.com/paper/1502.03167/) (2015)
- [Overcoming catastrophic forgetting in neural networks](https://mlsanity.com/paper/1612.00796/) (2016)
- [iCaRL: Incremental Classifier and Representation Learning](https://mlsanity.com/paper/1611.07725/) (2016)
- [Visualizing and Understanding Convolutional Networks](https://mlsanity.com/paper/1311.2901/) (2013)
- [Memory Aware Synapses: Learning what (not) to forget](https://mlsanity.com/paper/1711.09601/) (2017)
- [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)
- [Fine-Tuning can Distort Pretrained Features and Underperform Out-of-Distribution](https://mlsanity.com/paper/2202.10054/) (2022)
- [Large Scale Incremental Learning](https://mlsanity.com/paper/1905.13260/) (2019)
- [An Empirical Investigation of Catastrophic Forgetting in Gradient-Based Neural Networks](https://mlsanity.com/paper/1312.6211/) (2013)
- [Gradient based sample selection for online continual learning](https://mlsanity.com/paper/1903.08671/) (2019)
- [Online Continual Learning with Maximally Interfered Retrieval](https://mlsanity.com/paper/1908.04742/) (2019)
- [PathNet: Evolution Channels Gradient Descent in Super Neural Networks](https://mlsanity.com/paper/1701.08734/) (2017)
- [CORe50: a New Dataset and Benchmark for Continuous Object Recognition](https://mlsanity.com/paper/1705.03550/) (2017)
- [Semantic Drift Compensation for Class-Incremental Learning](https://mlsanity.com/paper/2004.00440/) (2020)
- [New Insights on Reducing Abrupt Representation Change in Online Continual Learning](https://mlsanity.com/paper/2104.05025/) (2021)
- [On the Theory of Transfer Learning: The Importance of Task Diversity](https://mlsanity.com/paper/2006.11650/) (2020)
- [FearNet: Brain-Inspired Model for Incremental Learning](https://mlsanity.com/paper/1711.10563/) (2017)
- [Continual Lifelong Learning in Natural Language Processing: A Survey](https://mlsanity.com/paper/2012.09823/) (2020)
- [Continual Prototype Evolution: Learning Online from Non-Stationary Data Streams](https://mlsanity.com/paper/2009.00919/) (2020)
- [Batch Renormalization: Towards Reducing Minibatch Dependence in Batch-Normalized Models](https://mlsanity.com/paper/1702.03275/) (2017)
- [Avalanche: an End-to-End Library for Continual Learning](https://mlsanity.com/paper/2104.00405/) (2021)
- [Probing Representation Forgetting in Supervised and Unsupervised Continual Learning](https://mlsanity.com/paper/2203.13381/) (2022)
- [AFEC: Active Forgetting of Negative Transfer in Continual Learning](https://mlsanity.com/paper/2110.12187/) (2021)
- [Sentence Embedding Alignment for Lifelong Relation Extraction](https://mlsanity.com/paper/1903.02588/) (2019)
- [CPR: Classifier-Projection Regularization for Continual Learning](https://mlsanity.com/paper/2006.07326/) (2020)
- [Rehearsal-Free Continual Learning over Small Non-I.I.D. Batches](https://mlsanity.com/paper/1907.03799/) (2019)

## Built on by (most cited first, 8 of 8)

- [Class-Incremental Learning: A Survey](https://mlsanity.com/paper/2302.03648/) (2023)
- [RanPAC: Random Projections and Pre-trained Models for Continual Learning](https://mlsanity.com/paper/2307.02251/) (2023)
- [Investigating Forgetting in Pre-Trained Representations Through Continual Learning](https://mlsanity.com/paper/2305.05968/) (2023)
- [On the Usage of Continual Learning for Out-of-Distribution Generalization in Pre-trained Language Models of Code](https://mlsanity.com/paper/2305.04106/) (2023)
- [CLR: Channel-wise Lightweight Reprogramming for Continual Learning](https://mlsanity.com/paper/2307.11386/) (2023)
- [Premonition: Using Generative Models to Preempt Future Data Changes in Continual Learning](https://mlsanity.com/paper/2403.07356/) (2024)
- [Demystifying the Base and Novel Performances for Few-shot Class-incremental Learning](https://mlsanity.com/paper/2206.10596/) (2022)
- [Evolving from Unknown to Known: Retentive Angular Representation Learning for Incremental Open Set Recognition](https://mlsanity.com/paper/2509.06570/) (2025)

## Abstract (arXiv)

We study a practical setting of continual learning: fine-tuning on a pre-trained model continually. Previous work has found that, when training on new tasks, the features (penultimate layer representations) of previous data will change, called representational shift. Besides the shift of features, we reveal that the intermediate layers' representational shift (IRS) also matters since it disrupts batch normalization, which is another crucial cause of catastrophic forgetting. Motivated by this, we propose ConFiT, a fine-tuning method incorporating two components, cross-convolution batch normalization (Xconv BN) and hierarchical fine-tuning. Xconv BN maintains pre-convolution running means instead of post-convolution, and recovers post-convolution ones before testing, which corrects the inaccurate estimates of means under IRS. Hierarchical fine-tuning leverages a multi-stage strategy to fine-tune the pre-trained network, preventing massive changes in Conv layers and thus alleviating IRS. Experimental results on four datasets show that our method remarkably outperforms several state-of-the-art methods with lower storage overhead.