# Reducing catastrophic forgetting of incremental learning in the absence of rehearsal memory with task-specific token

arXiv: https://arxiv.org/abs/2411.05846 · HTML: https://mlsanity.com/paper/2411.05846/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 12 of its 18 references that are in the map.
- First posted on arXiv 6 Nov 2024.

## Builds on (most cited first, 12 of 12)

- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [CBAM: Convolutional Block Attention Module](https://mlsanity.com/paper/1807.06521/) (2018)
- [CvT: Introducing Convolutions to Vision Transformers](https://mlsanity.com/paper/2103.15808/) (2021)
- [Riemannian Walk for Incremental Learning: Understanding Forgetting and Intransigence](https://mlsanity.com/paper/1801.10112/) (2018)
- [ConViT: Improving Vision Transformers with Soft Convolutional Inductive Biases](https://mlsanity.com/paper/2103.10697/) (2021)
- [Maintaining Discrimination and Fairness in Class Incremental Learning](https://mlsanity.com/paper/1911.07053/) (2019)
- [Rainbow Memory: Continual Learning with a Memory of Diverse Samples](https://mlsanity.com/paper/2103.17230/) (2021)
- [FOSTER: Feature Boosting and Compression for Class-Incremental Learning](https://mlsanity.com/paper/2204.04662/) (2022)
- [FeTrIL: Feature Translation for Exemplar-Free Class-Incremental Learning](https://mlsanity.com/paper/2211.13131/) (2022)
- [Resolving Task Confusion in Dynamic Expansion Architectures for Class Incremental Learning](https://mlsanity.com/paper/2212.14284/) (2022)
- [Precision Health Data: Requirements, Challenges and Existing Techniques for Data Security and Privacy](https://mlsanity.com/paper/2008.10733/) (2020)

## Abstract (arXiv)

Deep learning models generally display catastrophic forgetting when learning new data continuously. Many incremental learning approaches address this problem by reusing data from previous tasks while learning new tasks. However, the direct access to past data generates privacy and security concerns. To address these issues, we present a novel method that preserves previous knowledge without storing previous data. This method is inspired by the architecture of a vision transformer and employs a unique token capable of encapsulating the compressed knowledge of each task. This approach generates task-specific embeddings by directing attention differently based on the task associated with the data, thereby effectively mimicking the impact of having multiple models through tokens. Our method incorporates a distillation process that ensures efficient interactions even after multiple additional learning steps, thereby optimizing the model against forgetting. We measured the performance of our model in terms of accuracy and backward transfer using a benchmark dataset for different task-incremental learning scenarios. Our results demonstrate the superiority of our approach, which achieved the highest accuracy and lowest backward transfer among the compared methods. In addition to presenting a new model, our approach lays the foundation for various extensions within the spectrum of vision-transformer architectures.