Examining Changes in Internal Representations of Continual Learning Models Through Tensor Decomposition

Nishant Suresh Aswani, Amira Guesmi, Muhammad Abdullah Hanif, Muhammad Shafique
Proceedings of the 1st ContinualAI Unconference, 2023, PMLR 249:62-82, 2024.

Abstract

Continual learning (CL) has spurred the development of several methods aimed at consolidating previous knowledge across sequential learning. Yet, the evaluations of these methods have primarily focused on the final output, such as changes in the accuracy of predicted classes, overlooking the issue of representational forgetting within the model. In this paper, we propose a novel representation-based evaluation framework for CL models. This approach involves gathering internal representations from throughout the continual learning process and formulating three-dimensional tensors. The tensors are formed by stacking representations, such as layer activations, generated from several inputs and model ‘snapshots’, throughout the learning process. By conducting tensor component analysis (TCA), we aim to uncover meaningful patterns about how the internal representations evolve, expecting to highlight the merits or shortcomings of examined CL strategies. We conduct our analyses across different model architectures and importance-based continual learning strategies, with a curated task selection. Often, the results of our approach mirror the difference in performance of various CL strategies on various architectures. Ultimately, however, we found that our methodology did not directly highlight specialized clusters of neurons, nor provide an immediate understanding the evolution of filters. We believe a scaled down variation of our approach will provide insight into the benefits and pitfalls of using TCA to study continual learning dynamics.

Cite this Paper

BibTeX
@InProceedings{pmlr-v249-aswani24a, title = {Examining Changes in Internal Representations of Continual Learning Models Through Tensor Decomposition}, author = {Aswani, Nishant Suresh and Guesmi, Amira and Hanif, Muhammad Abdullah and Shafique, Muhammad}, booktitle = {Proceedings of the 1st ContinualAI Unconference, 2023}, pages = {62--82}, year = {2024}, editor = {Swaroop, Siddharth and Mundt, Martin and Aljundi, Rahaf and Khan, Mohammad Emtiyaz}, volume = {249}, series = {Proceedings of Machine Learning Research}, month = {09 Oct}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v249/main/assets/aswani24a/aswani24a.pdf}, url = {https://proceedings.mlr.press/v249/aswani24a.html}, abstract = {Continual learning (CL) has spurred the development of several methods aimed at consolidating previous knowledge across sequential learning. Yet, the evaluations of these methods have primarily focused on the final output, such as changes in the accuracy of predicted classes, overlooking the issue of representational forgetting within the model. In this paper, we propose a novel representation-based evaluation framework for CL models. This approach involves gathering internal representations from throughout the continual learning process and formulating three-dimensional tensors. The tensors are formed by stacking representations, such as layer activations, generated from several inputs and model ‘snapshots’, throughout the learning process. By conducting tensor component analysis (TCA), we aim to uncover meaningful patterns about how the internal representations evolve, expecting to highlight the merits or shortcomings of examined CL strategies. We conduct our analyses across different model architectures and importance-based continual learning strategies, with a curated task selection. Often, the results of our approach mirror the difference in performance of various CL strategies on various architectures. Ultimately, however, we found that our methodology did not directly highlight specialized clusters of neurons, nor provide an immediate understanding the evolution of filters. We believe a scaled down variation of our approach will provide insight into the benefits and pitfalls of using TCA to study continual learning dynamics.} }
Endnote
%0 Conference Paper %T Examining Changes in Internal Representations of Continual Learning Models Through Tensor Decomposition %A Nishant Suresh Aswani %A Amira Guesmi %A Muhammad Abdullah Hanif %A Muhammad Shafique %B Proceedings of the 1st ContinualAI Unconference, 2023 %C Proceedings of Machine Learning Research %D 2024 %E Siddharth Swaroop %E Martin Mundt %E Rahaf Aljundi %E Mohammad Emtiyaz Khan %F pmlr-v249-aswani24a %I PMLR %P 62--82 %U https://proceedings.mlr.press/v249/aswani24a.html %V 249 %X Continual learning (CL) has spurred the development of several methods aimed at consolidating previous knowledge across sequential learning. Yet, the evaluations of these methods have primarily focused on the final output, such as changes in the accuracy of predicted classes, overlooking the issue of representational forgetting within the model. In this paper, we propose a novel representation-based evaluation framework for CL models. This approach involves gathering internal representations from throughout the continual learning process and formulating three-dimensional tensors. The tensors are formed by stacking representations, such as layer activations, generated from several inputs and model ‘snapshots’, throughout the learning process. By conducting tensor component analysis (TCA), we aim to uncover meaningful patterns about how the internal representations evolve, expecting to highlight the merits or shortcomings of examined CL strategies. We conduct our analyses across different model architectures and importance-based continual learning strategies, with a curated task selection. Often, the results of our approach mirror the difference in performance of various CL strategies on various architectures. Ultimately, however, we found that our methodology did not directly highlight specialized clusters of neurons, nor provide an immediate understanding the evolution of filters. We believe a scaled down variation of our approach will provide insight into the benefits and pitfalls of using TCA to study continual learning dynamics.
APA
Aswani, N.S., Guesmi, A., Hanif, M.A. & Shafique, M.. (2024). Examining Changes in Internal Representations of Continual Learning Models Through Tensor Decomposition. Proceedings of the 1st ContinualAI Unconference, 2023, in Proceedings of Machine Learning Research 249:62-82 Available from https://proceedings.mlr.press/v249/aswani24a.html.

Related Material