Biological underpinnings for lifelong learning machines

By D. Kudithipudi, Mario Aguilar-Simon, Jonathan Babb, Maxim Bazhenov, Douglas Blackiston, Josh C. Bongard, Andrew P. Brna, Suraj Chakravarthi Raja, Nick Cheney, Jeff Clune and 33 more. 2022

294 citations recorded by the source corpus.

Subfields: Cellular and Molecular Neuroscience, Cognitive Neuroscience, Electrical and Electronic Engineering, Meta & Continual Learning

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Cellular and Molecular Neuroscience

1. Liyuan Wang
2. Jun Zhu
3. Soheil Kolouri
4. Hang Su
5. Bahram Zonooz
6. Jeff Clune
7. Xingxing Zhang
8. Prashant Bhat

In Cognitive Neuroscience

1. Prashant Bhat
2. Bahram Zonooz
3. Elahe Arani
4. Liyuan Wang
5. Jun Zhu
6. Soheil Kolouri
7. Hang Su
8. Hamed Pirsiavash

In Electrical and Electronic Engineering

1. Liyuan Wang
2. Jun Zhu
3. Hang Su
4. Bahram Zonooz
5. Hamed Pirsiavash
6. Jeff Clune
7. Xingxing Zhang
8. Bowen Baker

In Meta & Continual Learning

1. Xingxing Zhang
2. Liyuan Wang
3. Hang Su
4. Jun Zhu
5. Prashant Bhat
6. Bahram Zonooz
7. Elahe Arani
8. Soheil Kolouri

Citations in the corpus

• A Comprehensive Survey of Continual Learning: Theory, Method and Application (2023, background): “In contrast, the latest surveystypicallycollateonlyalocalaspectofcontinuallearning, withrespecttoitsbiologicalunderpinnings[1],[3],[9],[10], specializedsettingsforvisualclassification[11],[12],[13],[14], andspecificextensionsforNLP[15],[16]orRL[17].”
• Incorporating neuro-inspired adaptability for continual learning in artificial intelligence (2023, neutral): “However, the benefits of active forgetting for continual learning remain to be explored 3 .”
• Task-Aware Information Routing from Common Representation Space in Lifelong Learning (2023, background): “Evolving Architectures: In addition to experience rehearsal, CL in the brain is mediated by selfregulated neurogenesis that scale up the number of new memories that can be encoded and stored without catastrophic forgetting (Kudithipudi et al., 2022).”
• Hebbian Learning based Orthogonal Projection for Continual Learning of Spiking Neural Networks (2024, methodology): “To deal with this problem, biological underpinnings are investigated (Kudithipudi et al., 2022), and various machine learning methods are proposed.”
• Video PreTraining (VPT): Learning to Act by Watching Unlabeled Online Videos (2022, background): “Agents are optimized with the phasic policy gradient 65 A major problem when fine-tuning with RL is catastrophic forgetting 66,67 because previously learned skills can be lost before their value is realized.”
• Three types of incremental learning (2022, unknown): “During training on context K , the regularization term of FROMP was given by: with τ a hyperparameter controlling the regularization strength (which was set based on a grid search, see Supplementary Note 8) and ̂ θ ∗ the parameter vector with values as they were at the end of training on context K −”
• Is Multi-Task Learning an Upper Bound for Continual Learning? (2022, background): “Others have studied brain-inspired mechanisms [14] that allow for continual learning in mammalian brains and how to leverage them for continual machine learning.”