Sharing Less is More: Lifelong Learning in Deep Networks with Selective Layer Transfer

By Seungwon Lee, Sima Behpour, Eric Eaton. 2020

21 citations recorded by the source corpus.

How researchers see this paper

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

In Meta & Continual Learning

Citations in the corpus

Learning with Recoverable Forgetting (2022, positive): “As has been discussed in the previous work [26] that upper layers are preferable for transfer in life-long learning setting, T ( n − ) 0 is fully transferred to the target network.”
Architecture Matters in Continual Learning (2022, methodology): “There have been some eﬀorts on applying architecture search to continual learning (Xu and Zhu, 2018; Huang et al., 2019; Gao et al., 2020; Li et al., 2019; Lee et al., 2021).”
Recent Advances of Continual Learning in Computer Vision: An Overview (2021, positive): “A few approaches [60, 61] have been proposed to only transfer relevant knowledge.”
Towards Continual Reinforcement Learning: A Review and Perspectives (2020, neutral): “…and Kaelbling, 2018; Cases, Rosenbaum, Riemer, Geiger, Klinger, Tamkin, Li, Agarwal, Greene, Jurafsky, et al., 2019; Yang, Xu, Wu, and Wang, 2020; Lee, Behpour, and Eaton, 2021; Tseng, Lin, Feng, and Sun, 2021; Mendez and Eaton, 2022; Mendez, van Seijen, and Eaton, 2022; Gaya, Doan, Caccia,…”
How to Reuse and Compose Knowledge for a Lifetime of Tasks: A Survey on Continual Learning and Functional Composition (2022, background): “, 2020) or via expectation maximization (Lee et al., 2021a).”
Continual Object Detection via Prototypical Task Correlation Guided Gating Mechanism (2022, unknown): “Confronted with a relatively large gap, sharing too much previous knowledge would limit the performance gain in subsequent tasks [26] and more exclusive channels should be activated for new tasks.”