ORACLE: Order Robust Adaptive Continual LEarning

9 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.

Continual Learning with Adaptive Weights (CLAW) (2019, methodology): “Other recent algorithms include the ORACLE algorithm by Yoon et al. (2019), which addresses the sensitivity of a continual learner to the order of tasks it encounters by establishing an order robust learner that represents the parameters of each task as a sum of task-shared and task-specific…”
HyperGrid Transformers: Towards A Single Model for Multiple Tasks (2021, background): “Learning task-adaptive parameters to avoid catastrophic forgetting has also been a go-to strategy for continual learning (Yoon et al., 2019).”
HyperGrid: Efficient Multi-Task Transformers with Grid-wise Decomposable Hyper Projections (2020, background): “Learning task-adaptive parameters to avoid catastrophic forgetting has also been a go-to strategy for continual learning [Yoon et al., 2019].”
Variational Federated Multi-Task Learning (2019, background): “New approaches focused on other components, like architecture innovations, introducing lateral connections that allow new models to reuse knowledge from previously trained models with layer-wise adaptors [36, 38], and memory enhanced models with generative networks [45, 47].”