SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1803.10123 · 2018

Task Agnostic Continual Learning Using Online Variational Bayes

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 6 functions out of this paper's own repositories and ran 5 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
igolan/bgd canonical 5 of 6
FunctionStatusWhere it lives
adam Ran igolan/bgd/optimizers_lib/optimizers_lib.py
code served (permissive licence) · get_code("e4de4e90f9014c20")
calculate_gain Ran igolan/bgd/nn_utils/init_utils.py
code served (permissive licence) · get_code("35b9edc57d88d973")
kaiming_normal_std_ Ran igolan/bgd/nn_utils/init_utils.py
code served (permissive licence) · get_code("7017609a5ade3c7f")
labels_trick Ran igolan/bgd/nn_utils/labels_trick.py
code served (permissive licence) · get_code("2a7a7a7286986fde")
sgd Ran igolan/bgd/optimizers_lib/optimizers_lib.py
code served (permissive licence) · get_code("184f03d97f32de50")
bgd Not yet run igolan/bgd/optimizers_lib/optimizers_lib.py
code served (permissive licence) · get_code("996f532d849dc656")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Catastrophic forgetting is the notorious vulnerability of neural networks to the change of the data distribution while learning. This phenomenon has long been considered a major obstacle for allowing the use of learning agents in realistic continual learning settings. A large body of continual learning research assumes that task boundaries are known during training. However, research for scenarios in which task boundaries are unknown during training has been lacking. In this paper we present, for the first time, a method for preventing catastrophic forgetting (BGD) for scenarios with task boundaries that are unknown during training --- task-agnostic continual learning. Code of our algorithm is available at https://github.com/igolan/bgd.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("1803.10123")
get_code_for_paper("1803.10123")
have("1803.10123")

Connect an agent — have() is free.