SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1908.04742 · 2019

Online Continual Learning with Maximally Interfered Retrieval

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 13 functions out of this paper's own repositories and ran 7 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
optimass/Maximally_Interfered_Retrieval canonical 7 of 13
FunctionStatusWhere it lives
ResNet18 Ran optimass/Maximally_Interfered_Retrieval/model.py
code served (permissive licence) · get_code("c6b1a6798ffeb442")
conv3x3 Ran optimass/Maximally_Interfered_Retrieval/model.py
code served (permissive licence) · get_code("ec20f22a185cd708")
get_grad_vector Ran optimass/Maximally_Interfered_Retrieval/utils.py
code served (permissive licence) · get_code("947e501e9d13e0ce")
log_normal_diag Ran optimass/Maximally_Interfered_Retrieval/VAE/distributions.py
code served (permissive licence) · get_code("7d71e086676bc142")
log_normal_normalized Ran optimass/Maximally_Interfered_Retrieval/VAE/distributions.py
code served (permissive licence) · get_code("a811f7f779566e3e")
log_normal_standard Ran optimass/Maximally_Interfered_Retrieval/VAE/distributions.py
code served (permissive licence) · get_code("a39500bc34bfcdd6")
onehot Ran optimass/Maximally_Interfered_Retrieval/utils.py
code served (permissive licence) · get_code("651082fb231f9a1d")
get_cifar_buffer Not yet run optimass/Maximally_Interfered_Retrieval/buffer.py
code served (permissive licence) · get_code("7f8a097320e69838")
get_future_step_parameters Not yet run optimass/Maximally_Interfered_Retrieval/utils.py
code served (permissive licence) · get_code("a1dbf374743f5c7b")
get_permuted_mnist Not yet run optimass/Maximally_Interfered_Retrieval/data.py
code served (permissive licence) · get_code("93eb2ca19dee45bc")
get_split_cifar10 Not yet run optimass/Maximally_Interfered_Retrieval/data.py
code served (permissive licence) · get_code("de5833eda07e76b8")
get_split_mnist Not yet run optimass/Maximally_Interfered_Retrieval/data.py
code served (permissive licence) · get_code("a78b3966b0df22e0")
retrieve_hybrid Not yet run optimass/Maximally_Interfered_Retrieval/mir.py
code served (permissive licence) · get_code("952eca59dbb44579")

Repositories linked to this paper

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

Abstract

Continual learning, the setting where a learning agent is faced with a never ending stream of data, continues to be a great challenge for modern machine learning systems. In particular the online or "single-pass through the data" setting has gained attention recently as a natural setting that is difficult to tackle. Methods based on replay, either generative or from a stored memory, have been shown to be effective approaches for continual learning, matching or exceeding the state of the art in a number of standard benchmarks. These approaches typically rely on randomly selecting samples from the replay memory or from a generative model, which is suboptimal. In this work, we consider a controlled sampling of memories for replay. We retrieve the samples which are most interfered, i.e. whose prediction will be most negatively impacted by the foreseen parameters update. We show a formulation for this sampling criterion in both the generative replay and the experience replay setting, producing consistent gains in performance and greatly reduced forgetting. We release an implementation of our method at https://github.com/optimass/Maximally_Interfered_Retrieval.

For agents

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

get_harvested_code_for_paper("1908.04742")
get_code_for_paper("1908.04742")
have("1908.04742")

Connect an agent — have() is free.