SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1905.06922 · 2019

On Variational Bounds of Mutual Information

Ben Poole, Alexander Alemi, George Tucker, Sherjil Ozair, Aäron Van Den Oord

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 21 functions out of this paper's own repositories and ran 17 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
RSMI-NE/RSMI-NE pwc_unofficial 12 of 15
karlstratos/doe — 4 of 5
Linear95/CLUB — 1 of 1
FunctionStatusWhere it lives
CLUB Ran Linear95/CLUB/mi_estimators.py
pointer only (licence: NONE) · get_code("5d0591b944022c57")
Interpolated Ran karlstratos/doe/util.py
code served (permissive licence) · get_code("d12ee8538feed408")
LnConstant Ran karlstratos/doe/util.py
code served (permissive licence) · get_code("0025679c7875a0c7")
LnStandardNormal Ran karlstratos/doe/util.py
code served (permissive licence) · get_code("2e3c4ed92aa3a733")
RSMI_estimate Ran RSMI-NE/RSMI-NE/rsmine/coarsegrainer/cg_optimisers.py
code served (permissive licence) · get_code("6e2da11b3259438f")
RSMIdat_filename Ran RSMI-NE/RSMI-NE/rsmine/coarsegrainer/build_dataset.py
code served (permissive licence) · get_code("437b404161244763")
array2tensor Ran RSMI-NE/RSMI-NE/rsmine/coarsegrainer/cg_utils.py
code served (permissive licence) · get_code("4183e7ff2acbf08b")
connect_bbox Ran RSMI-NE/RSMI-NE/rsmine/coarsegrainer/plotter.py
code served (permissive licence) · get_code("aaece4f3948487e3")
construct_reference_graph Ran RSMI-NE/RSMI-NE/rsmine/coarsegrainer/cg_utils.py
code served (permissive licence) · get_code("addf3c7443f250de")
eids_from_edges Ran RSMI-NE/RSMI-NE/rsmine/coarsegrainer/analysis_utils.py
code served (permissive licence) · get_code("76a7c4299deb0214")
filename Ran RSMI-NE/RSMI-NE/rsmine/coarsegrainer/build_dataset.py
code served (permissive licence) · get_code("1bf1efc3800ca8ce")
get_ln_score_function Ran karlstratos/doe/util.py
code served (permissive licence) · get_code("f13e85f7d9e6310c")
iter_loadtxt Ran RSMI-NE/RSMI-NE/rsmine/coarsegrainer/build_dataset.py
code served (permissive licence) · get_code("7b89f9242dbba472")
loadNSplit_DimerandVBS Ran RSMI-NE/RSMI-NE/rsmine/coarsegrainer/cg_utils.py
code served (permissive licence) · get_code("9ef4718a1e2a6699")
non_collapsed_data Ran RSMI-NE/RSMI-NE/rsmine/coarsegrainer/analysis_utils.py
code served (permissive licence) · get_code("958a914e73170a8f")
non_collapsed_estimates Ran RSMI-NE/RSMI-NE/rsmine/coarsegrainer/analysis_utils.py
code served (permissive licence) · get_code("5c3a5208a2efd5f1")
round_up Ran RSMI-NE/RSMI-NE/rsmine/coarsegrainer/plotter.py
code served (permissive licence) · get_code("1b4c5ecd9ab8505e")
FF Not yet run karlstratos/doe/util.py
code served (permissive licence) · get_code("b344d686dec45a3b")
cg_configs Not yet run RSMI-NE/RSMI-NE/rsmine/coarsegrainer/cg_sequels.py
code served (permissive licence) · get_code("3a4cb53a40962c2b")
correlator Not yet run RSMI-NE/RSMI-NE/rsmine/coarsegrainer/cg_sequels.py
code served (permissive licence) · get_code("8c47f52617901f50")
mark_inset_hack Not yet run RSMI-NE/RSMI-NE/rsmine/coarsegrainer/plotter.py
code served (permissive licence) · get_code("1396b54f81fa12c7")

Repositories linked to this paper

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

Abstract

Estimating and optimizing Mutual Information (MI) is core to many problems in machine learning; however, bounding MI in high dimensions is challenging. To establish tractable and scalable objectives, recent work has turned to variational bounds parameterized by neural networks, but the relationships and tradeoffs between these bounds remains unclear. In this work, we unify these recent developments in a single framework. We find that the existing variational lower bounds degrade when the MI is large, exhibiting either high bias or high variance. To address this problem, we introduce a continuum of lower bounds that encompasses previous bounds and flexibly trades off bias and variance. On high-dimensional, controlled problems, we empirically characterize the bias and variance of the bounds and their gradients and demonstrate the effectiveness of our new bounds for estimation and representation learning.

For agents

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

get_harvested_code_for_paper("1905.06922")
get_code_for_paper("1905.06922")
have("1905.06922")

Connect an agent — have() is free.