SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1706.04599 · ICML · 2017

On Calibration of Modern Neural Networks

Geoff Pleiss, Kilian Weinberger, Yu Sun, Chuan Guo

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 4 functions out of this paper's own repositories and ran 4 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
saurabhgarg1996/calibration reimplementation 3 of 3
sirius8050/Expected-Calibration-Error reimplementation 1 of 1
FunctionStatusWhere it lives
add_softmax Ran saurabhgarg1996/calibration/calibration/calibrators.py
code served (permissive licence) · get_code("5ac25fc20a9cb76f")
cross_entropy_loss Ran saurabhgarg1996/calibration/calibration/calibrators.py
code served (permissive licence) · get_code("155f63d9cf89a950")
ece_loss Ran saurabhgarg1996/calibration/calibration/calibrators.py
code served (permissive licence) · get_code("2bc928c2a34c7e13")
ece_score Ran sirius8050/Expected-Calibration-Error/ECE.py
pointer only (licence: NONE) · get_code("aaabd8e8f8f13cc9")

Repositories linked to this paper

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

Abstract

Confidence calibration -the problem of predicting probability estimates representative of the true correctness likelihood -is important for classification models in many applications. We discover that modern neural networks, unlike those from a decade ago, are poorly calibrated. Through extensive experiments, we observe that depth, width, weight decay, and Batch Normalization are important factors influencing calibration. We evaluate the performance of various post-processing calibration methods on state-ofthe-art architectures with image and document classification datasets. Our analysis and experiments not only offer insights into neural network learning, but also provide a simple and straightforward recipe for practical settings: on most datasets, temperature scaling -a singleparameter variant of Platt Scaling -is surprisingly effective at calibrating predictions.

For agents

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

get_harvested_code_for_paper("1706.04599")
get_code_for_paper("1706.04599")
have("1706.04599")

Connect an agent — have() is free.