Identifying the Relevant Dependencies of the Neural Network Response on Characteristics of the Input Space

The relation between the input and output spaces of neural networks (NNs) is investigated to identify those characteristics of the input space that have a large influence on the output for a given task. For this purpose, the NN function is decomposed into a Taylor expansion in each element of the input space. The Taylor coefficients contain information about the sensitivity of the NN response to the inputs. A metric is introduced that allows for the identification of the characteristics that mostly determine the performance of the NN in solving a given task. Finally, the capability of this metric to analyze the performance of the NN is evaluated based on a task common to data analyses in high-energy particle physics experiments.

• Publication year

  2018

• Venue

  Computing and Software for Big Science

• Publication date

  2018-03-23

• Fields of study

  Mathematics, Physics, Computer Science

• Identifiers
  DOI 10.1007/s41781-018-0012-1arXiv 1803.08782
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Observation of the Higgs boson decay to a pair of τ leptons with the CMS detector

  2018cited by this paper
• Learning to classify from impure samples with high-dimensional data

  2018cited by this paper
• What is the Machine Learning

  2017cited by this paper
• (Machine) learning to do more with less

  2017cited by this paper
• Classification without labels: learning from mixed samples in high energy physics

  2017cited by this paper
• Automatic differentiation in PyTorch

  2017influential reference
• Weakly supervised classification in high energy physics

  2017cited by this paper
• On holographic entanglement entropy of Horndeski black holes

  2017cited by this paper
• TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems

  2016influential reference
• Learning to Pivot with Adversarial Networks

  2016cited by this paper
• On-shell diagrams for N$$ \mathcal{N} $$ = 8 supergravity amplitudes

  2016cited by this paper
• Evidence for the Higgs-boson Yukawa coupling to tau leptons with the ATLAS detector

  2015cited by this paper
• Evidence for the Higgs-boson Yukawa coupling to tau leptons with the ATLAS detector

  2015cited by this paper
• On Pixel-Wise Explanations for Non-Linear Classifier Decisions by Layer-Wise Relevance Propagation

  2015cited by this paper
• Jet-images — deep learning edition

  2015cited by this paper
• Learning to discover: the Higgs boson machine learning challenge

  2014influential reference
• Evidence for the 125 GeV Higgs boson decaying to a pair of τ leptons

  2014cited by this paper
• Adam: A Method for Stochastic Optimization

  2014cited by this paper
• Observation of a new particle in the search for the Standard Model Higgs boson with the ATLAS detector at the LHC

  2012cited by this paper
• Observation of a new boson at a mass of 125 GeV with the CMS experiment at the LHC

  2012cited by this paper
• Combined results of searches for the standard model Higgs boson in pp collisions at √s = 7 TeV

  2012cited by this paper
• Observation of a new boson at a mass of 125 GeV with the CMS experiment at the LHC

  2012cited by this paper
• Observation of a new particle in the search for the Standard Model Higgs boson with the ATLAS detector at the LHC

  2012cited by this paper
• Procedure for the LHC Higgs boson search combination in Summer 2011

  2011cited by this paper
• Author ' s personal copy A Fast Quartet tree heuristic for hierarchical clustering

  2010cited by this paper
• Theano: A CPU and GPU Math Compiler in Python

  2010cited by this paper
• Asymptotic formulae for likelihood-based tests of new physics

  2010cited by this paper
• Full regularization path for sparse principal component analysis

  2007cited by this paper
• Presentation of search results: the CLs technique

  2002cited by this paper
• Confidence Level Computation for Combining Searches with Small Statistics

  1999cited by this paper

• An Optimal Observable Machine for reinterpretable measurements in high-energy physics

  2026cites this paper
• Development of systematic uncertainty-aware neural network trainings for binned-likelihood analyses at the LHC

  2025cites this paper
• Searches for Higgs boson production through decays of heavy resonances

  2024cites this paper
• Tran-Switch: A transfer learning approach for sentence level cross-genre author profiling on code-switched English-RomanUrdu Text

  2023cites this paper
• Measurements of Higgs boson production in the decay channel with a pair of leptons in proton–proton collisions at TeV

  2023cites this paper
• Creating simple, interpretable anomaly detectors for new physics in jet substructure

  2022cites this paper
• Measurements of Higgs boson production in the decay channel with a pair of $$\uptau $$ τ leptons in proton–proton collisions at

  2022cites this paper
• Learning to isolate muons

  2021cites this paper
• Search for a heavy Higgs boson decaying into two lighter Higgs bosons in the ττbb final state at 13 TeV

  2021cites this paper
• Mapping machine-learned physics into a human-readable space

  2020cites this paper
• A guide for deploying Deep Learning in LHC searches: How to achieve optimality and account for uncertainty

  2019cites this paper
• Reducing the Dependence of the Neural Network Function to Systematic Uncertainties in the Input Space

  2019cites this paper
• on the CERN CDS information server CMS PAS HIG-18-032 CMS Physics Analysis Summary

  2019cites this paper
• A Complete Bibliography of Publications in Computing and Software for Big Science

  year unknowncites this paper