Identifying the relevant dependencies of the neural network response on characteristics of the input space

Abstract

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.

References in corpus (19)

• Adam: A Method for Stochastic Optimization
• Observation of a new particle in the search for the Standard Model Higgs boson with the ATLAS detector at the LHC
• Observation of a new boson at a mass of 125 GeV with the CMS experiment at the LHC
• TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems
• Asymptotic formulae for likelihood-based tests of new physics
• Confidence Level Computation for Combining Searches with Small Statistics
• Explaining NonLinear Classification Decisions with Deep Taylor Decomposition
• Combined results of searches for the standard model Higgs boson in pp collisions at sqrt(s) = 7 TeV
• Classification without labels: Learning from mixed samples in high energy physics
• Jet-Images -- Deep Learning Edition
• Evidence for the 125 GeV Higgs boson decaying to a pair of tau leptons
• Evidence for the Higgs-boson Yukawa coupling to tau leptons with the ATLAS detector
• Observation of the Higgs boson decay to a pair of tau leptons with the CMS detector
• Bounding the Space of Holographic CFTs with Chaos
• What is the Machine Learning?
• Learning to Classify from Impure Samples with High-Dimensional Data
• Weakly Supervised Classification in High Energy Physics
• (Machine) Learning to Do More with Less
• On holographic entanglement entropy of Horndeski black holes

Cited by in corpus (10)

• A guide for deploying Deep Learning in LHC searches: How to achieve optimality and account for uncertainty
• Mapping Machine-Learned Physics into a Human-Readable Space
• Search for a heavy Higgs boson decaying into two lighter Higgs bosons in the $\tau \tau$bb final state at 13 TeV
• Measurements of Higgs boson production in the decay channel with a pair of $\tau$ leptons in proton-proton collisions at $\sqrt{s}$ = 13 TeV
• Reducing the dependence of the neural network function to systematic uncertainties in the input space
• Searches for Higgs boson production through decays of heavy resonances
• Learning to Isolate Muons
• Measurement of the $\mathrm{t}\bar{\mathrm{t}}$H and tH production rates in the H $\to$ $\mathrm{b}\bar{\mathrm{b}}$ decay channel using proton-proton collision data at $\sqrt{s}$ = 13 TeV
• Development of systematic uncertainty-aware neural network trainings for binned-likelihood analyses at the LHC
• Creating Simple, Interpretable Anomaly Detectors for New Physics in Jet Substructure