Source author record

Michael Feindt

Michael Feindt appears in the imported research catalog. Authorship, coauthor and topic links are available while profile ownership is still unclaimed.

ResearcherUnclaimed source record

hep-ex physics.data-an physics.ins-det Machine Learning

Catalog footprint

What is connected

6works

4topics

4close collaborators

Actions

Connect this record

Open graph Browse works

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2021arXiv

Cyclic Boosting -- an explainable supervised machine learning algorithm

Supervised machine learning algorithms have seen spectacular advances and surpassed human level performance in a wide range of specific applications. However, using complex ensemble or deep learning algorithms typically results in black box models, where the path leading to individual predictions cannot be followed in detail. In order to address this issue, we propose the novel "Cyclic Boosting" machine learning algorithm, which allows to efficiently perform accurate regression and classification tasks while at the same time allowing a detailed understanding of how each individual prediction was made.

preprint2015arXiv

A neural network z-vertex trigger for Belle II

We present the concept of a track trigger for the Belle II experiment, based on a neural network approach, that is able to reconstruct the z (longitudinal) position of the event vertex within the latency of the first level trigger. The trigger will thus be able to suppress a large fraction of the dominating background from events outside of the interaction region. The trigger uses the drift time information of the hits from the Central Drift Chamber (CDC) of Belle II within narrow cones in polar and azimuthal angle as well as in transverse momentum (sectors), and estimates the z-vertex without explicit track reconstruction. The preprocessing for the track trigger is based on the track information provided by the standard CDC trigger. It takes input from the 2D ($r - φ$) track finder, adds information from the stereo wires of the CDC, and finds the appropriate sectors in the CDC for each track in a given event. Within each sector, the z-vertex of the associated track is estimated by a specialized neural network, with a continuous output corresponding to the scaled z-vertex. The input values for the neural network are calculated from the wire hits of the CDC.

preprint2015arXiv

A z-Vertex Trigger for Belle II

The Belle II experiment will go into operation at the upgraded SuperKEKB collider in 2016. SuperKEKB is designed to deliver an instantaneous luminosity $\mathcal{L}=8\times10^{35}\,\mathrm{cm}^{-2}\,\mathrm{s}^{-1}$. The experiment will therefore have to cope with a much larger machine background than its predecessor Belle, in particular from events outside of the interaction region. We present the concept of a track trigger, based on a neural network approach, that is able to suppress a large fraction of this background by reconstructing the $z$ (longitudinal) position of the event vertex within the latency of the first level trigger. The trigger uses the hit information from the Central Drift Chamber (CDC) of Belle II within narrow cones in polar and azimuthal angle as well as in transverse momentum ("sectors"), and estimates the $z$-vertex without explicit track reconstruction. The preprocessing for the track trigger is based on the track information provided by the standard CDC trigger. It takes input from the 2D track finder, adds information from the stereo wires of the CDC, and finds the appropriate sectors in the CDC for each track. Within the sector, the $z$-vertex is estimated by a specialized neural network, with the drift times from the CDC as input and a continuous output corresponding to the scaled $z$-vertex. The neural algorithm will be implemented in programmable hardware. To this end a Virtex 7 FPGA board will be used, which provides at present the most promising solution for a fully parallelized implementation of neural networks or alternative multivariate methods. A high speed interface for external memory will be integrated into the platform, to be able to store the $\mathcal{O}(10^9)$ parameters required. The contribution presents the results of our feasibility studies and discusses the details of the envisaged hardware solution.

preprint2014arXiv

An automated framework for hierarchical reconstruction of B mesons at the Belle II experiment

We present a software framework for Belle II that reconstructs B mesons in many decay modes with minimal user intervention. It does so by reconstructing particles in user-supplied decay channels, and then in turn using these reconstructed particles in higher-level decays. This hierarchical reconstruction allows one to cover a relatively high fraction of all B decays by specifying a limited number of particle decays. Multivariate classification methods are used to achieve a high signal-to-background ratio in each individual channel. The entire reconstruction, including the application of pre-cuts and classifier trainings, is automated to a high degree and will allow users to retrain to account for analysis-specific signal-side selections.

preprint2012arXiv

An algorithm for quantifying dependence in multivariate data sets

We describe an algorithm to quantify dependence in a multivariate data set. The algorithm is able to identify any linear and non-linear dependence in the data set by performing a hypothesis test for two variables being independent. As a result we obtain a reliable measure of dependence. In high energy physics understanding dependencies is especially important in multidimensional maximum likelihood analyses. We therefore describe the problem of a multidimensional maximum likelihood analysis applied on a multivariate data set with variables that are dependent on each other. We review common procedures used in high energy physics and show that general dependence is not the same as linear correlation and discuss their limitations in practical application. Finally we present the tool CAT, which is able to perform all reviewed methods in a fully automatic mode and creates an analysis report document with numeric results and visual review.

preprint2011arXiv

A Hierarchical NeuroBayes-based Algorithm for Full Reconstruction of B Mesons at B Factories

We describe a new B-meson full reconstruction algorithm designed for the Belle experiment at the B-factory KEKB, an asymmetric e+e- collider that collected a data sample of 771.6 x 10^6 BBbar pairs during its running time. To maximize the number of reconstructed B decay channels, it utilizes a hierarchical reconstruction procedure and probabilistic calculus instead of classical selection cuts. The multivariate analysis package NeuroBayes was used extensively to hold the balance between highest possible efficiency, robustness and acceptable consumption of CPU time. In total, 1104 exclusive decay channels were reconstructed, employing 71 neural networks altogether. Overall, we correctly reconstruct one B+/- or B0 candidate in 0.28% or 0.18% of the BBbar events, respectively. Compared to the cut-based classical reconstruction algorithm used at the Belle experiment, this is an improvement in efficiency by roughly a factor of 2, depending on the analysis considered. The new framework also features the ability to choose the desired purity or efficiency of the fully reconstructed sample freely. If the same purity as for the classical full reconstruction code is desired ~25%, the efficiency is still larger by nearly a factor of 2. If, on the other hand, the efficiency is chosen at a similar level as the classical full reconstruction, the purity rises from ~25% to nearly 90%.