Showing 1–10 of 10 results for author: Pollack, B

Explaining the Black-box Smoothly- A Counterfactual Approach

Authors: Sumedha Singla, Motahhare Eslami, Brian Pollack, Stephen Wallace, Kayhan Batmanghelich

Abstract: We propose a BlackBox Counterfactual Explainer, designed to explain image classification models for medical applications. Classical approaches (e.g., saliency maps) that assess feature importance do not explain "how" imaging features in important anatomical regions are relevant to the classification decision. Our framework explains the decision for a target class by gradually "exaggerating" the se… ▽ More We propose a BlackBox Counterfactual Explainer, designed to explain image classification models for medical applications. Classical approaches (e.g., saliency maps) that assess feature importance do not explain "how" imaging features in important anatomical regions are relevant to the classification decision. Our framework explains the decision for a target class by gradually "exaggerating" the semantic effect of the class in a query image. We adopted a Generative Adversarial Network (GAN) to generate a progressive set of perturbations to a query image, such that the classification decision changes from its original class to its negation. We used counterfactual explanations from our framework to audit a classifier trained on a chest x-ray dataset with multiple labels. We proposed clinically-relevant quantitative metrics such as cardiothoracic ratio and the score of a healthy costophrenic recess to evaluate our explanations. We conducted a human-grounded experiment with diagnostic radiology residents to compare different styles of explanations (no explanation, saliency map, cycleGAN explanation, and our counterfactual explanation) by evaluating different aspects of explanations: (1) understandability, (2) classifier's decision justification, (3) visual quality, (d) identity preservation, and (5) overall helpfulness of an explanation to the users. Our results show that our counterfactual explanation was the only explanation method that significantly improved the users' understanding of the classifier's decision compared to the no-explanation baseline. Our metrics established a benchmark for evaluating model explanation methods in medical images. Our explanations revealed that the classifier relied on clinically relevant radiographic features for its diagnostic decisions, thus making its decision-making process more transparent to the end-user. △ Less

Submitted 18 November, 2022; v1 submitted 11 January, 2021; originally announced January 2021.

Comments: Preprint Accepted in Medical image Analysis journal

HistomicsML2.0: Fast interactive machine learning for whole slide imaging data

Authors: Sanghoon Lee, Mohamed Amgad, Deepak R. Chittajallu, Matt McCormick, Brian P Pollack, Habiba Elfandy, Hagar Hussein, David A Gutman, Lee AD Cooper

Abstract: Extracting quantitative phenotypic information from whole-slide images presents significant challenges for investigators who are not experienced in develo** image analysis algorithms. We present new software that enables rapid learn-by-example training of machine learning classifiers for detection of histologic patterns in whole-slide imaging datasets. HistomicsML2.0 uses convolutional networks… ▽ More Extracting quantitative phenotypic information from whole-slide images presents significant challenges for investigators who are not experienced in develo** image analysis algorithms. We present new software that enables rapid learn-by-example training of machine learning classifiers for detection of histologic patterns in whole-slide imaging datasets. HistomicsML2.0 uses convolutional networks to be readily adaptable to a variety of applications, provides a web-based user interface, and is available as a software container to simplify deployment. △ Less

Submitted 30 January, 2020; originally announced January 2020.

Explanation by Progressive Exaggeration

Authors: Sumedha Singla, Brian Pollack, Junxiang Chen, Kayhan Batmanghelich

Abstract: As machine learning methods see greater adoption and implementation in high stakes applications such as medical image diagnosis, the need for model interpretability and explanation has become more critical. Classical approaches that assess feature importance (e.g. saliency maps) do not explain how and why a particular region of an image is relevant to the prediction. We propose a method that expla… ▽ More As machine learning methods see greater adoption and implementation in high stakes applications such as medical image diagnosis, the need for model interpretability and explanation has become more critical. Classical approaches that assess feature importance (e.g. saliency maps) do not explain how and why a particular region of an image is relevant to the prediction. We propose a method that explains the outcome of a classification black-box by gradually exaggerating the semantic effect of a given class. Given a query input to a classifier, our method produces a progressive set of plausible variations of that query, which gradually changes the posterior probability from its original class to its negation. These counter-factually generated samples preserve features unrelated to the classification decision, such that a user can employ our method as a "tuning knob" to traverse a data manifold while crossing the decision boundary. Our method is model agnostic and only requires the output value and gradient of the predictor with respect to its input. △ Less

Submitted 10 February, 2020; v1 submitted 1 November, 2019; originally announced November 2019.

Modeling Magnetic Fields with Helical Solutions to Laplace's Equation

Authors: Brian Pollack, Ryan Pellico, Cole Kampa, Henry Glass, Michael Schmitt

Abstract: The series solution to Laplace's equation in a helical coordinate system is derived and refined using symmetry and chirality arguments. These functions and their more commonplace counterparts are used to model solenoidal magnetic fields via linear, multidimensional curve-fitting. A judicious choice of functional forms, a small number of free parameters and sparse input data can lead to highly accu… ▽ More The series solution to Laplace's equation in a helical coordinate system is derived and refined using symmetry and chirality arguments. These functions and their more commonplace counterparts are used to model solenoidal magnetic fields via linear, multidimensional curve-fitting. A judicious choice of functional forms, a small number of free parameters and sparse input data can lead to highly accurate, fine-grained modeling of solenoidal magnetic fields, including helical features arising from the winding of the solenoid, with overall field accuracy at better than one part per million. △ Less

Submitted 6 July, 2020; v1 submitted 8 January, 2019; originally announced January 2019.

Comments: 22 pages, 16 figures

Report number: Report-no: FERMILAB-PUB-19-006-TD, NUHEP-EXP/19-02

Expression of Interest for Evolution of the Mu2e Experiment

Authors: F. Abusalma, D. Ambrose, A. Artikov, R. Bernstein, G. C. Blazey, C. Bloise, S. Boi, T. Bolton, J. Bono, R. Bonventre, D. Bowring, D. Brown, D. Brown, K. Byrum, M. Campbell, J. -F. Caron, F. Cervelli, D. Chokheli, K. Ciampa, R. Ciolini, R. Coleman, D. Cronin-Hennessy, R. Culbertson, M. A. Cummings, A. Daniel , et al. (103 additional authors not shown)

Abstract: We propose an evolution of the Mu2e experiment, called Mu2e-II, that would leverage advances in detector technology and utilize the increased proton intensity provided by the Fermilab PIP-II upgrade to improve the sensitivity for neutrinoless muon-to-electron conversion by one order of magnitude beyond the Mu2e experiment, providing the deepest probe of charged lepton flavor violation in the fores… ▽ More We propose an evolution of the Mu2e experiment, called Mu2e-II, that would leverage advances in detector technology and utilize the increased proton intensity provided by the Fermilab PIP-II upgrade to improve the sensitivity for neutrinoless muon-to-electron conversion by one order of magnitude beyond the Mu2e experiment, providing the deepest probe of charged lepton flavor violation in the foreseeable future. Mu2e-II will use as much of the Mu2e infrastructure as possible, providing, where required, improvements to the Mu2e apparatus to accommodate the increased beam intensity and cope with the accompanying increase in backgrounds. △ Less

Submitted 7 February, 2018; originally announced February 2018.

Comments: 17 pages, 4 figures, 1 table; Submitted to the Fermilab Physics Advisory Committee

Report number: Fermilab-FN-1052

Bayesian Block Histogramming for High Energy Physics

Authors: Brian Pollack, Saptaparna Bhattacharya, Michael Schmitt

Abstract: The Bayesian Block algorithm, originally developed for applications in astronomy, can be used to improve the binning of histograms in high energy physics. The visual improvement can be dramatic, as shown here with two simple examples. More importantly, this algorithm and the histogram is produces is a non-parametric density estimate, providing a description of background distributions that does no… ▽ More The Bayesian Block algorithm, originally developed for applications in astronomy, can be used to improve the binning of histograms in high energy physics. The visual improvement can be dramatic, as shown here with two simple examples. More importantly, this algorithm and the histogram is produces is a non-parametric density estimate, providing a description of background distributions that does not suffer from the arbitrariness of ad hoc analytical functions. The statistical power of an hypothesis test based on Bayesian Blocks is nearly as good as that obtained by fitting analytical functions. Two examples are provided: a narrow peak on a smoothly-falling background, and an excess in the tail of a background that falls rapidly over several orders of magnitude. These examples show the usefulness of the binning provided by the Bayesian Blocks algorithm both for presentation of data and when searching for new physics. △ Less

Submitted 13 June, 2019; v1 submitted 2 August, 2017; originally announced August 2017.

Comments: 12 pages

Report number: nuhep-exp/17-05

Design, Construction and Testing of the Digital Hadron Calorimeter (DHCAL) Electronics

Authors: C. Adams, A. Bambaugh, B. Bilki, J. Butler, F. Corriveau, T. Cundiff, G. Drake, K. Francis, V. Guarino, B. Haberichter, E. Hazen, J. Hoff, S. Holm, A. Kreps, P. DeLurgio, L. Dal Monte, N. Mucia, E. Norbeck, D. Northacker, Y. Onel, B. Pollack, J. Repond, J. Schlereth, J. R. Smith, D. Trojand , et al. (8 additional authors not shown)

Abstract: A novel hadron calorimeter is being developed for future lepton colliding beam detectors. The calorimeter is optimized for the application of Particle Flow Algorithms (PFAs) to the measurement of hadronic jets and features a very finely segmented readout with 1 x 1 cm2 cells. The active media of the calorimeter are Resistive Plate Chambers (RPCs) with a digital, i.e. one-bit, readout. To first ord… ▽ More A novel hadron calorimeter is being developed for future lepton colliding beam detectors. The calorimeter is optimized for the application of Particle Flow Algorithms (PFAs) to the measurement of hadronic jets and features a very finely segmented readout with 1 x 1 cm2 cells. The active media of the calorimeter are Resistive Plate Chambers (RPCs) with a digital, i.e. one-bit, readout. To first order the energy of incident particles in this calorimeter is reconstructed as being proportional to the number of pads with a signal over a given threshold. A large-scale prototype calorimeter with approximately 500,000 readout channels has been built and underwent extensive testing in the Fermilab and CERN test beams. This paper reports on the design, construction, and commissioning of the electronic readout system of this prototype calorimeter. The system is based on the DCAL front-end chip and a VME-based back-end. △ Less

Submitted 4 March, 2016; originally announced March 2016.

Design, Construction and Commissioning of the Digital Hadron Calorimeter - DHCAL

Authors: C. Adams, A. Bambaugh, B. Bilki, J. Butler, F. Corriveau, T. Cundiff, G. Drake, K. Francis, B. Furst, V. Guarino, B. Haberichter, E. Hazen, J. Hoff, S. Holm, A. Kreps, P. DeLurgio, Z. Matijas, L. Dal Monte, N. Mucia, E. Norbeck, D. Northacker, Y. Onel, B. Pollack, J. Repond, J. Schlereth , et al. (11 additional authors not shown)

Abstract: A novel hadron calorimeter is being developed for future lepton colliding beam detectors. The calorimeter is optimized for the application of Particle Flow Algorithms (PFAs) to the measurement of hadronic jets and features a very finely segmented readout with 1 x 1 cm2 cells. The active media of the calorimeter are Resistive Plate Chambers (RPCs) with a digital, i.e. one-bit, readout. To first ord… ▽ More A novel hadron calorimeter is being developed for future lepton colliding beam detectors. The calorimeter is optimized for the application of Particle Flow Algorithms (PFAs) to the measurement of hadronic jets and features a very finely segmented readout with 1 x 1 cm2 cells. The active media of the calorimeter are Resistive Plate Chambers (RPCs) with a digital, i.e. one-bit, readout. To first order the energy of incident particles in this calorimeter is reconstructed as being proportional to the number of pads with a signal over a given threshold. A large-scale prototype calorimeter with approximately 500,000 readout channels has been built and underwent extensive testing in the Fermilab and CERN test beams. This paper reports on the design, construction, and commissioning of this prototype calorimeter. △ Less

Submitted 4 March, 2016; originally announced March 2016.

Observation of the rare $B^0_s\toμ^+μ^-$ decay from the combined analysis of CMS and LHCb data

Authors: The CMS, LHCb Collaborations, :, V. Khachatryan, A. M. Sirunyan, A. Tumasyan, W. Adam, T. Bergauer, M. Dragicevic, J. Erö, M. Friedl, R. Frühwirth, V. M. Ghete, C. Hartl, N. Hörmann, J. Hrubec, M. Jeitler, W. Kiesenhofer, V. Knünz, M. Krammer, I. Krätschmer, D. Liko, I. Mikulec, D. Rabady, B. Rahbaran , et al. (2807 additional authors not shown)

Abstract: A joint measurement is presented of the branching fractions $B^0_s\toμ^+μ^-$ and $B^0\toμ^+μ^-$ in proton-proton collisions at the LHC by the CMS and LHCb experiments. The data samples were collected in 2011 at a centre-of-mass energy of 7 TeV, and in 2012 at 8 TeV. The combined analysis produces the first observation of the $B^0_s\toμ^+μ^-$ decay, with a statistical significance exceeding six sta… ▽ More A joint measurement is presented of the branching fractions $B^0_s\toμ^+μ^-$ and $B^0\toμ^+μ^-$ in proton-proton collisions at the LHC by the CMS and LHCb experiments. The data samples were collected in 2011 at a centre-of-mass energy of 7 TeV, and in 2012 at 8 TeV. The combined analysis produces the first observation of the $B^0_s\toμ^+μ^-$ decay, with a statistical significance exceeding six standard deviations, and the best measurement of its branching fraction so far. Furthermore, evidence for the $B^0\toμ^+μ^-$ decay is obtained with a statistical significance of three standard deviations. The branching fraction measurements are statistically compatible with SM predictions and impose stringent constraints on several theories beyond the SM. △ Less

Submitted 17 August, 2015; v1 submitted 17 November, 2014; originally announced November 2014.

Comments: Correspondence should be addressed to [email protected]

Report number: CERN-PH-EP-2014-220, CMS-BPH-13-007, LHCb-PAPER-2014-049

Journal ref: Nature 522, 68-72 (04 June 2015)

Fast Beam Condition Monitor for CMS: performance and upgrade

Authors: Jessica L. Leonard, Alan Bell, Piotr Burtowy, Anne Dabrowski, Maria Hempel, Hans Henschel, Wolfgang Lange, Wolfgang Lohmann, Nathaniel Odell, Marek Penno, Brian Pollack, Dominik Przyborowski, Vladimir Ryjov, David Stickland, Roberval Walsh, Weronika Warzycha, Agnieszka Zagozdzinska

Abstract: The CMS beam and radiation monitoring subsystem BCM1F (Fast Beam Condition Monitor) consists of 8 individual diamond sensors situated around the beam pipe within the pixel detector volume, for the purpose of fast bunch-by-bunch monitoring of beam background and collision products. In addition, effort is ongoing to use BCM1F as an online luminosity monitor. BCM1F will be running whenever there is b… ▽ More The CMS beam and radiation monitoring subsystem BCM1F (Fast Beam Condition Monitor) consists of 8 individual diamond sensors situated around the beam pipe within the pixel detector volume, for the purpose of fast bunch-by-bunch monitoring of beam background and collision products. In addition, effort is ongoing to use BCM1F as an online luminosity monitor. BCM1F will be running whenever there is beam in LHC, and its data acquisition is independent from the data acquisition of the CMS detector, hence it delivers luminosity even when CMS is not taking data. A report is given on the performance of BCM1F during LHC run I, including results of the van der Meer scan and on-line luminosity monitoring done in 2012. In order to match the requirements due to higher luminosity and 25 ns bunch spacing, several changes to the system must be implemented during the upcoming shutdown, including upgraded electronics and precise gain monitoring. First results from Run II preparation are shown. △ Less

Submitted 8 May, 2014; originally announced May 2014.

Comments: 10 pages, 8 figures. To be published in NIM A as proceedings for the 9th Hiroshima Symposium on Semiconductor Tracking Detectors (2013)

Report number: DESY 14-070

Journal ref: Nuclear Inst. and Methods in Physics Research, A (2014), pp. 235-239