Exploring Artificial Neural Networks to discover the Higgs boson at the LHC.

Slides:



Advertisements
Similar presentations
Martin zur Nedden, HU Berlin 1 WS 2007/08: Physik am LHC 5. Higgs Searches The Higgs-mechanism in the SM Yukava-coupling, masses of fermions Higgs Production:
Advertisements

Bruce Kennedy, RAL PPD Particle Physics 2 Bruce Kennedy RAL PPD.
Experimental Particle Physics PHYS6011 Joel Goldstein, RAL 1.Introduction & Accelerators 2.Particle Interactions and Detectors (2) 3.Collider Experiments.
5th May 2010Fergus Wilson, RAL1 Experimental Particle Physics PHYS6011 Looking for Higgs and SUSY at the LHC or...what can you get for $10,000,000,000.
Higgs physics theory aspects experimental approaches Monika Jurcovicova Department of Nuclear Physics, Comenius University Bratislava H f ~ m f.
Slides from: Doug Gray, David Poole
ACAT05 May , 2005 DESY, Zeuthen, Germany Search for the Higgs boson at LHC by using Genetic Algorithms Mostafa MJAHED Ecole Royale de l’Air, Mathematics.
Current limits (95% C.L.): LEP direct searches m H > GeV Global fit to precision EW data (excludes direct search results) m H < 157 GeV Latest Tevatron.
First look of ME study on ttH(  bb) channel Univ. of Tokyo Y.Kataoka
1 Analysis of Prompt Diphoton Production at the Large Hadron Collider. Andy Yen Mentor: Harvey Newman Co-Mentors: Marat Gataullin, Vladimir Litvine California.
Higgs Searches using Vector Boson Fusion. 2 Why a “Low Mass” Higgs (1) M H
THE SEARCH FOR THE HIGGS BOSON Aungshuman Zaman Department of Physics and Astronomy Stony Brook University October 11, 2010.
Artificial Neural Networks - Introduction -
Machine Learning Neural Networks
September 27, 2005FAKT 2005, ViennaSlide 1 B-Tagging and ttH, H → bb Analysis on Fully Simulated Events in the ATLAS Experiment A.H. Wildauer Universität.
Summary of Results and Projected Precision Rediscovering the Top Quark Marc-André Pleier, Universität Bonn Top Quark Pair Production and Decay According.
Introduction to Single-Top Single-Top Cross Section Measurements at ATLAS Patrick Ryan (Michigan State University) The measurement.
Search for resonances The fingerprints of the Top Quark Jessica Levêque, University of Arizona Top Quark Mass Measurement Top Turns Ten Symposium, Fermilab,
Single-Top Cross Section Measurements at ATLAS Patrick Ryan (Michigan State University) Introduction to Single-Top The measurement.
Top Quark Physics: An Overview Young Scientists’ Workshop, Ringberg castle, July 21 st 2006 Andrea Bangert.
Application of Neural Networks for Energy Reconstruction J. Damgov and L. Litov University of Sofia.
August 22, 2002UCI Quarknet The Higgs Particle Sarah D. Johnson University of La Verne August 22, 2002.
Neural Networks AI – Week 23 Sub-symbolic AI Multi-Layer Neural Networks Lee McCluskey, room 3/10
1 ZH Analysis Yambazi Banda, Tomas Lastovicka Oxford SiD Collaboration Meeting
LINEAR CLASSIFICATION. Biological inspirations  Some numbers…  The human brain contains about 10 billion nerve cells ( neurons )  Each neuron is connected.
1 Physics at Hadron Colliders Lecture III Beate Heinemann University of California, Berkeley and Lawrence Berkeley National Laboratory CERN, Summer Student.
Artificial Intelligence Methods Neural Networks Lecture 4 Rakesh K. Bissoondeeal Rakesh K. Bissoondeeal.
Artificial Intelligence Techniques Multilayer Perceptrons.
1 A Preliminary Model Independent Study of the Reaction pp  qqWW  qq ℓ qq at CMS  Gianluca CERMINARA (SUMMER STUDENT)  MUON group.
Associated top Higgs search: with ttH (H  bb) Chris Collins-Tooth, 17 June 2008.
Gideon Bella Tel Aviv University On behalf of the ATLAS collaboration ATL-PHYS-PUB ATL-PHYS-PUB Prospects of measuring ZZ and WZ polarization.
Higgs Properties Measurement based on HZZ*4l with ATLAS
Janice Drohan, Simon Dean, Nikos Konstantinidis UCL Christmas Meeting 12/12/2005 Search for Low Mass SM Higgs in the Channel ttH(H  bb)
Possibility of tan  measurement with in CMS Majid Hashemi CERN, CMS IPM,Tehran,Iran QCD and Hadronic Interactions, March 2005, La Thuile, Italy.
October 19, 2000ACAT 2000, Fermilab, Suman B. Beri Top Quark Mass Measurements Using Neural Networks Suman B. Beri, Rajwant Kaur Panjab University, India.
Moriond QCD 2001Enrico Piotto EP division CERN SM Higgs Search at LEP in Channels other than 4 Jets Enrico Piotto EP Division, CERN SM Higgs Search at.
Report on LHT at the LHC ~ Some results from simulation study ~ Shigeki Matsumoto (Univ. of Toyama) 1.What kinds of LHT signals are expected, and how accurately.
Precision Cross section measurements at LHC (CMS) Some remarks from the Binn workshop André Holzner IPP ETH Zürich DIS 2004 Štrbské Pleso Štrbské Pleso.
Back-Propagation Algorithm AN INTRODUCTION TO LEARNING INTERNAL REPRESENTATIONS BY ERROR PROPAGATION Presented by: Kunal Parmar UHID:
Physics at LHC Prague, 6-12 July, 2003 R. Kinnunen Helsinki Institute of Physics A/H ->  and H + ->  in CMS R. Kinnunen Physics at LHC Prague July 6.
Higgs Reach Through VBF with ATLAS Bruce Mellado University of Wisconsin-Madison Recontres de Moriond 2004 QCD and High Energy Hadronic Interactions.
BESS Model Resonance in the pp W + W tt + X Channel at LHC M. Gintner, I. Melo, B. Trpišová University of Žilina Herlany, September 2006.
Feasibility study of Higgs pair production in a Photon Collider Tohru Takahashi Hiroshima University for S.Kawada, N.Maeda, K.Ikematsu, K.Fujii,Y.Kurihara,,,
Top Quark Physics At TeVatron and LHC. Overview A Lightning Review of the Standard Model Introducing the Top Quark tt* Pair Production Single Top Production.
SEARCH FOR AN INVISIBLE HIGGS IN tth EVENTS T.L.Cheng, G.Kilvington, R.Goncalo Motivation The search for the Higgs boson is a window on physics beyond.
Analysis of H  WW  l l Based on Boosted Decision Trees Hai-Jun Yang University of Michigan (with T.S. Dai, X.F. Li, B. Zhou) ATLAS Higgs Meeting September.
Alexei Safonov (Texas A&M University) For the CDF Collaboration.
From the Standard Model to Discoveries - Physics with the CMS Experiment at the Dawn of the LHC Era Dimitri Bourilkov University of Florida CMS Collaboration.
Searching for the Higgs boson in the VH and VBF channels at ATLAS and CMS Dr. Adrian Buzatu Research Associate.
Particle Physics II Chris Parkes Top Quark Discovery Decay Higgs Searches Indirect mW and mt Direct LEP & LHC searches 2 nd Handout.
Régis Lefèvre (LPC Clermont-Ferrand - France)ATLAS Physics Workshop - Lund - September 2001 In situ jet energy calibration General considerations The different.
29 August, 2007 Ashfaq Ahmad, Search for Charged Higgs at the LHC 1 Search for Charged Higgs at the LHC Ashfaq Ahmad (Stony Brook)
Top Higgs Yukawa Coupling Analysis – Status Report Hajrah Tabassam Quai-i-Azam University, Islamabad ON BEHALF OF: R. Yonamine, T. Tanabe, K. Fujii, KEK.
La Thuile, March, 15 th, 2003 f Makoto Tomoto ( FNAL ) Prospects for Higgs Searches at DØ Makoto Tomoto Fermi National Accelerator Laboratory (For the.
SPS5 SUSY STUDIES AT ATLAS Iris Borjanovic Institute of Physics, Belgrade.
Background Shape Study for the ttH, H  bb Channel Catrin Bernius First year talk 15th June 2007 Background Shape Study for the ttH 0, H 0  bb Channel.
Backup slides Z 0 Z 0 production Once  s > 2M Z ~ GeV ÞPair production of Z 0 Z 0 via t-channel electron exchange. e+e+ e-e- e Z0Z0 Z0Z0 Other.
Kim HS Introduction considering that the amount of MRI data to analyze in present-day clinical trials is often on the order of hundreds or.
SEARCH FOR DIRECT PRODUCTION OF SUPERSYMMETRIC PAIRS OF TOP QUARKS AT √ S = 8 TEV, WITH ONE LEPTON IN THE FINAL STATE. Juan Pablo Gómez Cardona PhD Candidate.
Viktor Veszpremi Purdue University, CDF Collaboration Tev4LHC Workshop, Oct , Fermilab ZH->vvbb results from CDF.
Artificial Neural Networks An Introduction. Outline Introduction Biological and artificial neurons Perceptrons (problems) Backpropagation network Training.
Eric COGNERAS LPC Clermont-Ferrand Prospects for Top pair resonance searches in ATLAS Workshop on Top Physics october 2007, Grenoble.
1 Donatella Lucchesi July 22, 2010 Standard Model High Mass Higgs Searches at CDF Donatella Lucchesi For the CDF Collaboration University and INFN of Padova.
Exploring Artificial Neural Networks to Discover Higgs at LHC Using Neural Networks for B-tagging By Rohan Adur
Determining the CP Properties of a Light Higgs Boson
ttH (Hγγ) search and CP measurement
with Daniel L. Silver, Ph.D. Christian Frey, BBA April 11-12, 2017
SUSY SEARCHES WITH ATLAS
Experimental and theoretical Group Torino + Moscow
Search for a New Vector Resonance in the pp WWtt+X Channel at LHC
Presentation transcript:

Exploring Artificial Neural Networks to discover the Higgs boson at the LHC

Overview Introduction –The Standard Model and the mass problem –Higgs search at the LHC (and ANNs) ttH, H to bb and other channels –Production process –Decay –Experimental signatures –Background processes ANNs a possible solution (theory) ANN development issues (on simple 2-D classification problem + results) ANNs applied to Higgs data (results) Summary

Introduction Origin of mass is last big unanswered question in the SM. The Standard Model –To make physical equations in SM gauge invariant we require new terms, these correspond directly to gauge bosons. (eg photon) –Massless particles would preserve SM’s gauge symmetry (easiest, but not the case) – Higgs mechanism allows generation of mass in the SM (by breaking gauge invariance of vacuum) (spontaneous symmetry breaking) – Needs further particle: HIGGS BOSON!! So search for Higgs is important to our understanding of particle interactions. It may be that nature has chosen another mass-generating mechanism, but whatever this mechanism is, it should show itself at the LHC.

Search at the Large Hadron Collider (LHC) (Higgs discovery one of its main aims) H mass not predicted by the SM but production and decay rates can be predicted as function of m H. –From LEP: 114.4GeV < m H (SM). –The LHC, with its detectors ATLAS and CMS, (due to go online in 2007) will collide p-p at 14TeV Higgs mass reach to about 1TeV. –High Luminosity; But Higgs production v. rare! (10 16 proton-proton interactions will occur per year, but less than 100,000 Higgs bosons will be produced ) –As well as Higgs, LHC hopes to find evidence for new physics Supersymmetry (SUSY) (modifies the SM to include a whole new series of particles, supersymmetric partners of all the particles so far known. Has many desirable features, mending some shortcomings of the SM. If SUSY is the theory, we do not know how many Higgs bosons we would see (minimum 5)

ttH, H to bb Channel H production processes: –Gluon fusion is dominant Higgs production process, gg to H (but difficult to separate signal from large QCD background) –Associated Production! ttH, Lower cross section but has leptonic final states Dominant decay mode at m H <130GeV is H to bb bb WW ZZ

ttH, H to bb could account for half the Higgs discovery potential at ATLAS (Cammin) Background; –ttjj (most important, 94% after TDR analysis) Full reconstruction of final state is necessary to minimize combinatorial background and to discriminate signal from large bg. TDR analysis has 3 steps: Preselection -1 isolated lepton, -At least 6 jets, -Exactly 4 tagged as b-jets. Reconstruction -reconstruction of 2 top quarks, minimise: Δ 2 = (m lvb – m t ) 2 + (m jjb – m t ) 2 Cuts on the reconstructed t and H masses (where ANNs come in) -Reconstructed top masses must be within  20GeV of m t. -m bb = m H  30GeV

After this TDR analysis, significance S/√B = 1.94 (for 120GeV Higgs) Could increase significance by: –Better jet pairing –Improving ‘final selection’ (after t reconstruction, apply to events in m bb = m H  30GeV) Applying ANNs promising as makes use of event topology, not just mass cuts! Ie minimising (Δ 2 eqn) does not take into account additional info such as spatial differences between jets! I looked at final selection! Used 10 variables generated by Pythia. (which gave separation in signal and background distributions) Fed variables into a neural network. (to classify event as signal or background)

ANNs Artificial Neural Networks (ANNs) are computational modelling tools Inspired by biological nervous system Good at: –generalization, –non-linear, –learn by example. Want to train network with examples to recognise right data( classification task) and reject rest (ANNs perform better than cut based in theory because can separate classes in feature space non linearly) (but training is difficult, optimisation harder than for cut based methods) x1x1 x2x2 x1x1 x2x2

oioi w ij hjhj w jk xkxk Response function: o i =g(∑ i w ij g(∑ k w jk x k )) Which is non-linear so network able to perform non-linear mappings Architecture and weight settings are what change classification! We want network to output 1 for signal and 0 for all background A neural node A neural network How do ANNs Work?

Weights are changed in proportion to the difference (error) between target output and actual network output for each example. Minimize summed square error function: E = 1/2 ∑ p ∑ i (o i (p) - t i (p) ) 2 with respect to the weights. Error is function of all the weights and forms an irregular multidimensional complex hyperplane with many peaks, saddle points and minima. Error minimized by finding set of weights that correspond to global minimum. (ie get close to 1 for signal and close to 0 for background) Done with gradient descent method – (weights incrementally updated in proportion to δE/δ w ij ) Error Surface

Summary of learning algorithm 1.Initialize w ij and w jk with random values. 2.Pick pattern p from training set. Present input and calculate the output from: o i =g(∑ i w ij g(∑ k w jk x k )) Update weights according to: w ij (t + 1) = w ij (t) – Δw ij w jk (t + 1) = w jk (t) – Δw jk where Δw = -η δE/δw. (…etc…for extra hidden layers). When no change (within some accuracy) occurs, the weights are frozen and network is ready to use on data it has never seen.

2-D problem Initially looked at simple ANN classification problem; –Separate out a single point in a 2-D plane of randomly generated numbers. Generated 2 sets of random numbers Fed network (using SNNS)(2 input 1 output) (show diag!!) examples of signal and background data. (desired output 1 and 0 respectively) Used 300 patterns in both tr. And val sets. Background to signal ratio was 3 to 1. Looked at various net architectures. Results: –Learning shown by error curves –Projections show hyperplanes –3 hidden nodes solve classification task fully! (effectively 1 hidden node is equiv. of 1 linear hyperplane)

–Got spiking behaviour of some error curves. Showed inconsistent learning (updating of weights) Was solved by adjusting some network params (made learning more stable!!!) –Learning parameter, η. –dmax. –Shuffle option. To get a deeper understanding of learning, also looked at weight and bias variables.

Using ANNs for Higgs search Worked with data after reconstruction of top quarks. Variables used; –mbb: the invariant mass of the two b-jets assigned to the Higgs boson, –Δη(tnear, bb): the difference in pseudo rapidity between the bb-system and the reconstructed top quark nearest ΔR. –cos  b,b*: the cosine of the decay angle of the two b-jets from the Higgs boson in the rest frame of the bb-system, –Δη(b,b): the difference in pseudo rapidity between the two b-jets from the Higgs boson, –mbb(1): the combination with the smallest invariant mass mbb out of the six combintations which are possible when selecting two b-jets out of four b-jets, –mbb(2): the combination with the second smallest invariant mass mbb out of the six combinations which are possible when selecting two b-jets out of four b-jets, –  t1-  t2: the difference in phi between the reconstructed top quarks, –pTt1+pTt2: the sum of the transverse momenta of the reconstructed top quarks.

Signal is RED

(only ttjj background used) Rescaled data to [0,1] Separated data into tr and val sets Used 1:1 for signal to background. Looked at various archs (1 and 2 hidden layers) Weak generalization:

Output for best architecture ( ) gave: Signal is RED

Summary Optimisation difficulties and solutions have been identified in net development Some classification produced for Higgs data More work on arch could be needed (more data, lack of generalization) s/√B as fn. of cut on output.

ANN development Require training and validation sets Difficulties in optimisation: –Several factors need to be considered Architecture Learning params (ie stepwidth, local minima) Data size –Finding optimum (of parameter settings) is largely trial and error (rules of thumb)!!! Optimum network Testing Training Error (eg SSE) No. of hidden nodes or training cycles