1. Using jet substructure and boosted objects: Measurements, searches, coping with pileup And something on measurements in general Jonathan Butterworth UCL MCnet School, Ambleside 29/8/2014Jon Butterworth, UCL1

2. 29/8/2014Jon Butterworth, UCL2

3. 1.Detectors meeting new challenges Unprecedented experimental environment at the LHC 29/8/2014Jon Butterworth, UCL3

4. 29/8/2014Jon Butterworth, UCL4

5. 1.Detectors meeting new challenges 2.Energy! Unprecedented experimental environment at the LHC 29/8/2014Jon Butterworth, UCL5

7. Jet properties Final stage of jet structure is “soft” non-perturbative QCD. Formation of hadrons from gluons, 100 MeV energy scales (  QCD ) Vast phase space between quark-gluon scatter (100’s GeV, few TeV) and  QCD Most of jet substructure can be analysed perturbatively EWSB scale (~100 GeV) lies in this region Jets may contain objects with EW-scale mass (W,Z,H,t,?) 29/8/2014Jon Butterworth, UCL7

8. 1.Detectors meeting new challenges 2.Energy! In other words, for the the first time, boosted electroweak-scale objects are “commonplace” (see Simone’s talk) Unprecedented experimental environment at the LHC 29/8/2014Jon Butterworth, UCL8

9. So where are we now? Happily, the Boost 2014 workshop was last week at UCL Borrowing liberally from the speakers there, especially Nhan Tran (FNAL) who gave the excellent experimental summary excellent experimental summary

10. Where are we now? You heard from Simone that the idea of interrogating jet substructure to identify decaying massive particles has sparked a lot of interest (phenomenology, theory, experiment) How far has it really gone with data?

11. Measuring and calibrating variables and taggers

12. Measurements with inclusive jets arXiv:1203.4606

13. Measurements with inclusive jets k T scale, N-subjettiness arXiv:1203.4606

14. 29/8/2014Jon Butterworth, UCL14 N. Tran

15. 29/8/2014Jon Butterworth, UCL15 N. Tran

16. 29/8/2014Jon Butterworth, UCL16 N.Tran

17. Shower deconstruction (Soper & Spannowsky) and ATLAS-CONF-2014-003 (relatively new example) 29/8/2014Jon Butterworth, UCL17

18. Shower deconstruction (Soper & Spannowsky) and ATLAS-CONF-2014-003 29/8/2014Jon Butterworth, UCL18

19. 29/8/2014Jon Butterworth, UCL19 Shower deconstruction (Soper & Spannowsky) and ATLAS-CONF-2014-003

20. 29/8/2014Jon Butterworth, UCL20 Shower deconstruction (Soper & Spannowsky) and ATLAS-CONF-2014-003

21. 29/8/2014Jon Butterworth, UCL21 Shower deconstruction (Soper & Spannowsky) and ATLAS-CONF-2014-003

22. 29/8/2014Jon Butterworth, UCL22 Shower deconstruction (Soper & Spannowsky) and ATLAS-CONF-2014-003

23. 29/8/2014Jon Butterworth, UCL23 Shower deconstruction (Soper & Spannowsky) and ATLAS-CONF-2014-003

24. 29/8/2014Jon Butterworth, UCL24 Shower deconstruction (Soper & Spannowsky) and ATLAS-CONF-2014-003

25. 29/8/2014Jon Butterworth, UCL25 Shower deconstruction (Soper & Spannowsky) and ATLAS-CONF-2014-003

26. Substructure in searches 29/8/2014Jon Butterworth, UCL26

27. 29/8/2014Jon Butterworth, UCL27 N.Tran

28. Handling pile-up

29. 29/8/2014Jon Butterworth, UCL29 N.Tran

30. 29/8/2014Jon Butterworth, UCL30 N. Tran

31. Applications in searches

32. 29/8/2014Jon Butterworth, UCL32 N. Tran

33. 29/8/2014Jon Butterworth, UCL33 N. Tran

34. 29/8/2014Jon Butterworth, UCL34 N. Tran

35. 29/8/2014Jon Butterworth, UCL35 N. Tran

36. 29/8/2014Jon Butterworth, UCL36 M.Selvaggi

37. 29/8/2014Jon Butterworth, UCL37 M.Selvaggi

38. 29/8/2014Jon Butterworth, UCL38 M.Selvaggi

39. Jet shapes, Q-jets, jet charge, jet pull, jet superstructure… Moving on from the “two goals” (tagging & grooming) to finding still more information; identifying jet properties in context, learning more about the short-distance physics. … also 29/8/2014Jon Butterworth, UCL39

40. Gone from wacky new idea to obviously essential item in the toolkit (~5 years) Studies with data show: modelling is adequate but can be improved in some cases Grooming is robust against pile up and important for controlling it in all jet measurements Not seen full potential yet in Higgs searches (cf 14 TeV) Subjets and measuring boosted objects 29/8/2014Jon Butterworth, UCL40

41. Already widely used in searches (mainly boosted tops, but results on H, W/Z coming) Calibration work on of novel observables is intense & interesting Theoretical work on developing (analytic) understanding (See Simone’s talk) Boost meetings 2009 Stanford 2010 Oxford 2011 Princeton 2012 Valencia 2013 Arizona 2014 UCL (Last week) Subjets and measuring boosted objects 29/8/2014Jon Butterworth, UCL41

42. 29/8/2014Jon Butterworth, UCL42 Or, as Nhan put it…

43. And finally… What do we actually measure? Difference between “efficiency corrections” or “unfolding”, and “acceptance corrections”. The first two generally mean correction for detector effects which no one but the experimentalists can do. The third means extrapolating into kinematic regions which have not been measured at all Beware of the third, especially as we go to higher energies… 29/8/2014Jon Butterworth, UCL43

44. 29/8/2014Jon Butterworth, UCL44 Unfold

45. 29/8/2014Jon Butterworth, UCL45

46. 29/8/2014Jon Butterworth, UCL46 Increase acceptance

47. 29/8/2014Jon Butterworth, UCL47 Increase acceptance

48. 29/8/2014Jon Butterworth, UCL48 Extrapolate

49. 29/8/2014Jon Butterworth, UCL49 Extrapolate

50. 29/8/2014Jon Butterworth, UCL50 But how reliably?

51. 29/8/2014Jon Butterworth, UCL51

52. Concept of a “fiducial” cross section Defines a region in which acceptance is ~100% Implies that some kinematic cuts must be implemented in whatever theory is compared to (easy for MC, less so for some high-order calculations) 29/8/2014Jon Butterworth, UCL52

53. 29/8/2014Jon Butterworth, UCL53 Inaccessible. Removed by kinematics cuts, and not part of the fiducial cross section

54. Concept of a “fiducial” cross section Defines a region in which acceptance is ~100% Implies that some kinematic cuts must be implemented in whatever theory is compared to (easy for MC, less so for some high-order calculations) Ideally of course, build an experiment which covers all the phase space of interest… 29/8/2014Jon Butterworth, UCL54

55. 29/8/2014Jon Butterworth, UCL55

56. 29/8/2014Jon Butterworth, UCL56

57. Concept of a “fiducial” cross section Defines a region in which acceptance is ~100% Implies that some kinematic cuts must be implemented in whatever theory is compared to (easy for MC, less so for some high-order calculations) Ideally of course, build an experiment which covers all the phase space of interest… Fiducial cross defined in terms of the “ideal” or “true” final state 29/8/2014Jon Butterworth, UCL57

58. Real example: ATLAS WW cross section (to e,  ), 7 TeV Efficiency/detector corrections to obtain fiducial cross section: 0.4-0.7 (defined in terms of lepton kinematics) Acceptance (accessible phase space compared to include WW): 0.07-0.16 That missing 90% is stuff we don’t measure The efficiency/detector efficiency won’t change much at 13 TeV, but the acceptance may well drop further 29/8/2014Jon Butterworth, UCL58

59. For example… Top cross section Total cross section measurements extrapolate to 4 , 4 TeV>p T >0 Often not even possible to extract the acceptance from the papers (convoluted with efficiencies and migrations) Means for some, non-trivially-different, regions of phase space, we are just buying the theory Will be even more of a problem at higher beam energies. 29/8/2014Jon Butterworth, UCL59

60. Garbage in, garbage out 29/8/2014Jon Butterworth, UCL60

61. THEORY in, THEORY out 29/8/2014Jon Butterworth, UCL61

62. THEORY in, THEORY out 29/8/2014Jon Butterworth, UCL62 … so the experiment is a waste of time

63. Important considerations (for searches too) Will be residual model dependence in any measurement, but experimentalists should minimise it, make assumptions clear, and control it as far possible within some given systematic uncertainty Theorists should look for measurements of physical final states and where possible calculate them Of course we all want to learn about the underlying theory by interpreting what we see, but try to keep interpretation separate from measurement: interpretations can, and do, change, but we won’t build an LHC (Tevatron, LEP, HERA…) again.

64. Important considerations (for searches too) Will be residual model dependence in any measurement, but experimentalists should minimise it, make assumptions clear, and control it as far possible within some given systematic uncertainty Theorists should look for measurements of physical final states and where possible calculate them Of course we all want to learn about the underlying theory by interpreting what we see, but try to keep interpretation separate from measurement: interpretations can, and do, change, but we won’t build an LHC (Tevatron, LEP, HERA…) again.

65. Important considerations (for searches too) Will be residual model dependence in any measurement, but experimentalists should minimise it, make assumptions clear, and control it as far possible within some given systematic uncertainty Theorists should look for measurements of physical final states and where possible calculate them Of course we all want to learn about the underlying theory by interpreting what we see, but try to keep interpretation separate from measurement: interpretations can, and do, change, but we won’t build an LHC (Tevatron, LEP, HERA…) again.

66. Important considerations (for searches too) What is your final state? quarks, gluons? (top?) W, Z, H? Hadrons? (lifetime cut?) Photons? Taus? Electrons, muons (what about FSR?) Jets (what are the input objects?) Neutrinos? All of them? Missing E T