Newer
Older
Presentations / tau23mu_lhcb / 09_10_2014 / mchrzasz.tex
@Marcin Chrzaszcz Marcin Chrzaszcz on 3 Mar 2015 27 KB massive update
  1. \documentclass[xcolor=svgnames]{beamer}
  2. \usepackage[utf8]{inputenc}
  3. \usepackage[english]{babel}
  4. \usepackage{polski}
  5. %\usepackage{amssymb,amsmath}
  6. %\usepackage[latin1]{inputenc}
  7. %\usepackage{amsmath}
  8. %\newcommand\abs[1]{\left|#1\right|}
  9. \usepackage{amsmath}
  10. \newcommand\abs[1]{\left|#1\right|}
  11. \usepackage{hepnicenames}
  12. \usepackage{hepunits}
  13. \usepackage{color}
  14. \usepackage{feynmp}
  15. \usepackage{pst-pdf}
  16. \usepackage{hyperref}
  17.  
  18.  
  19.  
  20. \setbeamertemplate{footline}{\insertframenumber/\inserttotalframenumber}
  21. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5
  22. \definecolor{mygreen}{cmyk}{0.82,0.11,1,0.25}
  23.  
  24.  
  25. \usetheme{Sybila}
  26. \title[$\Ptau \to \Pmu \Pmu \Pmu$ approval presentation ]{$\Ptau \to \Pmu \Pmu \Pmu$ approval presentation}
  27.  
  28. \author[Paul Seyfert]{
  29. Johannes Albrecht\inst{1}, Marta Calvi\inst{2}, \underline{Marcin Chrzaszcz\inst{3,4}}, \\Laura Gavardi\inst{1}, Jon Harrison\inst{5}, Basem Khanji\inst{2}, George Lafferty\inst{5}, Tatiana Likhomanenko\inst{6}, Eduardo Rodrigues\inst{5}, \\Nicola Serra\inst{3}, Paul Seyfert\inst{7}\\{~}\\
  30. \textbf{Referees}:\\ Benoit Viaud (Chair) , Matteo Rama, Frederic Machefert (EB) }
  31. \institute[Uni Heidelberg]{
  32. \inst{1}Dortmund,
  33. \inst{2}Milano Bicocca,
  34. \inst{3}Zurich, \inst{4}IFJ Cracow, \inst{5}Manchester, \inst{6}Yandex, \inst{7}Uni. Heidelberg
  35. }
  36. \date{\today}
  37. \begin{document}
  38. % --------------------------- SLIDE --------------------------------------------
  39. \frame[plain]{\titlepage}
  40. \author{Marcin Chrz\k{a}szcz}
  41. % ------------------------------------------------------------------------------
  42. % --------------------------- SLIDE --------------------------------------------
  43.  
  44. \institute{~(UZH, IFJ)}
  45.  
  46.  
  47. % \begin{frame}\frametitle{Outline}
  48. % \begin{enumerate}
  49. % \item introduction\vspace{.5em}
  50. % \item multivariate technique\vspace{.5em}
  51. % \item normalisation\vspace{.5em}
  52. % % \item backgrounds\vspace{.5em}
  53. % \item expected sensitivity\vspace{.5em}
  54. % \item model dependence\vspace{.5em} data from Reco14Stripping20(r1)
  55. % \end{enumerate}
  56. % Major news wrt.\ the $1~fb^{-1}$ analysis are highlighted in \textcolor{mygreen}{green}
  57. % \end{frame}
  58.  
  59. \begin{frame}\frametitle{Outline}
  60. \tableofcontents
  61. \end{frame}
  62.  
  63.  
  64.  
  65.  
  66. \begin{frame}
  67. \frametitle{Yellow pages}
  68. \begin{itemize}
  69. \item TWiki: \href{https://twiki.cern.ch/twiki/bin/viewauth/LHCbPhysics/Tau_LFV_3fb}{\url{https://twiki.cern.ch/twiki/bin/viewauth/LHCbPhysics/Tau_LFV_3fb}}
  70. \item ANA note: \href{https://twiki.cern.ch/twiki/pub/LHCbPhysics/Tau_LFV_3fb/v8.pdf}{LHCb-ANA-2014-005}
  71. \item Paper draft: \href{https://twiki.cern.ch/twiki/pub/LHCbPhysics/Tau_LFV_3fb/paper_v1.pdf}{LHCb-PAPER-2014-X}
  72. \item Target journal: JHEP
  73. \item Conference: Tau 2014
  74. \end{itemize}
  75. \end{frame}
  76.  
  77.  
  78.  
  79.  
  80. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  81. \section{Introduction}
  82.  
  83. \begin{frame}
  84. \frametitle{Status of $\Ptau\to\Pmu\Pmu\Pmu$}
  85. \begin{columns}
  86. \begin{column}{.62\textwidth}
  87.  
  88. \includegraphics[width=.95\textwidth]{feymn.png}
  89.  
  90. {{
  91. \begin{itemize}
  92. \item Charged Lepton Flavour Violation process
  93. \item Possible as penguin with neutrino oscillation
  94. \item SM prediction is beyond experimental reach~$O(10^{-40})$
  95. \end{itemize}
  96. }}
  97. \end{column}
  98. \begin{column}{.45\textwidth}
  99. \begin{alertblock}{current limits ($ 90\,\%$ CL)}
  100.  
  101. \begin{description}
  102. \item[BaBar] $3.3\times 10^{-8}$
  103. \item[Belle] $2.1\times 10^{-8}$
  104. \item[LHCb] $8.0\times 10^{-8}$
  105. \end{description}
  106. \end{alertblock}
  107. \begin{alertblock}{BSM predictions}
  108. \begin{description}
  109. \item[var.\ SUSY] $10^{-10}$
  110. \item[non universal $\PZprime$] $10^{-8}$
  111. \item[mSUGRA+seesaw] $10^{-9}$
  112. \item[and many more...]
  113. \end{description}
  114. \end{alertblock}
  115. \end{column}
  116. \end{columns}
  117. \end{frame}
  118. %%%%%%%%%%%%%%%%%
  119. \begin{frame}
  120. \frametitle{Strategy}
  121. \begin{itemize}
  122. \item Following same approach as other RD searches.
  123. \item Loose stripping selection
  124. \item Multivariate classification in: mass, PID, ``geometry/topology''
  125. \item Binning optimisation.
  126. \item Relative normalisation ($\PDs\to\Pphi(\Pmu\Pmu)\Ppi$)
  127. \item Invariant mass fit for expected background in each likelihood bin: fit in $\left| m-m_{\Ptau} \right| >\unit{30}{\MeV}$
  128. \item ``middle sidebands'' for classifier evaluation and tests.($\unit{20}{\MeV}<\left| m-m_{\Ptau}\right| <\unit{30}{\MeV}$).
  129. \item CLs for limit calculation
  130. \end{itemize}
  131. \end{frame}
  132.  
  133. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  134. \begin{frame}
  135. \frametitle{$\Ptau$ production}
  136. \begin{itemize}
  137. \item Consider five production channels (fractions at $\unit{8}{\TeV}$):\begin{itemize}
  138. \item Prompt $\PDs\to\Ptau$ ($72.4\pm2.7\,\%$)
  139. \item Prompt $\PDplus\to\Ptau$ ($4.2\pm0.7\,\%$)
  140. \item Non-prompt $\PDs\to\Ptau$ ($8.5\pm1.7\,\%$)
  141. \item Non-prompt $\PDplus\to\Ptau$ ($0.17\pm0.04\,\%$)
  142. \item $X_{\Pbottom}\to\Ptau$ (meson or baryon) ($14.7\pm2.3\,\%$)
  143. \end{itemize}
  144. \item Use $\sigma(\Pbottom\APbottom)$ at $\unit{8}{\TeV}$ from LHCb
  145. \item Use Pythia scaling for $\sigma(\Pcharm\APcharm)$ at $\unit{8}{\TeV}$
  146. \end{itemize}
  147. \begin{columns}
  148. \column{0.8\textwidth}
  149. \begin{exampleblock}{$\mathcal{B}(\PDplus\to\Ptau)$}
  150. \begin{itemize}
  151. \item There is no measurement of $\mathcal{B}(\PDplus\to\Ptau)$
  152. \item One can calculate it form: $\mathcal{B}(\PDplus\to\Pmu\Pnum)$ + helicity suppression + phase space
  153. \item \texttt{hep-ex:0604043}
  154. \item $\mathcal{B}(\PDplus\to\Ptau\Pnut)=1.0\times10^{-3}$
  155. \end{itemize}
  156. \end{exampleblock}
  157. \column{0.2\textwidth}
  158. {~}
  159. \end{columns}
  160. \end{frame}
  161. \begin{frame}
  162. \frametitle{Datasets}
  163. \begin{itemize}
  164. \item Data from Reco14Stripping20(r1)
  165. \item Much MC\begin{itemize}
  166. \item 24M Inclusive background events ($\Pbottom\APbottom$ and $\Pcharm\APcharm$)
  167. \item 10M Exclusive background events ($\PDs\to\Peta(\Pmu\Pmu\Pphoton)\Pmu\Pnum$)
  168. \item 2M Signal events (split over 5 production channels)
  169. \item 12M $\PD \to \PK \Ppi \Ppi$ (missID studies)
  170. \item 10M $\PDstar \to \PD(\PK \Pmu \Pnum) \Ppi$ (missID studies)
  171. \end{itemize}
  172. \item[$\Rightarrow$] Generator level cuts for improved use of computing resources
  173. \begin{itemize}
  174. \item $\sim 14$ times more signal statistics after stripping
  175. \item $\sim 2$ times more background statistics
  176. \end{itemize}
  177. \item Mix $\Ptau$ production on ntuple level instead of reweighting.
  178. \newline$\Rightarrow$ Ease up ntuple usage (no forgotten weighting, no double weighting, \dots)
  179. \end{itemize}
  180. \end{frame}
  181. \section{Selection}
  182. \begin{frame}
  183. \frametitle{Stripping and selection}
  184. {\footnotesize{
  185. \begin{tabular}{|c|cc|}
  186. \hline
  187. &$\Ptau\to\Pmu\Pmu\Pmu$&$\PDs\to\Pphi\Ppi$\\
  188. \hline
  189. $\mu^\pm$ , $ \pi^\pm$ &\multicolumn{2}{c|}{} \\
  190. $p_T$ &\multicolumn{2}{c|}{$>300\MeV$} \\
  191. Track $\chi^2$/ndf &\multicolumn{2}{c|}{$<3 $} \\
  192. IP $\chi^2$/ndf &\multicolumn{2}{c|}{$>9 $} \\
  193. track ghost probability &\multicolumn{2}{c|}{$<0.3 $} \\
  194. \hline
  195. $\mu$ pairs &\multicolumn{2}{c|}{} \\
  196. $m_{\mu^+\mu^-} - m_{\phi}$ & $>20\MeV$ & $<20\MeV$\\
  197. $m_{\mu^+\mu^-}$ & $> 450\MeV$ & - \\
  198. $m_{\mu^+\mu^+}$ & $> 250\MeV$ & - \\
  199. \hline
  200. $\tau^\pm$ and \PDs &\multicolumn{2}{c|}{} \\
  201. $\Delta m$ & $<400\MeV$ & $<50\MeV$\\
  202. Vertex $\chi^2$ &\multicolumn{2}{c|}{$<15$} \\
  203. IP $\chi^2$ &\multicolumn{2}{c|}{$<225 $} \\
  204. $\cos\alpha$ &\multicolumn{2}{c|}{$>0.99$} \\
  205. $c\tau$ (stripping) &\multicolumn{2}{c|}{$>\unit{100}{\mu m}$} \\
  206. &\multicolumn{2}{c|}{no PV refitting}\\
  207. decay time (offline) &\multicolumn{2}{c|}{$> -0.01$ ns \& $< 0.025$ ns}\\
  208. &\multicolumn{2}{c|}{PV refitting}\\
  209. \hline
  210. \end{tabular}
  211. }}
  212.  
  213. \end{frame}
  214. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  215. \begin{frame}
  216. \frametitle{Triggers}
  217. {\footnotesize{
  218. \begin{tabular}{l|c|c}
  219. & signal & normalisation \\\hline\hline
  220. L0$^1$ & \multicolumn{2}{c}{L0Muon TOS}\\\hline
  221. Hlt1$^1$ & \multicolumn{2}{c}{Hlt1TrackMuon TOS}\\\hline
  222. Hlt2 2011 & Hlt2CharmSemilepD2HMuMu TOS & Hlt2DiMuonDetached$^2$ TOS \\
  223. & $||$ Hlt2TriMuonTau TOS & \\\hline
  224. Hlt2 2012 & Hlt2TriMuonTau$^1$ TOS & Hlt2DiMuonDetached$^2$ TOS\\\hline
  225. \end{tabular}
  226. }
  227. }
  228. \only<1>{
  229. \begin{block}{$^1$ triggers in 2012}
  230. \begin{itemize}
  231. \item Cuts changed through 2012
  232. \item[$\rightarrow$]emulated two different TCKs for 2012
  233. \item[$\rightarrow$] Found negligible differences
  234. \item Choice of triggers were optimised based on $\dfrac{s}{\sqrt{b}}$ FOM.
  235. \end{itemize}
  236. \end{block}}
  237. % \only<2>{
  238. % \begin{block}{$^2$ word on Hlt2DiMuonDetached}
  239. % \begin{itemize}
  240. % \item keep it simple here
  241. % \item line unchanged in 2012
  242. % \item[$\rightarrow$] choice keeps Hlt2 trigger efficiency stable
  243. % \item $\PDs\to\Pphi\Ppi$ anyhow doesn't behave like $\Ptau\to\Pmu\Pmu\Pmu$ in the TriMuon trigger (requires misidentification)
  244. % \end{itemize}
  245. % \end{block}}
  246. \end{frame}
  247. %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
  248. \section{Multivariate technique}
  249.  
  250. \begin{frame}
  251. \frametitle{Geometric likelihood}
  252. Much work has been put in improving our geometric and kinematic classifier:
  253. \begin{itemize}
  254. \item Classify the displaced 3-body decay properties of a signal candidate
  255. \item Revisit variable choice
  256. \item Revisit classification technique
  257. \item More toolkits tried: MatrixNet, NeuroBayes, TMVA
  258. \item Retune input variables\newline($\PBs\to\Pmu\Pmu$ isolation $\rightarrow$ BDT isolation: CERN-THESIS-2013-259)
  259. \item Blending technique.
  260. \end{itemize}
  261.  
  262. \end{frame}
  263. \begin{frame}
  264. \frametitle{Basic Setup - Step I}
  265. \begin{itemize}
  266. \item Train $1/3$ signal MC against $1/2$ background MC
  267. \item variables \begin{itemize}
  268. \item DOCA
  269. \item Vertex $\chi^2$
  270. \item $\tau$ decay time
  271. \item $\tau$ IP$\chi^2$
  272. \item min.\ $\mu$ IP$\chi^2$
  273. \item $\Ptau$ pointing angle
  274. \item $\tau$ $p_T$
  275. \item max.\ track $\chi^2$
  276. \item $\PBs\to\Pmu\Pmu$ track isolation
  277. \item Cone isolation
  278. \item BDT isolation
  279. \end{itemize}
  280. \item Using these variables, train several classifiers (''Base'') \\for each of the $\Ptau$ source
  281. \end{itemize}
  282. \end{frame}
  283.  
  284. \begin{frame}
  285. \frametitle{Step II}
  286. \begin{itemize}
  287. \item Train using second $1/3$ signal MC against second $1/2$ background MC.
  288. \item Introduce Blending technique
  289. \end{itemize}
  290. \begin{exampleblock}{Blending technique}
  291. \begin{itemize}
  292. \item For each signal channel we train: one BDT, three Fisher classifier, four MLPs, one FDA, one LD classifier and MatrixNet classifier.
  293. \item One final MatrixNet classifier using the 13 base variables and the base classifiers as input
  294. \end{itemize}
  295. \end{exampleblock}
  296. \begin{itemize}
  297. \item All evaluation is done on $3rd$ $1/3$ signal sample and middle side-bands.
  298. \item Splitting into independent samples makes the procedure insensitive to
  299. overtraining.
  300. \end{itemize}
  301. \end{frame}
  302. \begin{frame}
  303. \frametitle{Performance of Blend classifier}
  304. \begin{itemize}
  305. \item Classifier prefers $\Ptau$ from prompt $\PDs$
  306. \end{itemize}
  307. \begin{columns}
  308. \begin{column}{.48\textwidth}
  309. \begin{exampleblock}{MC response for different\newline $\Ptau$ production channels}
  310. \includegraphics[width=.95\textwidth]{./mn_sig.png}
  311. \end{exampleblock}
  312. \end{column}
  313. \begin{column}{.48\textwidth}
  314. \begin{exampleblock}{response for $\PDs\to\Pphi\Ppi$\newline data and MC}
  315. \includegraphics[width=.95\textwidth]{./MN_BLEND_FLAT.png}
  316. \end{exampleblock}
  317. \end{column}
  318. \end{columns}
  319. \end{frame}
  320.  
  321. \begin{frame}
  322. \frametitle{Calibration}
  323. \begin{itemize}
  324. \item Assume all differences between $\Ptau\to\Pmu\Pmu\Pmu$ and $\PDs\to\Pphi\Ppi$ come from kinematics (mass, resonance, decay time)
  325. \item Get correction $\PDs\leadsto\Ptau$ from MC
  326. \item Apply corrections to $\PDs\to\Pphi\Ppi$ on data
  327. \end{itemize}
  328. \begin{block}{validation}
  329. \begin{itemize}
  330. \item done for 2011 analysis, treating smeared MC as data
  331. \end{itemize}
  332. \end{block}
  333. \begin{columns}
  334. \begin{column}{.45\textwidth}
  335. \begin{itemize}
  336. \item $\PDs\to\Pphi\Ppi$ well modelled in MC
  337. % \item[$\rightarrow$] i.e.\ also badly pointing non-prompt $\PDs$
  338. \end{itemize}
  339. \end{column}
  340. \begin{column}{.45\textwidth}
  341. \includegraphics[width=.95\textwidth]{MN_BLEND_FLAT.png}
  342. \end{column}
  343. \end{columns}
  344. \end{frame}
  345.  
  346. \begin{frame}
  347. \frametitle{PID}
  348. \begin{itemize}
  349. \item We used ProbNNmu already in the previous round of the analysis
  350. \item Now use MC12TuneV2 (latest)
  351. \item Two-fold reason:\begin{itemize}
  352. \item Expect better performance than CombDLL variables
  353. \item ``one variable for everything'':\newline with CombDLL we needed both CombDLL($\mu-\pi$) and CombDLL($\mu-K$)
  354. \end{itemize}
  355. \end{itemize}
  356. \end{frame}
  357.  
  358. \begin{frame}
  359. \frametitle{PID calibration }
  360. \begin{itemize}
  361. \item Since PIDCalib tool didn't work for us, we choose a phenomenological approach.
  362. \item Many thanks to Barbara Sciascia for help understanding this problem.
  363. \end{itemize}
  364. \begin{exampleblock}{phenomenologic treatment}
  365. \begin{itemize}
  366. \item correlations are small in $\PDs\to\Pphi\Ppi$ data and MC
  367. \newline $\varepsilon(\text{cut on one muon})^2 = \varepsilon(\text{cut on two muons})$
  368. \item[$\Rightarrow$] use $c^3=(\varepsilon(\text{cut and fit})/\varepsilon(\text{PIDCalib}))^3$ as correction to PIDCalib for $\Ptau\to\Pmu\Pmu\Pmu$
  369. \item assign error of $0.02$ for $c$
  370. \end{itemize}
  371. \end{exampleblock}
  372. \begin{itemize}
  373. \item Many x-checks done.
  374. \item Everything works fine.
  375. \end{itemize}
  376.  
  377. \end{frame}
  378.  
  379.  
  380.  
  381. \begin{frame}
  382. \frametitle{Binning optimisation}
  383. \begin{itemize}
  384. \item How to optimise the binning in two classifiers?
  385. \item $\unit{1}{\reciprocal\femtobarn}$ CONF note: two one-dimensional optimisations as in $\PBs\to\Pmu\Pmu$
  386. \item $\unit{1}{\reciprocal\femtobarn}$ PAPER: iterative loop of one-dimensional optimisations\newline optimising one classifier on the sensitive range of the other classifier
  387. \item Now: optimise two-dimensions (optimise bin boundaries in both dimensions at the same time)
  388. \item Unchanged: don't use lowest likelihood bins\newline(reflection backgrounds, no sensitivity gain)
  389. \end{itemize}
  390. \end{frame}
  391. \begin{frame}
  392. \frametitle{Impact of new binning optimisation}
  393. \begin{itemize}
  394. \item Removal of tiny bins which contribute negligible sensitivity
  395. \item Colour: limit obtained, using only this particular bin
  396. \item Number: rank of that bin (1=best sensitivity bin)
  397. \end{itemize}
  398. ~
  399.  
  400. \begin{columns}
  401. \begin{column}{.5\textwidth}
  402. old analysis
  403.  
  404. ~
  405.  
  406. \includegraphics[width=.95\textwidth]{./90CLonebinlimit.pdf}
  407. \end{column}
  408. \begin{column}{.5\textwidth}
  409. new analysis
  410. (2011 data)
  411. \includegraphics[width=.95\textwidth]{./rank.pdf}
  412. \end{column}
  413. \end{columns}
  414. \end{frame}
  415.  
  416. \begin{frame}
  417. \frametitle{Mass shape}
  418. \begin{itemize}
  419. \item Double-Gaussian with fixed fraction ($70\,\%$ inner Gaussian)
  420. \item Fix fraction to ease calibration:\newline
  421. $\sigma_{data}^{\Ptau} = \frac{\sigma_{MC}^{\Ptau}}{\sigma_{MC}^{\PDs}}\times\sigma_{data}^{\PDs}$
  422. \end{itemize}
  423. \includegraphics[width=.44\textwidth]{./Ds_data_2011.pdf}
  424. \includegraphics[width=.44\textwidth]{./Ds_data_2012.pdf}
  425.  
  426. {\footnotesize{
  427. \begin{tabular}{|c|c|c|}
  428. \hline
  429. Calibrated $\Ptau$ Mass shape & 7~TeV & 8~TeV\\
  430. \hline
  431. Mean ($\MeV$) & $1779.1 \pm 0.1$ & $1779.0 \pm 0.1$\\
  432. \hline
  433. $\sigma_1$ ($\MeV$) & $7.7 \pm 0.1$ & $7.6 \pm 0.1$\\
  434. \hline
  435. $\sigma_2$ ($\MeV$) & $12.0 \pm 0.8$ & $11.5 \pm 0.5$\\
  436. \hline
  437. \end{tabular}
  438. }
  439. }
  440. \end{frame}
  441.  
  442.  
  443. \section{Normalisation}
  444.  
  445. \begin{frame}
  446. \frametitle{Relative normalisation}
  447. $\mathcal{B}(\Ptau\to\Pmu\Pmu\Pmu) = \frac{\mathcal{B}(\PDs\to\Pphi\Ppi)}{\mathcal{B}(\PDs\to\Ptau\Pnut)} \times f_{\PDs}^{\Ptau} \times \frac{\varepsilon_\text{norm} }{\varepsilon_\text{sig} } \times \frac{N_\text{sig}}{N_\text{norm}} = \alpha\times N_\text{sig}$
  448. \begin{itemize}
  449. \item where $\varepsilon$ stands for trigger, reconstruction, selection
  450. \item $\text{norm}$ = normalisation channel $\PDs\to\Pphi\Ppi$
  451. \item $f_{\PDs}^{\Ptau}$ is the fraction of $\Ptau$ coming from $\PDs$
  452. \newline i.e.\ $(83\pm3)\,\%$ for 2012
  453. \end{itemize}
  454. \includegraphics[width=.47\textwidth]{./Ds_data_2011.pdf}
  455. \includegraphics[width=.47\textwidth]{./Ds_data_2012.pdf}
  456. \end{frame}
  457.  
  458. \begin{frame}[allowframebreaks]
  459. \frametitle{Normalisation in numbers}
  460. {\footnotesize{
  461. $\begin{array}{c|c|c}
  462. & \rm{7~TeV} & \rm{8~TeV}\\
  463. \hline
  464. \rm{\epsilon\mathstrut_{sig}}^{GEN} (\%) & 8.989 \pm 0.40 & 9.21 \pm 0.35\\
  465. \hline
  466. \rm{\epsilon_{cal}}^{GEN} (\%) & 11.19 \pm 0.34 & 11.53 \pm 0.32\\
  467. \hline
  468. \rm{\epsilon_{sig}}^{REC,isMuon,SEL} (\%) & 9.927 \pm 0.028 & 9.261 \pm 0.023 \\
  469. \hline
  470. \rm{\epsilon_{cal}}^{REC,isMuon,SEL} (\%) & 7.187 \pm 0.022 & 6.690 \pm 0.022 \\
  471. \hline
  472. \frac{\rm{c_{cal}}^{track}}{\rm{c_{sig}}^{track}} & 0.997 \pm 0.009 \pm 0.026 & 0.996 \pm 0.009 \pm 0.026 \\
  473. \hline
  474. \frac{\rm{c_{cal}}^{\mu ID}}{\rm{c_{sig}}^{\mu ID}} & 0.9731 \pm 0.0031 \pm 0.0264 & 1.0071 \pm 0.0022 \pm 0.0204 \\
  475. \hline
  476. \rm{c}^{\phi} & \multicolumn{2}{c}{0.98 \pm 0.01} \\
  477. \hline
  478. \rm{c}^{\tau} & 1.032 \pm 0.006 & 1.026 \pm 0.006\\
  479. \hline
  480. \rm{c}^{trash} & 1.89 \pm 0.12 & 1.96 \pm 0.12\\
  481. \hline
  482. \rm{\epsilon\mathstrut_{sig}}^{TRIG} (\%) & 35.52 \pm 0.14 \pm 0.14 & 39.3 \pm 1.7 \pm 2.0 \\
  483. \hline
  484. \rm{\epsilon\mathstrut_{cal}}^{TRIG} (\%) & 23.42 \pm 0.14 \pm 0.09 & 20.62 \pm 0.76 \pm 1.07 \\
  485. \end{array}$
  486. }}
  487.  
  488. \framebreak
  489.  
  490. {\footnotesize{
  491. $\begin{array}{c|c|c}
  492. & \rm{7~TeV} & \rm{8~TeV}\\
  493. \hline
  494. \mathcal{B}(PDs \to \Pphi \Ppi) & \multicolumn{2}{c}{(1.317 \pm 0.099) \times 10^{-5}}\\
  495. \hline
  496. f^{\tau}_{D_{s}} & 0.78 \pm 0.04 & 0.80 \pm 0.03 \\
  497. \hline
  498. \mathcal{B} (\PDs \to \Ptau \Pnut) & \multicolumn{2}{c}{0.0561 \pm 0.0024}\\
  499. \hline
  500. \rm{\epsilon\mathstrut_{cal}}^{REC\&SEL}/
  501. \rm{\epsilon\mathstrut_{sig}}^{REC\&SEL}
  502. & 0.898 \pm 0.060 & 0.912 \pm 0.054 \\
  503. \hline
  504. \rm{\epsilon\mathstrut_{cal}}^{TRIG}/
  505. \rm{\epsilon\mathstrut_{sig}}^{TRIG}
  506. & 0.6593 \pm 0.0058 & 0.525 \pm 0.040\\
  507. \hline
  508. N_{cal} & 28,207 \pm 440 & 52,131 \pm 695\\
  509. \hline & \\[-0.8em]\hline
  510. \alpha & (3.81 \pm 0.46) \times 10^{-9} & (1.72 \pm 0.23) \times 10^{-9}\\
  511. \alpha^{trash} & (7.20 \pm 0.98) \times 10^{-9} & (3.37 \pm 0.50) \times 10^{-9}\\
  512. \end{array}$
  513. }}
  514. \end{frame}
  515.  
  516.  
  517. \section{Backgrounds}
  518.  
  519. \begin{frame}
  520. \frametitle{Misidentification}
  521. \begin{itemize}
  522. \item Most dominant: $\PDplus\to\PK\Ppi\Ppi$
  523. \item Experience from last round: cut away low ProbNNmu range
  524. \item Check remaining data under $\PK\Ppi\Ppi$ hypothesis for $\PDplus$ peak
  525. \item[$\Rightarrow$] misid safely contained in ``trash'' bin
  526. \item $\PDplus\to\Ppi\Ppi\Ppi$ and $\PDs\to\Ppi\Ppi\Ppi$ start to become visible in 2012
  527. \end{itemize}
  528. \includegraphics[width=.45\textwidth]{./Dp2Kpipi_all_2012_senseBins.pdf}
  529. \includegraphics[width=.45\textwidth]{./FittoD23pi_2012.pdf}
  530. \end{frame}
  531.  
  532. \begin{frame}
  533. \frametitle{Evil backgrounds}
  534. \begin{itemize}
  535. \item $\Pphi\to\Pmu\Pmu + X$: narrow veto on dimuon mass
  536. \item $\PDs\to\Peta(\Pmu\Pmu\Pphoton)\Pmu\Pnum$: not so easy
  537. \begin{itemize}
  538. \item Modelled in CONF note
  539. \item Optimised veto in PAPER
  540. \item Right now: both versions in the ANA note
  541. \end{itemize}
  542. \item Baseline: veto $m_{\APmuon\Pmuon} < \unit{450}{\MeV}$
  543. \begin{itemize}
  544. \item Fits better understood
  545. \item Sensitivity unchanged when removing veto
  546. \item Smaller uncertainty on expected background
  547. \end{itemize}
  548. \end{itemize}
  549. \end{frame}
  550.  
  551. \begin{frame}
  552. \frametitle{Remaining backgrounds}
  553. \begin{itemize}
  554. \item Fit exponential to invariant mass spectrum in each likelihood bin
  555. \item Don't use $\pm \unit{30}{\MeV}$ in the fit
  556. \item[$\rightarrow$] Compatible results blinding only $\pm \unit{20}{\MeV}$\footnote{partially used in classifier developement}
  557. \end{itemize}
  558. {\begin{center}
  559. Example of most sensitive regions in 2011 and 2012
  560. \includegraphics[width=0.9\textwidth]{./fits.png}
  561.  
  562. \end{center}}
  563. \end{frame}
  564.  
  565. \section{Expected limit}
  566.  
  567. \begin{frame}
  568. \frametitle{expected limit}
  569. \begin{itemize}
  570. \item Consider nuisance parameters from background fit, signal pdf calibration, normalisation
  571. \item Nuisance parameters due to $\Ptau$ production, normalization.
  572. \item Limit for combined 2011+2012 analysis
  573. \end{itemize}
  574. \end{frame}
  575.  
  576. \begin{frame}
  577. \frametitle{Sensitivity}
  578. $\mathcal{B}(\Ptau\to\Pmu\Pmu\Pmu)<5.0 \times 10^{-8}$ at 90\% CL
  579.  
  580. \includegraphics[width=.8\textwidth]{./banana.png}
  581. \end{frame}
  582.  
  583.  
  584.  
  585. \section{Model dependence}
  586.  
  587. \begin{frame}
  588. \frametitle{Model dependence}
  589. \begin{itemize}
  590. \item $\Peta$ veto $\Rightarrow$ our limit not applicable to New Physics with small $m_{\APmuon\Pmuon}$
  591. \item Model description in \texttt{arXiv:0707.0988}
  592. \item 5 relevant Dalitz distributions: 2 four-point operators, 1 radiative operator, 2 interference terms
  593. \end{itemize}
  594. \only<2->{
  595. \begin{itemize}
  596. \item With radiative distribution limit gets worse by $51\,\%$ (dominantly from the $\Peta$ veto)
  597. \item The other four Dalitz distributions behave nicely (within $7\,\%$)
  598. \end{itemize}
  599. }
  600. \only<1>{
  601. \includegraphics[width=.331\textwidth]{./gammallll.pdf}
  602. \includegraphics[width=.331\textwidth]{./gammallrr.pdf}
  603. \includegraphics[width=.331\textwidth]{./gammarad.pdf}
  604.  
  605. \includegraphics[width=.331\textwidth]{./gammarad-llll.pdf}
  606. \includegraphics[width=.331\textwidth]{./gammarad-llrr.pdf}
  607. }
  608.  
  609. \end{frame}
  610.  
  611.  
  612. % \begin{frame}
  613. % \frametitle{Conclusion}
  614. % \begin{columns}
  615. % \begin{column}{.55\textwidth}
  616. % \begin{itemize}
  617. % \item finally all pieces put together
  618. % \item model (in)dependence of $\Peta$ veto investigated
  619. % \item expected sensitivity computed\newline $5.6\times 10^{-8}$
  620. % \end{itemize}
  621. % \end{column}
  622. % \begin{column}{.45\textwidth}
  623. % \includegraphics[width=\textwidth]{party-music-hd-wallpaper-1920x1200-3850.jpg}
  624. % \end{column}
  625. % \end{columns}
  626.  
  627. % \end{frame}
  628.  
  629.  
  630. \section{Unblinded results}
  631.  
  632. \begin{frame}
  633. \frametitle{Unblinding 1}
  634. \begin{columns}
  635. \column{1in}{~}
  636. \column{3in}
  637. ''
  638. THERE came a day at summers full \\
  639. Entirely for us \\
  640. I thought that such were for the saints, \\
  641. Where revelations be. ''\footnote{E.Dickinson} \\
  642. \column{1in}{~}
  643. \end{columns}
  644. {~}\\
  645. {~}\\
  646. \begin{Large}
  647. On Monday $4^{th}$ of August we were given the permission to unblind.
  648. \end{Large}
  649. \end{frame}
  650. \begin{frame}
  651. \frametitle{Unblinding 2}
  652. \begin{itemize}
  653. \item Unfortunately no big ''revelations'' were there:
  654. \item 2011 numbers:
  655. \end{itemize}
  656. \includegraphics[width=1.\textwidth]{2011.png}
  657. \end{frame}
  658. \begin{frame}
  659. \frametitle{Unblinding 3}
  660. \begin{itemize}
  661. \item Unfortunately no big ''revelations'' were also in 2012 data:
  662.  
  663. \end{itemize}
  664. \includegraphics[width=1.1\textwidth]{2012.png}
  665. \end{frame}
  666.  
  667. \begin{frame}
  668. \frametitle{Unblinding 4}
  669.  
  670. \begin{center}
  671. \includegraphics[width=0.7\textwidth]{banana_line.pdf}
  672. \end{center}
  673. \begin{columns}
  674.  
  675. \column{0.2in}{~}
  676. \column{2in}
  677. Limits(PHSP):\\
  678. Observed(Expected)\\
  679. $4.6~(5.0)\times 10^{-8}$ at $90\%$ CL\\
  680. $5.6~(6.1)\times 10^{-8}$ at $95\%$ CL\\
  681.  
  682. \column{3in}
  683. \includegraphics[width=0.5\textwidth]{model.png}
  684. \end{columns}
  685. \end{frame}
  686. \begin{frame}
  687. \frametitle{Conclusions}
  688. \begin{columns}
  689. \column{2.5in}
  690. \begin{itemize}
  691. \item We didn't find NP (yet).
  692. \item Limits set with full LHCb dataset.
  693. \item Awaiting for the future data!
  694. \end{itemize}
  695. \column{2.5in}
  696. \includegraphics[width=1\textwidth]{TauLFV_UL_2013001.pdf}
  697.  
  698.  
  699. \end{columns}
  700. \begin{itemize}
  701. \item We would like to thank our referees for very friendly,thorough and fruitful review.
  702. \item With this presentation we ask collaboration for approval.
  703. \end{itemize}
  704.  
  705. \end{frame}
  706. \end{document}