Newer
Older
Presentations / tau23mu_lhcb / APROVAL / tmp.tex
@mchrzasz mchrzasz on 13 Aug 2014 28 KB update
  1. \documentclass[xcolor=dvipsnames,table]{beamer}
  2.  
  3.  
  4.  
  5.  
  6.  
  7. \author[Paul Seyfert]{
  8. Johannes Albrecht\inst{1}, Marta Calvi\inst{2}, Marcin Chrz\k{a}szcz\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{4}, \underline{Paul Seyfert\inst{7}}}
  9.  
  10. \institute[Uni Heidelberg]{
  11. \inst{1}Dortmund,
  12. \inst{2}Milano Bicocca,
  13. \inst{3}Cracow, \inst{4}Zurich, \inst{5}Manchester, \inst{6}Yandex, \inst{7}Heidelberg University
  14. }
  15.  
  16. \date{\today}
  17. \subject{}
  18.  
  19.  
  20. \AtBeginSection[]
  21. {
  22. \begin{frame}<beamer>{}
  23. \tableofcontents[currentsection,currentsubsection]
  24. \end{frame}
  25. }
  26.  
  27.  
  28. \begin{document}
  29. \begin{frame}
  30. \titlepage
  31. \end{frame}
  32.  
  33.  
  34. \begin{frame}
  35. \begin{enumerate}
  36. \item introduction\vspace{.5em}
  37. \item multivariate technique\vspace{.5em}
  38. \item choice of triggers\vspace{.5em}
  39. \item normalisation\vspace{.5em}
  40. \item backgrounds\vspace{.5em}
  41. \item expected sensitivity\vspace{.5em}
  42. \item model dependence\vspace{.5em}
  43. \end{enumerate}
  44. Major news wrt.\ the previous analysis rounds highlighted in \textcolor{darkgreen}{green}
  45. \end{frame}
  46.  
  47. \section{introduction}
  48.  
  49. \begin{frame}
  50. \frametitle{Status $\Ptau\to\Pmu\Pmu\Pmu$}
  51. \begin{columns}
  52. \begin{column}{.62\textwidth}
  53. \begin{fmffile}{sm}
  54. \begin{fmfgraph*}(200,150)
  55. \fmfstraight
  56. \fmfleft{is3,is2,is1,i1,i2,i3,i4,i5,i6}
  57. \fmfright{os3,os2,os1,o1,o2,o3,o4,o5,o6}
  58.  
  59.  
  60. \fmf{phantom}{o1,g1}
  61. \fmf{phantom,tension=1.5}{g1,i1}
  62.  
  63. \fmffreeze
  64.  
  65. \fmf{photon, tension=2.,lab.side=left, lab=$\Pphoton$}{g1,g2}
  66. \fmf{fermion,lab.side=left,label=$\APmuon$}{o3,g2}
  67. \fmf{fermion,lab.side=left,label=$\Pmuon$}{g2,o5}
  68. \fmf{phantom,tension=1.5}{i6,g2}
  69.  
  70. \fmffreeze
  71.  
  72.  
  73. \fmf{fermion,lab.side=left,label=$\Pgtm$}{i1,w1}
  74. \fmf{dashes,lab.side=right,label=$\PWminus$}{w1,w2}
  75. \fmf{fermion,lab.side=left,label=$\Pmuon$}{w2,o1}
  76.  
  77.  
  78. \fmfipath{p[]}
  79. \fmf{plain,tension=0,right,label={$\Pgngt\to\Pgngm$},tag=1}{w1,w2}
  80. \fmffreeze
  81. \fmfiset{p1}{vpath1(__w1,__w2)}
  82. \fmfiv{d.sh=cross,d.ang=0,d.size=5thick}{point length(p1)/2 of p1}
  83.  
  84.  
  85. \end{fmfgraph*}
  86.  
  87. \end{fmffile}
  88. {\footnotesize{
  89. \begin{itemize}
  90. \item charged Lepton Flavour Violation process
  91. \item possible as penguin with neutrino oscillation
  92. \item unmeasurable small
  93. \end{itemize}
  94. }}
  95. \end{column}
  96. \begin{column}{.37\textwidth}
  97. \begin{block}{current limits ($ 90\,\%$ CL)}
  98.  
  99. \begin{description}
  100. \item[BaBar] $3.3\times 10^{-8}$
  101. \item[Belle] $2.1\times 10^{-8}$
  102. \item[LHCb] $8.0\times 10^{-8}$
  103. \end{description}
  104. \end{block}
  105. \begin{block}{BSM predictions}
  106. \begin{description}
  107. \item[var.\ SUSY] $10^{-10}$
  108. \item[non universal $\PZprime$] $10^{-8}$
  109. \item[mSUGRA+seesaw] $10^{-9}$
  110. \end{description}
  111. \end{block}
  112. \end{column}
  113. \end{columns}
  114. \end{frame}
  115.  
  116. \begin{frame}
  117. \frametitle{$\Ptau$ production}
  118. \begin{itemize}
  119. \item consider five production channels (fractions at $\unit{8}{\TeV}$):\begin{itemize}
  120. \item prompt $\PDs\to\Ptau$ ($73.1\pm3.9\,\%$)
  121. \item prompt $\PDplus\to\Ptau$ ($0.42\pm0.43\,\%$)
  122. \item non-prompt $\PDs\to\Ptau$ ($9.7\pm2.1\,\%$)
  123. \item non-prompt $\PDplus\to\Ptau$ ($0.02\pm0.02\,\%$)
  124. \item $X_{\Pbottom}\to\Ptau$ (meson or baryon) ($16.8\pm3.0\,\%$)
  125. \end{itemize}
  126. \item \textcolor{darkgreen}{use $\sigma(\Pbottom\APbottom)$ at $\unit{8}{\TeV}$ from LHCb}
  127. \item \textcolor{darkgreen}{use Pythia scaling for $\sigma(\Pcharm\APcharm)$ at $\unit{8}{\TeV}$}
  128. \end{itemize}
  129. \begin{exampleblock}{$\mathcal{B}(\PDplus\to\Ptau)$}
  130. \begin{itemize}
  131. \item old analysis: used upper limit
  132. \item now: $\mathcal{B}(\PDplus\to\Pmu\Pnum)$ + helicity suppression + phase space
  133. \item \texttt{hep-ex:0604043}
  134. \item $\mathcal{B}(\PDplus\to\Ptau\Pnut)=1\pm1\times10^{-4}$
  135. \end{itemize}
  136. \end{exampleblock}
  137. \end{frame}
  138.  
  139. \begin{frame}
  140. \frametitle{Strategy}
  141. \begin{itemize}
  142. \item mostly as in many rare decay searches:
  143. \item loose stripping selection
  144. \item multivariate classification in: mass, PID, ``geometry/topology''
  145. \item relative normalisation ($\PDs\to\Pphi(\Pmu\Pmu)\Ppi$)
  146. \item invariant mass fit for expected background in each likelihood bin\newline fit in $m-m_{\Ptau}>\unit{30}{\MeV}$
  147. \item ``middle sidebands'' for classifier evaluation ($\unit{20}{\MeV}<m-m_{\Ptau}<\unit{30}{\MeV}$)
  148. \item CLs for limit calculation
  149. \end{itemize}
  150. \begin{block}{today}
  151. \begin{itemize}
  152. \item $\unit{3}{\reciprocal\femtobarn}$ analysis ``as is''
  153. \item nice \& interesting things not included so far e.g.\ 2011 reanalysed
  154. \end{itemize}
  155. \end{block}
  156. \end{frame}
  157.  
  158. \begin{frame}
  159. \frametitle{Datasets}
  160. \begin{itemize}
  161. \item data from \textcolor{darkgreen}{Reco14Stripping20(r1)}
  162. \item much MC\begin{itemize}
  163. \item \textcolor{darkgreen}{24M} inclusive background events ($\Pbottom\APbottom$ and $\Pcharm\APcharm$)
  164. \item \textcolor{darkgreen}{10M} exclusive background events ($\PDs\to\Peta(\Pmu\Pmu\Pphoton)\Pmu\Pnum$)
  165. \item \textcolor{darkgreen}{2M} signal events (split over 5 production channels)
  166. \end{itemize}
  167. \item[$\Rightarrow$] \textcolor{darkgreen}{generator level cuts} for improved use of computing resources
  168. \begin{itemize}
  169. \item \textcolor{darkgreen}{$\sim 14$ times more} signal statistics after stripping
  170. \item \textcolor{darkgreen}{$\sim 2$ times more} background statistics
  171. \end{itemize}
  172. \item \textcolor{darkgreen}{mix $\Ptau$ production on ntuple level} instead of reweighting.
  173. \newline$\Rightarrow$ ease up ntuple usage (no forgotten weighting, no double weighting, \dots)
  174. \end{itemize}
  175. \end{frame}
  176.  
  177. \begin{frame}
  178. \frametitle{(Stripping) selection}
  179. {\footnotesize{
  180. \begin{tabular}{|c|cc|}
  181. \hline
  182. &$\Ptau\to\Pmu\Pmu\Pmu$&$\PDs\to\Pphi\Ppi$\\
  183. \hline
  184. $\mu^\pm$ , $ \pi^\pm$ &\multicolumn{2}{c|}{} \\
  185. $p_T$ &\multicolumn{2}{c|}{$>300\MeV$} \\
  186. Track $\chi^2$/ndf &\multicolumn{2}{c|}{$<3 $} \\
  187. IP $\chi^2$/ndf &\multicolumn{2}{c|}{$>9 $} \\
  188. \textcolor{darkgreen}{track ghost probability} &\multicolumn{2}{c|}{\textcolor{darkgreen}{$<0.3 $}} \\
  189. \hline
  190. $\mu$ pairs &\multicolumn{2}{c|}{} \\
  191. $m_{\mu^+\mu^-} - m_{\phi}$ & $>20\MeV$ & $<20\MeV$\\
  192. $m_{\mu^+\mu^-}$ & $> 450\MeV$ & - \\
  193. $m_{\mu^+\mu^+}$ & $> 250\MeV$ & - \\
  194. \hline
  195. $\tau^\pm$ and \PDs &\multicolumn{2}{c|}{} \\
  196. $\Delta m$ & $<400\MeV$ & $<50\MeV$\\
  197. Vertex $\chi^2$ &\multicolumn{2}{c|}{$<15$} \\
  198. \textcolor{darkgreen}{IP $\chi^2$} &\multicolumn{2}{c|}{$<225^1 $} \\
  199. $\cos\alpha$ &\multicolumn{2}{c|}{$>0.99$} \\
  200. $c\tau$ (stripping) &\multicolumn{2}{c|}{$>\unit{100}{\mu m}$} \\
  201. &\multicolumn{2}{c|}{no PV refitting}\\
  202. decay time (offline) &\multicolumn{2}{c|}{$> -0.01$ ns \& $< 0.025$ ns}\\
  203. &\multicolumn{2}{c|}{PV refitting}\\
  204. \hline
  205. \end{tabular}
  206. }}
  207.  
  208. {\footnotesize{$^1$ different LoKi functor}}
  209. \end{frame}
  210.  
  211. \section{multivariate technique}
  212.  
  213. \begin{frame}
  214. \frametitle{``geometric likelihood''}
  215. \begin{itemize}
  216. \item classify the displaced 3-body decay properties of a signal candidate
  217. \item revisit variable choice
  218. \item revisit classification technique
  219. \item \textcolor{darkgreen}{more toolkits tried: MatrixNet, NeuroBayes}, TMVA
  220. \item \textcolor{darkgreen}{retune input variables\newline($\PBs\to\Pmu\Pmu$ isolation $\rightarrow$ Laura's BDT isolation: CERN-THESIS-2013-259)}
  221. \end{itemize}
  222.  
  223. \end{frame}
  224. \begin{frame}
  225. \frametitle{setup}
  226. \begin{itemize}
  227. \item train $1/3$ signal MC against $1/2$ background MC
  228. \item variables \begin{itemize}
  229. \item $3\times$ DOCA
  230. \item vertex $\chi^2$
  231. \item $\tau$ decay time
  232. \item $\tau$ IP$\chi^2$
  233. \item min.\ $\mu$ IP$\chi^2$
  234. \item $\Ptau$ pointing angle
  235. \item $\tau$ $p_T$
  236. \item max.\ track $\chi^2$
  237. \item $\PBs\to\Pmu\Pmu$ track isolation
  238. \item cone isolation
  239. \item \textcolor{darkgreen}{BDT isolation}
  240. \end{itemize}
  241. \end{itemize}
  242. \end{frame}
  243. \begin{frame}
  244. \frametitle{futher tweaking}
  245. \begin{itemize}
  246. \item \textcolor{darkgreen}{remove fully reconstructed 3-body decays from background sample\newline (don't expect to be able to discriminate these)}
  247. \item don't apply trigger prior to training
  248. \end{itemize}
  249.  
  250. \begin{exampleblock}{``blending'' technique}
  251. \begin{itemize}
  252. \item for each signal channel we train: one BDT, three Fisher classifier, four MLPs, one FDA, and one LD classifier
  253. \item[$\Rightarrow$] 50 classifiers
  254. \item one final MatrixNet classifier using the 13 base variables and the 50 classifiers as input
  255. \newline(trained on the second $1/3$ of signal MC and the second $1/2$ of background MC)
  256. \end{itemize}
  257. \end{exampleblock}
  258. \end{frame}
  259.  
  260. \begin{frame}
  261. \frametitle{performance}
  262. \begin{itemize}
  263. \item classifier prefers $\Ptau$ from prompt $\PDs$
  264. \end{itemize}
  265. \begin{columns}
  266. \begin{column}{.48\textwidth}
  267. \begin{block}{MC response for different\newline $\Ptau$ production channels}
  268. \includegraphics[width=.95\textwidth]{./for_paul.png}
  269. \end{block}
  270. \end{column}
  271. \begin{column}{.48\textwidth}
  272. \begin{block}{response for $\PDs\to\Pphi\Ppi$\newline data and MC}
  273. \includegraphics[width=.95\textwidth]{./MN_BLEND_FLAT.png}
  274. \end{block}
  275. \end{column}
  276. \end{columns}
  277. \end{frame}
  278.  
  279. \begin{frame}
  280. \frametitle{calibration}
  281. \begin{itemize}
  282. \item assume all differences between $\Ptau\to\Pmu\Pmu\Pmu$ and $\PDs\to\Pphi\Ppi$ come from kinematics (mass, resonance, decay time)
  283. \item get correction $\PDs\leadsto\Ptau$ from MC
  284. \item apply corrections to $\PDs\to\Pphi\Ppi$ on data
  285. \end{itemize}
  286. \begin{block}{validation}
  287. \begin{itemize}
  288. \item done for 2011 analysis, treating smeared MC as data
  289. \end{itemize}
  290. \end{block}
  291. \begin{columns}
  292. \begin{column}{.45\textwidth}
  293. \begin{itemize}
  294. \item $\PDs\to\Pphi\Ppi$ well modelled in MC
  295. \item until the very low likelihood end of the distribution
  296. \item[$\rightarrow$] i.e.\ also badly pointing non-prompt $\PDs$
  297. \end{itemize}
  298. \end{column}
  299. \begin{column}{.45\textwidth}
  300. \includegraphics[width=.95\textwidth]{MN_BLEND_FLAT.png}
  301. \end{column}
  302. \end{columns}
  303. \end{frame}
  304.  
  305. \begin{frame}
  306. \frametitle{PID}
  307. \begin{itemize}
  308. \item we used ProbNNmu already in the previous round of the analysis
  309. \item now use MC12TuneV2 (latest)
  310. \item two-fold reason:\begin{itemize}
  311. \item expect better performance than CombDLL variables
  312. \item ``one variable for everything'':\newline with CombDLL we needed both CombDLL($\mu-\pi$) and CombDLL($\mu-K$)
  313. \end{itemize}
  314. \end{itemize}
  315. \end{frame}
  316.  
  317. \begin{frame}
  318. \frametitle{PIDCalib}
  319. \begin{itemize}
  320. \item calibration strategy: use PIDCalib
  321. \item confirm with $\PDs\to\Pphi\Ppi$ if everything is fine
  322. \end{itemize}
  323. \begin{columns}
  324. \begin{column}{.45\textwidth}
  325. cut\&fit:
  326. \begin{itemize}
  327. \item fit $\PDs\to\Pphi\Ppi$ with a TIS muon in data
  328. \item cut on ProbNNmu of one muon
  329. \item fit again
  330. \item[$\rightarrow$] ratio is ``true'' cut efficiency
  331. \end{itemize}
  332. \begin{block}{ProbNNmu>0.4}
  333. $\varepsilon=86.3\,\%$
  334. \end{block}
  335. \end{column}
  336. \begin{column}{.45\textwidth}
  337. PIDCalib
  338. \begin{itemize}
  339. \item apply full selection (incl. trigger) to $\PDs\to\Pphi\Ppi$ MC reference sample
  340. \item avoid IsMuon bias
  341. \end{itemize}
  342. \begin{block}{ProbNNmu>0.4}
  343. $\varepsilon=89.8\,\%$
  344. \end{block}
  345. \end{column}
  346. \end{columns}
  347. \end{frame}
  348.  
  349. \begin{frame}
  350. \frametitle{?}
  351. \begin{itemize}
  352. \item first shown at \myhref{https://indico.cern.ch/event/291727/}{charming VRD meeting}
  353. \item also mentionned at last \myhref{https://indico.cern.ch/event/298021/}{LHCb week}
  354. \item many emails exchanged with Barbara Sciascia
  355. \item mistakes on user side found
  356. \item still no agreement
  357. \end{itemize}
  358. \begin{exampleblock}{phenomenologic treatment}
  359. \begin{itemize}
  360. \item correlations are small in $\PDs\to\Pphi\Ppi$ data and MC
  361. \newline $\varepsilon(\text{cut on one muon})^2 = \varepsilon(\text{cut on two muons})$
  362. \item[$\Rightarrow$] use $c^3=(\varepsilon(\text{cut and fit})/\varepsilon(\text{PIDCalib}))^3$ as correction to PIDCalib for $\Ptau\to\Pmu\Pmu\Pmu$
  363. \item assign error of $0.02$ for $c$
  364. \end{itemize}
  365. \end{exampleblock}
  366. \begin{itemize}
  367. \item planned: investigate further (usage/bug/samples)
  368. \item planned: use muons from $\PDs\to\Pphi\Ppi$ directly
  369. \end{itemize}
  370. \end{frame}
  371.  
  372. \begin{frame}
  373. \frametitle{binning optimisation}
  374. \begin{itemize}
  375. \item how to optimise the binning in two classifiers?
  376. \item $\unit{1}{\reciprocal\femtobarn}$ CONF note: two one-dimensional optimisations as in $\PBs\to\Pmu\Pmu$
  377. \item $\unit{1}{\reciprocal\femtobarn}$ PAPER: iterative loop of one-dimensional optimisations\newline optimising one classifier on the sensitive range of the other classifier
  378. \item \textcolor{darkgreen}{now: optimise two-dimensions (optimise bin boundaries in both dimensions at the same time)}
  379. \item unchanged: don't use lowest likelihood bins\newline(reflection backgrounds, no sensitivity gain)
  380. \end{itemize}
  381. \end{frame}
  382. \begin{frame}
  383. \frametitle{impact of new binning optimisation}
  384. \begin{itemize}
  385. \item removal of tiny bins which contribute negligible sensitivity
  386. \item colour: limit obtained, using only this particular bin
  387. \item number: rank of that bin (1=best sensitivity bin)
  388. \end{itemize}
  389. ~
  390.  
  391. \begin{columns}
  392. \begin{column}{.5\textwidth}
  393. old analysis
  394.  
  395. ~
  396.  
  397. \includegraphics[width=.95\textwidth]{./90CLonebinlimit.eps}
  398. \end{column}
  399. \begin{column}{.5\textwidth}
  400. new analysis
  401. (2011 data, not final calibration)
  402. \includegraphics[width=.95\textwidth]{./rank.eps}
  403. \end{column}
  404. \end{columns}
  405. \end{frame}
  406.  
  407.  
  408. \begin{frame}
  409. \frametitle{mass shape}
  410. \begin{itemize}
  411. \item double-Gaussian with fixed fraction ($70\,\%$ inner Gaussian)
  412. \item fix fraction to ease calibration:\newline
  413. $\sigma_{data}^{\Ptau} = \frac{\sigma_{MC}^{\Ptau}}{\sigma_{MC}^{\PDs}}\times\sigma_{data}^{\PDs}$
  414. \end{itemize}
  415. \includegraphics[width=.44\textwidth]{./Ds_data_2011.pdf}
  416. \includegraphics[width=.44\textwidth]{./Ds_data_2012.pdf}
  417.  
  418. {\footnotesize{
  419. \begin{tabular}{|c|c|c|}
  420. \hline
  421. calibrated $\Ptau$ mass shape & 7~TeV & 8~TeV\\
  422. \hline
  423. Mean ($\MeV$) & $1779.1 \pm 0.1$ & $1779.0 \pm 0.1$\\
  424. \hline
  425. $\sigma_1$ ($\MeV$) & $7.7 \pm 0.1$ & $7.6 \pm 0.1$\\
  426. \hline
  427. $\sigma_2$ ($\MeV$) & $12.0 \pm 0.8$ & $11.5 \pm 0.5$\\
  428. \hline
  429. \end{tabular}
  430. }
  431. }
  432. \end{frame}
  433.  
  434. \beamertemplateshadingbackground{darkgreen!30}{PineGreen!30}
  435. \section{choice of triggers}
  436. \begin{frame}
  437. \frametitle{the story}
  438. \begin{itemize}
  439. \item first look at $\PDs\to\Pphi\Ppi$ revealed: signal/background much worse in 2012
  440. \item[$\rightarrow$] charm got more bandwidth in 2012
  441. \item taking all events quite unsatisfactory:\begin{itemize}
  442. \item strictly speaking: we don't know why the events ended up on tape
  443. \item trigger efficiencies unstable (TIS/TPS/TOS efficiencies for lots of lines and TCKs)
  444. \item signal anyhow mostly TOS in a muon trigger
  445. \end{itemize}
  446. \item[$\Rightarrow$] if an event was background in the trigger, then don't consider it signal afterwards.
  447. \end{itemize}
  448. \end{frame}
  449.  
  450. \begin{frame}
  451. \frametitle{strategy}
  452. \begin{itemize}
  453. \item back of the envelope: limit scales with $1/\varepsilon_\text{sig}$
  454. \item back of the envelope: limit scales with $\sqrt{\varepsilon_\text{bkg}}$
  455. \item[$\Rightarrow$] minimise $FOM=\frac{\sqrt{\varepsilon_\text{bkg}}}{\varepsilon_\text{sig}}$
  456. \item running TCKsh and some combinatorics: $\mathcal{O}(2^{256}\approx 10^{75})$ possibilities of using HLT2 line (ignoring that there are different TCKs, TIS/TOS/TPS/DEC, \dots)
  457. \item[$\rightarrow$] be more pragmatic!
  458. \end{itemize}
  459. \begin{exampleblock}{trigger optimisation}
  460. \begin{itemize}
  461. \item sort trigger lines (Punzi FoM)
  462. \item start with events from the best trigger line and compute $FOM$
  463. \item add events from the next trigger and recompute $FOM$
  464. \item iterate and converge to $FOM(\text{all triggers})$
  465. \item use all triggers until $FOM$ starts rising.
  466. \end{itemize}
  467. \end{exampleblock}
  468. \end{frame}
  469.  
  470. \begin{frame}
  471. \frametitle{our triggers}
  472. {\footnotesize{
  473. \begin{tabular}{l|c|c}
  474. & signal & normalisation \\\hline\hline
  475. L0$^1$ & \multicolumn{2}{c}{L0Muon TOS}\\\hline
  476. Hlt1$^1$ & \multicolumn{2}{c}{Hlt1TrackMuon TOS}\\\hline
  477. Hlt2 2011 & Hlt2CharmSemilepD2HMuMu TOS & Hlt2DiMuonDetached$^2$ TOS \\
  478. & || Hlt2TriMuonTau TOS & \\\hline
  479. Hlt2 2012 & Hlt2TriMuonTau$^1$ TOS & Hlt2DiMuonDetached$^2$ TOS\\\hline
  480. \end{tabular}
  481. }
  482. }
  483. \only<1>{
  484. \begin{block}{$^1$ triggers in 2012}
  485. \begin{itemize}
  486. \item cuts changed through 2012
  487. \item[$\rightarrow$] \textcolor{darkgreen}{emulated two different TCKs for 2012}
  488. \end{itemize}
  489. \end{block}}
  490. \only<2>{
  491. \begin{block}{$^2$ word on Hlt2DiMuonDetached}
  492. \begin{itemize}
  493. \item keep it simple here
  494. \item line unchanged in 2012
  495. \item[$\rightarrow$] choice keeps Hlt2 trigger efficiency stable
  496. \item $\PDs\to\Pphi\Ppi$ anyhow doesn't behave like $\Ptau\to\Pmu\Pmu\Pmu$ in the TriMuon trigger (requires misidentification)
  497. \end{itemize}
  498. \end{block}}
  499. \end{frame}
  500.  
  501. \begin{frame}
  502. \frametitle{cross check (not in the note)}
  503. \begin{itemize}
  504. \item shouldn't the multivariate classifier do better than any trigger?
  505. \item back-of-the-envelope overestimates the improvement (expect $\sim 9\,\%$ improvement)
  506. \item[$\rightarrow$] add events back to the ntuple, recalculate normalisation, redid fits
  507. \item[$\Rightarrow$] \textcolor{darkgreen}{restricting the triggers gains $\sim 3\,\%$ sensitivity wrt.\ previous round}
  508. \end{itemize}
  509. \end{frame}
  510. \beamertemplateshadingbackground{White}{White}
  511.  
  512. \section{normalisation}
  513.  
  514. \begin{frame}
  515. \frametitle{relative normalisation}
  516. $\mathcal{B}(\Ptau\to\Pmu\Pmu\Pmu) = \frac{\mathcal{B}(\PDs\to\Pphi\Ppi)}{\mathcal{B}(\PDs\to\Ptau\Pnut)} \times f_{\PDs}^{\Ptau} \times \frac{\varepsilon_\text{norm} }{\varepsilon_\text{sig} } \times \frac{N_\text{sig}}{N_\text{norm}} = \alpha\times N_\text{sig}$
  517. \begin{itemize}
  518. \item where $\varepsilon$ stands for trigger, reconstruction, selection
  519. \item $\text{norm}$ = normalisation channel $\PDs\to\Pphi\Ppi$
  520. \item $f_{\PDs}^{\Ptau}$ is the fraction of $\Ptau$ coming from $\PDs$
  521. \newline i.e.\ $(83\pm3)\,\%$ for 2012
  522. \end{itemize}
  523. \includegraphics[width=.47\textwidth]{./Ds_data_2011.pdf}
  524. \includegraphics[width=.47\textwidth]{./Ds_data_2012.pdf}
  525. \end{frame}
  526.  
  527. \begin{frame}[allowframebreaks]
  528. \frametitle{normalisation in numbers}
  529. {\footnotesize{
  530. $\begin{array}{c|c|c}
  531. & \rm{7~TeV} & \rm{8~TeV}\\
  532. \hline
  533. \rm{\epsilon\mathstrut_{sig}}^{GEN} & 8.98 \pm 0.41 & 9.20 \pm 0.36\\
  534. \hline
  535. \rm{\epsilon_{cal}}^{GEN} & 11.19 \pm 0.34 & 11.53 \pm 0.32\\
  536. \hline
  537. \rm{\epsilon_{sig}}^{REC,isMuon,SEL} & 9.794 \pm 0.028 & 9.134 \pm 0.023 \\
  538. \hline
  539. \rm{\epsilon_{cal}}^{REC,isMuon,SEL} & 7.187 \pm 0.022 & 6.690 \pm 0.022 \\
  540. \hline
  541. \frac{\rm{c_{cal}}^{track}}{\rm{c_{sig}}^{track}} & 0.997 \pm 0.009 \pm 0.026 & 0.996 \pm 0.009 \pm 0.026 \\
  542. \hline
  543. \frac{\rm{c_{cal}}^{\mu ID}}{\rm{c_{sig}}^{\mu ID}} & 0.9731 \pm 0.0031 \pm 0.0264 & 1.0071 \pm 0.0022 \pm 0.0204 \\
  544. \hline
  545. \rm{c}^{\Pphi} & \multicolumn{2}{c}{0.98 \pm 0.01} \\
  546. \hline
  547. \rm{c}^{\Ptau} & 1.032 \pm 0.006 & 1.026 \pm 0.006\\
  548. \hline
  549. \rm{c}^{trash} & 1.95 \pm 0.12 & 2.05 \pm 0.12\\
  550. \hline
  551. \rm{\epsilon\mathstrut_{sig}}^{TRIG} & 35.45 \pm 0.11 \pm 0.14 & 39.1 \pm 1.7 \pm 2.0 \\
  552. \hline
  553. \rm{\epsilon\mathstrut_{cal}}^{TRIG} & 23.42 \pm 0.14 \pm 0.09 & 20.62 \pm 0.76 \pm 1.07 \\
  554. \end{array}$
  555. }}
  556.  
  557. \framebreak
  558.  
  559. {\footnotesize{
  560. $\begin{array}{c|c|c}
  561. & \rm{7~TeV} & \rm{8~TeV}\\
  562. \hline
  563. \mathcal{B}(\PDs\to\Pphi\Ppi) & \multicolumn{2}{c}{(1.317 \pm 0.099) \times 10^{-5}}\\
  564. \hline
  565. f^{\Ptau}_{\PDs} & 0.82 \pm 0.03 & 0.83 \pm 0.03 \\
  566. \hline
  567. \mathcal{B}(\PDs\to\Ptau\Pnut) & \multicolumn{2}{c}{0.0561 \pm 0.0024}\\
  568. \hline
  569. \rm{\epsilon\mathstrut_{norm}}^{REC\&SEL}/
  570. \rm{\epsilon\mathstrut_{sig}}^{REC\&SEL}
  571. & 0.897 \pm 0.061 & 0.926 \pm 0.056 \\
  572. \hline
  573. \rm{\epsilon\mathstrut_{norm}}^{TRIG}/
  574. \rm{\epsilon\mathstrut_{sig}}^{TRIG}
  575. & 0.6606 \pm 0.0059 & 0.527 \pm 0.041\\
  576. \hline
  577. N_{norm} & 28,162 \pm 434 & 51,998 \pm 684\\
  578. \hline & \\[-1.5em]\hline
  579. \alpha & (4.05 \pm 0.48) \times 10^{-9} & (1.83 \pm 0.25) \times 10^{-9}\\
  580. \alpha^{trash} & (7.90 \pm 0.49) \times 10^{-9} & (3.75 \pm 0.27) \times 10^{-9}\\
  581. \end{array}$
  582. }}
  583. \end{frame}
  584.  
  585.  
  586. \section{backgrounds}
  587.  
  588. \begin{frame}
  589. \frametitle{misidentification}
  590. \begin{itemize}
  591. \item most dominant: $\PDplus\to\PK\Ppi\Ppi$
  592. \item experience from last round: cut away low ProbNNmu range
  593. \item check remaining data under $\PK\Ppi\Ppi$ hypothesis for $\PDplus$ peak
  594. \item[$\Rightarrow$] misid safely contained in ``trash'' bin
  595. \item \textcolor{darkgreen}{$\PDplus\to\Ppi\Ppi\Ppi$ and $\PDs\to\Ppi\Ppi\Ppi$ start to become visible in 2012}
  596. \end{itemize}
  597. \includegraphics[width=.45\textwidth]{./Dp2Kpipi_all_2012_senseBins.pdf}
  598. \includegraphics[width=.45\textwidth]{./FittoD23pi_2012.pdf}
  599. \end{frame}
  600.  
  601. \begin{frame}
  602. \frametitle{evil backgrounds}
  603. \begin{itemize}
  604. \item $\Pphi\to\Pmu\Pmu + X$: narrow veto on dimuon mass
  605. \item $\PDs\to\Peta(\Pmu\Pmu\Pphoton)\Pmu\Pnum$: not so easy
  606. \begin{itemize}
  607. \item modelled in CONF note
  608. \item optimised veto in PAPER
  609. \item right now: both versions in the ANA note
  610. \end{itemize}
  611. \item baseline: veto $m_{\APmuon\Pmuon} < \unit{450}{\MeV}$
  612. \begin{itemize}
  613. \item fits better understood
  614. \item sensitivity unchanged when removing veto
  615. \item smaller uncertainty on expected background
  616. \end{itemize}
  617. \end{itemize}
  618. \end{frame}
  619.  
  620. \begin{frame}
  621. \frametitle{remaining backgrounds}
  622. \begin{itemize}
  623. \item fit exponential to invariant mass spectrum in each likelihood bin
  624. \item don't use $\pm \unit{30}{\MeV}$ in the fit
  625. \item[$\rightarrow$] compatible results blinding only $\pm \unit{20}{\MeV}$\footnote{partially used in classifier developement}
  626. \end{itemize}
  627. {\begin{center}
  628. most sensitive bins in 2011 and 2012
  629. \includegraphics[width=.4\textwidth]{./fit2011.png}
  630. \includegraphics[width=.4\textwidth]{./fit2012.png}
  631. \end{center}}
  632. \end{frame}
  633.  
  634. \section{results}
  635.  
  636. \begin{frame}
  637. \frametitle{expected limit}
  638. \begin{itemize}
  639. \item still blinded
  640. \item consider nuisance parameters from background fit, signal pdf calibration, normalisation
  641. \item nuisance parameters due to $\Ptau$ production not included in signal pdf shape, yet
  642. \item limit for combined 2011+2012 analysis
  643. \end{itemize}
  644. \end{frame}
  645.  
  646. \begin{frame}
  647. \frametitle{sensitivity}
  648. $\mathcal{B}(\Ptau\to\Pmu\Pmu\Pmu)<5.6 \times 10^{-8}$ at 90\% CL
  649.  
  650. \includegraphics[width=.8\textwidth]{./banana.png}
  651. \end{frame}
  652.  
  653. \beamertemplateshadingbackground{PineGreen!30}{White}
  654. \section{model dependence}
  655.  
  656. \begin{frame}
  657. \frametitle{model dependence}
  658. \begin{itemize}
  659. \item $\Peta$ veto $\Rightarrow$ our limit not applicable to New Physics with small $m_{\APmuon\Pmuon}$
  660. \item model independent description in \texttt{arXiv:0707.0988}
  661. \item 5 relevant Dalitz distributions: 2 four-point operators, 1 radiative operator, 2 interference terms
  662. \end{itemize}
  663. \only<2->{
  664. \begin{itemize}
  665. \item with radiative distribution limit gets worse by $51\,\%$ (dominantly from the $\Peta$ veto)
  666. \item the other four Dalitz distributions behave nicely (within $7\,\%$)
  667. \end{itemize}
  668. }
  669. \only<1>{
  670. \includegraphics[width=.331\textwidth]{./gammallll.eps}
  671. \includegraphics[width=.331\textwidth]{./gammallrr.eps}
  672. \includegraphics[width=.331\textwidth]{./gammarad.eps}
  673.  
  674. \includegraphics[width=.331\textwidth]{./gammarad-llll.eps}
  675. \includegraphics[width=.331\textwidth]{./gammarad-llrr.eps}
  676. }
  677.  
  678. \end{frame}
  679.  
  680. \beamertemplateshadingbackground{White}{White}
  681.  
  682. \begin{frame}
  683. \frametitle{Conclusion}
  684. \begin{columns}
  685. \begin{column}{.55\textwidth}
  686. \begin{itemize}
  687. \item finally all pieces put together
  688. \item model (in)dependence of $\Peta$ veto investigated
  689. \item expected sensitivity computed\newline $5.6\times 10^{-8}$
  690. \end{itemize}
  691. \end{column}
  692. \begin{column}{.45\textwidth}
  693. \includegraphics[width=\textwidth]{party-music-hd-wallpaper-1920x1200-3850.jpg}
  694. \end{column}
  695. \end{columns}
  696.  
  697. \end{frame}
  698. \end{document}