dg-fisma-2026.tex (22195B)
1 \documentclass[aspectratio=169,t]{beamer} 2 \input taler-macros 3 \usepackage{eurosym} 4 \usepackage{array} 5 6 \newcommand{\TITLE}{NEXT \\ GENERATION \\ INTERNET} 7 \newcommand{\SUB}{Digital Euro: privacy as a property} 8 \newcommand{\AUTHOR}{Tanja Lange, Christian Grothoff, Leon Schumacher} 9 \newcommand{\SPEAKER}{Lange, Grothoff, Schumacher} 10 \newcommand{\INST}{NGI TALER} 11 \newcommand{\DATE}{\\ \ \vspace{0.5cm} \\ October 20\textsuperscript{th} 2026} 12 13 % Do not edit this part 14 \title{\TITLE} 15 \subtitle{\SUB} 16 \date{\DATE} 17 \author[\SPEAKER]{\AUTHOR} 18 \institute{\INST} 19 20 \begin{document} 21 22 \begin{frame}[plain] 23 \maketitle 24 \end{frame} 25 26 27 % ---------------------------------------------------------------------- 28 \begin{frame}{Five disputes, one root cause} 29 Still open in trilogue: {\bf privacy}, {\bf holding limits}, 30 {\bf compensation}, {\bf distribution}, {\bf offline}. 31 \vfill 32 Each is being negotiated as {\bf policy constraints on data the 33 architecture collects anyway.} 34 \vfill 35 \begin{itemize} 36 \item The Regulation governs {\bf what may be done} with payment data. 37 \item The Rulebook already fixed {\bf how payments work}. Its glossary 38 defines the user identifier as {\bf pseudonymous}, ``attributable to 39 an identifiable natural or legal person by the payment service 40 provider''. 41 \item Across the whole draft Rulebook the words {\bf anonymity} and 42 {\bf blind signature} do not appear. 43 \end{itemize} 44 \vfill 45 \begin{center} 46 The architecture was settled {\bf before} the privacy question was 47 answered. No drafting raises that ceiling. 48 \end{center} 49 \end{frame} 50 51 52 % ---------------------------------------------------------------------- 53 \begin{frame}{Pseudonymous is not cash-like} 54 \begin{itemize} 55 \item One intermediary necessarily knows {\bf both} the payment 56 identifier and the bank account --- it must, to fund the wallet. 57 \item Correlated with location or device data, a pseudonymous history 58 {\bf de-anonymises retroactively} --- and no procedure exists to 59 obtain a new identifier. 60 \item Payment data stops being siloed across thousands of institutions 61 and becomes {\bf one queryable history}. 62 \item The payer {\bf cannot verify} any of it. 63 \end{itemize} 64 \vfill 65 \begin{center} 66 {\small ECB progress report, July 2025: ``{\bf Only PSPs will be able to 67 identify the end user and link that identity to individual 68 transactions.}''}\\[4pt] 69 Privacy that depends on data being {\bf held but not used} is a policy 70 setting, not a property. 71 \end{center} 72 \end{frame} 73 74 75 % ---------------------------------------------------------------------- 76 \begin{frame}{The demand and the design are on the record} 77 \small 78 \begin{itemize} 79 \item {\bf 43\%} of respondents to the ECB's own consultation ranked 80 privacy the single most important feature. In the Bundesbank's 2020 81 study, {\bf 94\%} called it very important and {\bf 59\%} decisive. 82 \item {\bf 14 civil-society organisations}, including EDRi, wrote to the 83 co-legislators in June: privacy {\bf technically enforced}, open 84 standards, {\bf source code published}. 85 \item The {\bf Austrian} central bank has called bearer-token digital 86 cash a possible template; the {\bf BIS} called for exactly this work 87 in 2021. 88 \item {\bf Banca d'Italia} specified an eCash component using blind 89 signatures --- in a paper the {\bf ECB republished on its own site}. 90 \end{itemize} 91 \vfill 92 \begin{center} 93 The demand exists and the design is known. 94 {\bf The architecture is what blocks it.} 95 \end{center} 96 \end{frame} 97 98 99 % ---------------------------------------------------------------------- 100 \begin{frame}{The other way to build it} 101 \framesubtitle{Digital cash, not an account} 102 \begin{enumerate} 103 \item {\bf Withdraw.} The wallet obtains digital coins from a regulated 104 operator against a debit of the customer's account. The operator signs 105 coins it {\bf cannot recognise later}. 106 \item {\bf Pay.} The wallet spends coins directly. No account, no 107 credential, nothing reusable is shown to the merchant. 108 \item {\bf Deposit.} The merchant deposits the coins with the 109 {\bf operator}, which detects any double-spend and credits the 110 merchant's account. 111 \end{enumerate} 112 \vfill 113 Not a currency. Not a speculative asset. Not decentralised. Not a 114 blockchain. No proof-of-work, no proof-of-stake. 115 \end{frame} 116 117 118 % ---------------------------------------------------------------------- 119 \begin{frame}{Architecture} 120 \begin{center} 121 \includegraphics[height=6.0cm]{images/operations.png} 122 \end{center} 123 \end{frame} 124 125 126 % ---------------------------------------------------------------------- 127 \begin{frame}{Asymmetric privacy answers AML} 128 \begin{minipage}{0.47\textwidth} 129 {\bf Payer} 130 \begin{itemize} 131 \item Unlinkable by construction 132 \item No identity in the payment 133 \item Carries nobody else's data 134 \end{itemize} 135 \end{minipage} 136 \hfill 137 \begin{minipage}{0.47\textwidth} 138 {\bf Payee} 139 \begin{itemize} 140 \item Always identified 141 \item Every receipt attributable 142 \item Income auditable and taxable 143 \end{itemize} 144 \end{minipage} 145 146 \vfill 147 KYC/AML applies where it works: at {\bf withdrawal} and at {\bf deposit}, 148 inside the regulated perimeter. Anonymity is {\bf one-sided}, so there is 149 no anonymous \emph{recipient} --- the laundering direction is closed. 150 \vfill 151 \begin{center} 152 Cash protects the payer \emph{and} hides the merchant.\\ 153 This protects the payer and {\bf exposes} the merchant.\\[4pt] 154 {\small Not only a rights argument: asymmetric privacy is 155 {\bf welfare-optimal} --- Tinn, \emph{Management Science}, 2025.} 156 \end{center} 157 \end{frame} 158 159 160 % ---------------------------------------------------------------------- 161 \begin{frame}{Offline: what cryptography cannot deliver} 162 \begin{itemize} 163 \item Consistency, availability and partition-tolerance cannot be had 164 together. Offline double-spending can be {\bf punished}, never 165 {\bf prevented}. 166 \item So an offline bearer balance has exactly two options: trust 167 tamper-resistant hardware {\bf against its own owner}, or 168 {\bf de-anonymise retroactively}. There is no third. 169 \item Damage scales with {\bf transaction volume}, not with the balance 170 held --- and an attacker can {\bf provoke} the outage that enables it. 171 \end{itemize} 172 \vfill 173 \begin{center} 174 {\bf The risk in this trilogue:} conceding online privacy in exchange 175 for an offline promise hardware cannot keep.\\ 176 Online is where almost every payment happens. 177 \end{center} 178 \end{frame} 179 180 181 % ---------------------------------------------------------------------- 182 \begin{frame}[fragile]{Securing hardware against its owner} 183 \framesubtitle{Only if history has nothing to teach us} 184 \def\svgwidth{0.9\textwidth} 185 186 \begingroup% 187 \makeatletter% 188 \providecommand\color[2][]{% 189 \errmessage{(Inkscape) Color is used for the text in Inkscape, but the package 'color.sty' is not loaded}% 190 \renewcommand\color[2][]{}% 191 }% 192 \providecommand\transparent[1]{% 193 \errmessage{(Inkscape) Transparency is used (non-zero) for the text in Inkscape, but the package 'transparent.sty' is not loaded}% 194 \renewcommand\transparent[1]{}% 195 }% 196 \providecommand\rotatebox[2]{#2}% 197 \newcommand*\fsize{\dimexpr\f@size pt\relax}% 198 \newcommand*\lineheight[1]{\fontsize{\fsize}{#1\fsize}\selectfont}% 199 \ifx\svgwidth\undefined% 200 \setlength{\unitlength}{345bp}% 201 \ifx\svgscale\undefined% 202 \relax% 203 \else% 204 \setlength{\unitlength}{\unitlength * \real{\svgscale}}% 205 \fi% 206 \else% 207 \setlength{\unitlength}{\svgwidth}% 208 \fi% 209 \global\let\svgwidth\undefined% 210 \global\let\svgscale\undefined% 211 \makeatother% 212 \begin{picture}(1,0.53623188)% 213 \lineheight{1}% 214 \setlength\tabcolsep{0pt}% 215 \put(0,0){\includegraphics[width=\unitlength,page=1]{images/offline-timeline.pdf}}% 216 \put(0.10624514,0.04914349){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l }\tiny Feb 2017\end{tabular}}}}% 217 \put(0,0){\includegraphics[width=\unitlength,page=2]{images/offline-timeline.pdf}}% 218 \put(0.28309276,0.44884928){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny Mar, Jun 2020\end{tabular}}}}% 219 \put(0,0){\includegraphics[width=\unitlength,page=3]{images/offline-timeline.pdf}}% 220 \put(0.18177392,0.17262607){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny Aug 2017\end{tabular}}}}% 221 \put(0,0){\includegraphics[width=\unitlength,page=4]{images/offline-timeline.pdf}}% 222 \put(0.08024638,0.44884927){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny Dec 2017\end{tabular}}}}% 223 \put(0,0){\includegraphics[width=\unitlength,page=5]{images/offline-timeline.pdf}}% 224 \put(0.02512174,0.36103189){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny May, Aug 2016\end{tabular}}}}% 225 \put(0,0){\includegraphics[width=\unitlength,page=6]{images/offline-timeline.pdf}}% 226 \put(0.34119422,0.08566952){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny Aug 2019\end{tabular}}}}% 227 \put(0,0){\includegraphics[width=\unitlength,page=7]{images/offline-timeline.pdf}}% 228 \put(0.3956406,0.36131883){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny Oct 2019\end{tabular}}}}% 229 \put(0,0){\includegraphics[width=\unitlength,page=8]{images/offline-timeline.pdf}}% 230 \put(0.38164733,0.17343635){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny Jun 2020\end{tabular}}}}% 231 \put(0,0){\includegraphics[width=\unitlength,page=9]{images/offline-timeline.pdf}}% 232 \put(0.59564059,0.44827535){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny Sep 2022\end{tabular}}}}% 233 \put(0,0){\includegraphics[width=\unitlength,page=10]{images/offline-timeline.pdf}}% 234 \put(0.55650392,0.05335936){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny Mar 2023\end{tabular}}}}% 235 \put(0,0){\includegraphics[width=\unitlength,page=11]{images/offline-timeline.pdf}}% 236 \put(0.80271684,0.06728262){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny Sep 2024\end{tabular}}}}% 237 \put(0,0){\includegraphics[width=\unitlength,page=12]{images/offline-timeline.pdf}}% 238 \put(0.59062556,0.17319998){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny Feb 2023\end{tabular}}}}% 239 \put(0,0){\includegraphics[width=\unitlength,page=13]{images/offline-timeline.pdf}}% 240 \put(0.78984349,0.17291304){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny Oct 2024\end{tabular}}}}% 241 \put(0,0){\includegraphics[width=\unitlength,page=14]{images/offline-timeline.pdf}}% 242 \put(0.7963995,0.44796104){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny Aug 2023\end{tabular}}}}% 243 \put(0,0){\includegraphics[width=\unitlength,page=15]{images/offline-timeline.pdf}}% 244 \put(0.59916521,0.36103188){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny Apr 2023\end{tabular}}}}% 245 \put(0,0){\includegraphics[width=\unitlength,page=16]{images/offline-timeline.pdf}}% 246 \put(0.79631301,0.36103188){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}\tiny Aug 2024\end{tabular}}}}% 247 \put(0,0){\includegraphics[width=\unitlength,page=17]{images/offline-timeline.pdf}}% 248 \put(0.123203,0.24410952){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}2015\end{tabular}}}}% 249 \put(0.5000146,0.24410952){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}2020\end{tabular}}}}% 250 \put(0.87682623,0.24410952){\color[rgb]{0,0,0}\makebox(0,0)[lt]{\lineheight{1.25}\smash{\begin{tabular}[t]{l}2025\end{tabular}}}}% 251 \put(0,0){\includegraphics[width=\unitlength,page=18]{images/offline-timeline.pdf}}% 252 \end{picture}% 253 \endgroup% 254 \end{frame} 255 256 257 % ---------------------------------------------------------------------- 258 \begin{frame}{Holding limits: the wrong control variable} 259 \small 260 \begin{minipage}{0.47\textwidth} 261 {\bf One negotiated ceiling} 262 \begin{itemize} 263 \item \EUR{3{,}000} retail, zero for business 264 \item A political number, for everyone 265 \item Invites gaming of the allocation 266 \item Leaves the balance a {\bf direct central-bank liability} 267 \end{itemize} 268 \end{minipage} 269 \hfill 270 \begin{minipage}{0.47\textwidth} 271 {\bf Four independent levers} 272 \begin{itemize} 273 \item Withdrawal limits at the distributing bank 274 \item Limits set by the wallet provider 275 \item Self-custody risk borne by the holder 276 \item Coin expiry and refresh fees discourage hoarding 277 \end{itemize} 278 \end{minipage} 279 280 \vfill 281 \begin{center} 282 Banca d'Italia, on eCash: withdrawal {\bf applies all the controls of an 283 account-based system}.\\ 284 SNB: backed by reserves, it replicates {\bf cash, not deposits}. 285 \end{center} 286 \end{frame} 287 288 289 % ---------------------------------------------------------------------- 290 \begin{frame}{Cost and competition} 291 \begin{itemize} 292 \item {\bf \EUR{1.3}\,bn} to first issuance, then {\bf \EUR{320}\,m} a 293 year --- before a single bank or merchant has adapted. 294 \item Bank-side adaptation: the banks' own study says 295 {\bf \EUR{18}\,bn}; the ECB says {\bf \EUR{4}--5.8\,bn}. A factor of 296 three to four, unresolved --- and merchants bear integration costs 297 that mandated acceptance forces on them. 298 \item Tender eligibility has required turnover in the order of 299 {\bf \EUR{100}\,m} overall and {\bf \EUR{10}\,m} in comparable 300 services. 301 \end{itemize} 302 \vfill 303 \begin{center} 304 Those thresholds exclude {\bf every European open-source provider}, 305 including us, before the technical comparison begins.\\[4pt] 306 Lowering the system's {\bf own} cost is not the same as having the ECB 307 {\bf absorb} it. 308 \end{center} 309 \end{frame} 310 311 312 % ---------------------------------------------------------------------- 313 \begin{frame}{Sovereignty is a property of the licence} 314 \begin{itemize} 315 \item A proprietary national scheme is adopted {\bf nowhere else}: no 316 state cedes its payment rails to another's vendor. 317 \item Free software defines a {\bf non-discriminatory standard} any 318 country can adopt and operate {\bf without} depending on Europe's 319 goodwill --- which is precisely why it travels. 320 \item The alternative is familiar: infrastructure controlled elsewhere, 321 reachable by foreign law. {\bf What happened with e-mail can happen 322 with payments.} 323 \item Offline on proprietary secure elements hard-wires dependence on 324 {\bf two non-European platform vendors}. 325 \end{itemize} 326 \vfill 327 \begin{center} 328 DG CONNECT funds the European sovereign mobile stack 329 ({\bf \EUR{32}\,m} to 2029). Taler is already its payment layer. 330 \end{center} 331 \end{frame} 332 333 334 % ---------------------------------------------------------------------- 335 \begin{frame}{Inclusion is decided by the architecture too} 336 \begin{itemize} 337 \item A wallet that is {\bf software} needs no account to hold or spend 338 --- so children, tourists, refugees and the unbanked are not excluded 339 by onboarding they cannot complete. 340 \item Over {\bf five million} children in the EU aged 10--18 cannot open 341 a bank account. An account-based instrument excludes them by design. 342 \item Field testing with users who could not read a five-digit number: 343 {\bf 20 of 21} completed every payment task unaided after a day's gap. 344 \item Browser and terminal support matter: an instrument that lives only 345 on two mobile platforms is not universal. 346 \end{itemize} 347 \vfill 348 \begin{center} 349 Legal tender that some citizens cannot obtain is not legal tender for 350 them. 351 \end{center} 352 \end{frame} 353 354 355 % ---------------------------------------------------------------------- 356 \begin{frame}{Where the software actually is} 357 \framesubtitle{So that nothing here is taken on trust} 358 \small 359 \begin{minipage}{0.47\textwidth} 360 {\bf Done} 361 \begin{itemize} 362 \item Live in Switzerland in CHF under the FINMA {\bf sandbox 363 exemption} --- {\bf not} FINMA-supervised; AML through a 364 recognised SRO 365 \item v1.6; independent external security audits 366 \item {\bf 28{,}500} transactions/s measured at reduced key sizes 367 --- Switzerland needs $\approx$2{,}000, the euro area 368 $\approx$100{,}000 369 \item Android, iOS, browser; EBICS/ISO 20022 370 \end{itemize} 371 \end{minipage} 372 \hfill 373 \begin{minipage}{0.47\textwidth} 374 {\bf Not done} 375 \begin{itemize} 376 \item No euro-area deployment yet; Hungary in internal alpha, with 377 no launch date announced 378 \item Sandbox caps the Swiss operator at {\bf CHF 1\,m} in total 379 deposits 380 \item Fully offline (both sides) {\bf unsolved} --- as it is for 381 everyone 382 \item Direct debit and pull payments need a different mechanism 383 \end{itemize} 384 \end{minipage} 385 \end{frame} 386 387 388 % ---------------------------------------------------------------------- 389 \begin{frame}{The objections, and the answers} 390 \small 391 \begin{itemize} 392 \item {\em ``User anonymity is not a desirable feature.''} --- ECB to 393 the Eurogroup, 2022.\\ 394 {\bf Payer} anonymity is not system anonymity. The payee is always 395 identified, and every withdrawal and deposit sits inside the 396 regulated perimeter. 397 \vspace{2pt} 398 \item {\em Full anonymity ``could facilitate illegal activity''.} --- 399 BIS Annual Report, 2021.\\ 400 Agreed --- which is why no serious bearer design offers it on 401 {\bf both} sides. One-sided anonymity closes the laundering 402 direction. 403 \vspace{2pt} 404 \item {\em A ledger defeats privacy by definition; token designs are 405 ``misguided''.} --- Riksbank staff memo, 2021.\\ 406 The ledger records {\bf spent coins, not identities}. That is 407 precisely what a blind signature achieves --- and the memo engages no 408 blind-signature design. 409 \end{itemize} 410 \end{frame} 411 412 413 % ---------------------------------------------------------------------- 414 \begin{frame}{Four things for the compromise text} 415 {\small We are not asking for a competing currency, a reopening of the 416 monetary-policy case, a delay, or a procurement decision in our 417 favour.} 418 \vspace{4pt} 419 \begin{enumerate} 420 \item {\bf Unlinkability, not confidentiality.} A payer's transactions 421 must be {\bf technically unlinkable} by whoever holds the 422 infrastructure. Today's text protects the data; this prevents its 423 creation. 424 \item {\bf Properties, not mechanisms.} Keep the account/token choice 425 {\bf out} of the Regulation; hold the Rulebook to demonstrating 426 conformance. 427 \item {\bf Open specification, verifiable implementation.} An open, 428 royalty-free specification and one {\bf free and open-source} 429 reference implementation --- as 14 organisations have already asked. 430 \item {\bf Proportionate eligibility.} Procurement thresholds that do 431 not foreclose European providers before the comparison begins. 432 \end{enumerate} 433 \end{frame} 434 435 436 % ---------------------------------------------------------------------- 437 \begin{frame}{And one thing for the pilot} 438 \vfill 439 \begin{center} 440 {\bf Benchmark one bearer-token architecture alongside the 441 account-based one.} 442 \end{center} 443 \vfill 444 \begin{itemize} 445 \item On the trilogue's {\bf own} contested metrics: payer privacy, cost 446 per transaction, inclusion without an account, offline risk exposure, 447 low-value payments. 448 \item The software is {\bf already written, deployed and supervised}. 449 This is a comparison, not a development programme. 450 \item It replaces a negotiation over assertions with {\bf evidence}, 451 before the privacy article has to be settled. 452 \end{itemize} 453 \vfill 454 \begin{center} 455 Cheap, reversible, and it strengthens whichever design wins. 456 \end{center} 457 \end{frame} 458 459 460 % ---------------------------------------------------------------------- 461 \begin{frame}{Summary} 462 \framesubtitle{What design settles, and what text must do} 463 \scriptsize 464 \renewcommand{\arraystretch}{1.25} 465 \begin{tabular}{% 466 >{\raggedright\arraybackslash}p{0.235\textwidth}% 467 >{\raggedright\arraybackslash}p{0.32\textwidth}% 468 >{\raggedright\arraybackslash}p{0.32\textwidth}} 469 {\bf Dispute} & {\bf Settled by design} & {\bf Still needs text} 470 \\ 471 \hline 472 Privacy vs.\ AML & 473 Payer unlinkable, payee always identified & 474 Unlinkability as a requirement \\ 475 Inclusion vs.\ KYC & 476 No account needed to hold or spend & 477 Controls fixed at the fiat perimeter \\ 478 Holding limits vs.\ stability & 479 No direct retail central-bank liability & 480 Limits by risk tier, not one ceiling \\ 481 Fees vs.\ compensation & 482 Low marginal cost, aggregated settlement & 483 No mandated interchange layer \\ 484 Distribution vs.\ gatekeepers & 485 Any operator, any wallet, one protocol & 486 Open, royalty-free specification \\ 487 Sovereignty vs.\ centralisation & 488 Inspectable, deployable, forkable & 489 FLOSS reference implementation \\ 490 \end{tabular} 491 \end{frame} 492 493 494 % This should be last... 495 \begin{frame}{Acknowledgements} 496 497 \begin{minipage}{0.45\textwidth} \ \\ 498 {\tiny Funded by the European Union (Project 101135475).} 499 500 \begin{center} 501 \includegraphics[width=0.5\textwidth]{images/bandera.jpg} 502 \end{center} 503 \end{minipage} 504 \hfill 505 \begin{minipage}{0.45\textwidth} 506 {\tiny Funded by SERI (HEU-Projekt 101135475-TALER).} 507 508 \begin{center} 509 \includegraphics[width=0.65\textwidth]{images/sbfi.jpg} 510 \end{center} 511 \end{minipage} 512 513 \vfill 514 515 {\tiny 516 517 Views and opinions expressed are however those of the author(s) only 518 and do not necessarily reflect those of the European Union. Neither the 519 European Union nor the granting authority can be held responsible for 520 them. 521 522 } 523 \end{frame} 524 525 \end{document}