YouTube2Text

Challenge Briefing and Live Q&A — Transcript

by ADB Digital Sandbox · 6,056 words · 893 segments · language en · Watch on YouTube

Full transcript

  1. 0:02So, there's a bit of a miss messages,
  2. 0:04but um yeah, I you saw my
  3. 0:08>> Yeah. Yeah. Okay. Don't worry, Jeff.
  4. 0:09We'll talk about it later. It's good.
  5. 0:11>> Good. Good. Thank you. Thank you very
  6. 0:13much.
  7. 0:15>> Of course.
  8. 0:26Just waiting for more people to join.
  9. 1:30Hi Verdon. Good morning. Good afternoon.
  10. 1:33Thanks everyone for joining.
  11. 1:38>> Hello.
  12. 1:41>> Good morning.
  13. 1:44I think we should wait for a couple of
  14. 1:45minutes for more people to join. I I got
  15. 1:47rece I keep receiving emails at like
  16. 1:494:00 a.m. my time that people were
  17. 1:51already waiting to join. So I see the
  18. 1:53enthusiasm which is really exciting.
  19. 1:58It's
  20. 2:16maybe we can start in about
  21. 2:19two
  22. 2:22two minutes.
  23. 3:07Yes, and thank you Eugene for raising
  24. 3:09the question. uh if everyone's looking
  25. 3:11to join a team still uh yes I think it's
  26. 3:15yeah if anyone wants to form teams
  27. 3:17outside of this call of course you're
  28. 3:18more than welcome to get uh to know each
  29. 3:21other and uh yeah so if anyone can wants
  30. 3:24to reach out to Eugene you're more than
  31. 3:26welcome
  32. 3:47Okay. Should we perhaps
  33. 3:50start in one minute? Is that okay for
  34. 3:52everyone?
  35. 3:55>> Yes, we're done.
  36. 3:56James. Okay.
  37. 3:57>> Yeah.
  38. 3:58>> Yeah. Fab Fabian um as well is on the
  39. 4:02call. I think actually he might want to
  40. 4:04start with a couple of slides as well.
  41. 4:08>> Oh, he would like to share some slides
  42. 4:10or
  43. 4:12>> Yeah. Is that Can we just check if it's
  44. 4:14possible?
  45. 4:14>> Yes, I I'm here. I don't know if you see
  46. 4:17my video actually.
  47. 4:19>> Uh
  48. 4:19>> hi James.
  49. 4:21>> Yeah, I can see you now.
  50. 4:24>> Hi. Hi Febian. Hi.
  51. 4:25>> Hi. Good morning.
  52. 4:27>> Good morning. Good morning. Um, let me
  53. 4:29see if I'm gonna stop sharing my slides
  54. 4:33and then uh
  55. 4:35um
  56. 4:40stop sharing. Uh do you can you share
  57. 4:42your from your side?
  58. 4:46>> Yes. I I will just quickly set up the
  59. 4:49the sharing of my the the presentation
  60. 4:53mode which always screws up my windows.
  61. 4:57So, but now I think I got it.
  62. 5:00So, let me just
  63. 5:07Okay.
  64. 5:10Very good.
  65. 5:16Yes. So, yes. So, uh I think do you want
  66. 5:19to do another introduction or should I
  67. 5:21just jump in
  68. 5:23>> from our side?
  69. 5:25>> Um you can jump in and then we can we
  70. 5:28can do the industry challenge later. But
  71. 5:30yeah, because we're done, we'll take
  72. 5:31over.
  73. 5:32>> Mhm.
  74. 5:32>> Perfect. So, yeah. So, I I'm I'm from my
  75. 5:35name is Fabian. I'm based out of
  76. 5:37Switzerland, Zurich. I'm I'm at AWS like
  77. 5:41in a team that's called advanced
  78. 5:42solutions lab. We're working with with
  79. 5:45customers on on quantum computation and
  80. 5:47I'm I'm basically here because also like
  81. 5:50AWS is kind of like like helping out
  82. 5:52here on this quantum challenge and with
  83. 5:54some credits because as you may know
  84. 5:57like Amazon bracket is our quantum
  85. 6:00service and I quickly just take a few
  86. 6:02slides to kind of show what we're doing
  87. 6:04at AWS in in the area of quantum
  88. 6:06computation.
  89. 6:13Yes. So we have basically three pillar.
  90. 6:15Main pillar of course is our service
  91. 6:17Amazon bracket as I already said where
  92. 6:18we our goal is to give access of quantum
  93. 6:21computers to a broad community as as as
  94. 6:25scientists as you that want to
  95. 6:27experiment with quantum computers. Then
  96. 6:30we have the AWS center for quantum
  97. 6:32computation which is our research
  98. 6:34facility where we do active research on
  99. 6:37quantum algorithms hardware and also
  100. 6:40networking and we have the Amazon
  101. 6:42advanced solutions lab that's where I'm
  102. 6:43in. We're working actually with
  103. 6:45enterprise customers uh in order to kind
  104. 6:48of help them to get started with quantum
  105. 6:50computation run research engagement with
  106. 6:52them and also build PC's out on Amazon
  107. 6:56bracket.
  108. 6:59So maybe first on Amazon bracket. Amazon
  109. 7:02bracket is part of AWS cloud. It's
  110. 7:04integrated in it. It provides you with
  111. 7:07several ways on connecting to it and to
  112. 7:09to leverage quantum computers. it uh you
  113. 7:12can you can work like from manage
  114. 7:14notebooks where you get the compute in
  115. 7:16on AWS uh in the AWS cloud through
  116. 7:20Amazon bracket or you can use your local
  117. 7:22ID for the for code development
  118. 7:26and then you can basically access like
  119. 7:29different simulators we have like state
  120. 7:31vector simulator tensor network
  121. 7:32simulator and densy matrix simulators uh
  122. 7:35simulators running on AWS cloud so we
  123. 7:38also have a simulator that runs directly
  124. 7:40on your laptop or on the notebooks if
  125. 7:42you uh like do that for for testing. But
  126. 7:46if you want to do it on a larger scale
  127. 7:48like uh this managed simulators actually
  128. 7:51help you to do that it's basically in
  129. 7:54the cloud starting like spinning up the
  130. 7:56compute that is required. You will do
  131. 7:58your computation and once the
  132. 8:00computation is done it you will uh also
  133. 8:03kind of depreciate all the kind of
  134. 8:05resources that has been used and we'll
  135. 8:07give you the the the results back. Of
  136. 8:10course that that will incur some costs
  137. 8:12on your side because these are managed
  138. 8:14compute. Then we have uh like like
  139. 8:17quantum hardware which we come into it
  140. 8:19as well. The quantum hardware usually
  141. 8:21don't live in our AWS uh like like uh
  142. 8:25data centers. They are usually accessed
  143. 8:28through API and standing at the
  144. 8:30different locations.
  145. 8:33So we have different quantum hardware
  146. 8:35that we we provide for you. is like the
  147. 8:38from IMQ with ion trap devices. We have
  148. 8:41like superc conducting with regetti and
  149. 8:43IMQ and we also have quer which is a
  150. 8:47like a neutral atom device which is a
  151. 8:50special purpose kind of Hamilton
  152. 8:52simulation uh QPU.
  153. 8:56So you will see like we have different
  154. 8:58numbers of cubits like for instance we
  155. 9:01also like the the INQ comes with kind of
  156. 9:03error mitigation techniques and then we
  157. 9:06have up to 108 cubits for
  158. 9:07superconducting with the newest rigetti.
  159. 9:14Good. And maybe highlighting one thing
  160. 9:16like we are actually approaching the
  161. 9:18area of kind of like uh quasi fall
  162. 9:22tolerant quantum computation
  163. 9:24which we also engaged in with our
  164. 9:26strategic partner quer which will not be
  165. 9:29at the first place like any accessed
  166. 9:32through the Amazon bracket will be kind
  167. 9:33of like tailor because these are kind of
  168. 9:35very advanced. We are we are planning to
  169. 9:38start off in 2028 with the first kind of
  170. 9:41quasi fall tolerant quantum comput
  171. 9:45uh computer computing device which gives
  172. 9:48you about 25
  173. 9:5156 cubits together with about 10 to the
  174. 9:54six uh operations that you can do in
  175. 9:56depth. So we are very excited actually
  176. 9:59on this journey and it also shows how
  177. 10:01how we progressing in that field.
  178. 10:07Good. So the Yeah. So not much to say on
  179. 10:10on on the AWS center for quantum
  180. 10:13computing. I mean their their uh
  181. 10:16location is is close to is it uh at
  182. 10:20close to Caltech and they also working
  183. 10:23on uh our own hardware.
  184. 10:27Maybe quickly to our team what we doing.
  185. 10:29So it's the Amazon advanced solutions
  186. 10:32lab. So we basically work with customers
  187. 10:34in order to search for uh quantum
  188. 10:37applications.
  189. 10:39As you know this these are and this is
  190. 10:41also part of this challenge. Finding
  191. 10:44industry relevant quantum application is
  192. 10:46still one of the holy grails and quantum
  193. 10:48computation. We are still working uh
  194. 10:50with custom on that especially bridging
  195. 10:52the gap between academia and and and the
  196. 10:55industries and for that we combine like
  197. 10:58different kind of like fields like
  198. 11:00quantum computing at the heart of it
  199. 11:03where we actually bring deep expertise
  200. 11:05and then but combining it with machine
  201. 11:07learning and optimization so we can
  202. 11:09benchmark the different approaches which
  203. 11:11is very important because like if you
  204. 11:14cannot show that it brings a quantum
  205. 11:16advantage against classical then then
  206. 11:19then the the the valid validity of
  207. 11:22results are is always questionable. So
  208. 11:24we we really want to combine these
  209. 11:26experties
  210. 11:27and I I I think like I'm stopping here
  211. 11:31and handing back to
  212. 11:33>> S. Yes. Thank you. Thank you so much
  213. 11:35Fab. Perfect. Thank you so much. Uh now
  214. 11:38I will share um our slides for this um
  215. 11:43challenge. So
  216. 11:46um
  217. 11:47yes, welcome everyone. Thank you so much
  218. 11:49for joining and thank you Fabin for your
  219. 11:51introduction. Um yeah so the agenda for
  220. 11:55today is more about uh getting to know
  221. 11:57what the project is about more in diving
  222. 12:00into the technical um implications of
  223. 12:03this deliverables and then of course for
  224. 12:05those who want to participate ask
  225. 12:07questions you're actually going to be
  226. 12:08asking questions for the team that's
  227. 12:10actually running this challenge. So just
  228. 12:12a brief hello um Fabian thank you yeah
  229. 12:15for your introduction. Now from my side
  230. 12:18um business development manager at Kai
  231. 12:20Ventures uh Verdon is a CTO and Tobias
  232. 12:24is our global head of technology and we
  233. 12:26also have James from ADB side on this
  234. 12:29call.
  235. 12:30Hello everyone and um James I will leave
  236. 12:35it to you to actually do an introduction
  237. 12:36on the challenge itself.
  238. 12:39>> Excellent. Uh thank you SA. So hi
  239. 12:41everyone, my name is James. I come from
  240. 12:43ADB uh from our digital innovation team.
  241. 12:46And this challenge specifically we're
  242. 12:48looking at something called a joint loss
  243. 12:50model or how we can replicate joint loss
  244. 12:52modeling using quantum systems. Uh very
  245. 12:56very quickly a joint loss model can be
  246. 12:58used to describe a set of uh
  247. 13:00interrelated or dependent events. Um,
  248. 13:04and often times when you design a
  249. 13:06traditional joint loss model, um, you're
  250. 13:08looking at, uh, what the likelihood of
  251. 13:10two or more related processes triggering
  252. 13:12simultaneously,
  253. 13:13um, might be. And specifically to us at
  254. 13:16at ADB, we're interested in this because
  255. 13:19uh these style of uh events,
  256. 13:22probabilistic events are very common
  257. 13:24across uh the problems that we look at
  258. 13:28uh that are related to um the financial
  259. 13:31services that we offer as a development
  260. 13:33bank. So you can think problems in the
  261. 13:36insurance industry uh such as insurance
  262. 13:38link securities or catastrophe bonds but
  263. 13:41also other models that we might look at
  264. 13:44such as bond market and looking at the
  265. 13:47secondary or tertiary uh spillover
  266. 13:50effects of certain bond market triggers
  267. 13:52that might happen. So there's a lot of
  268. 13:55macroeconomics um as well as a sort of
  269. 13:58smaller more traditional economic
  270. 14:00problems that we look at where we're
  271. 14:02trying to understand and trying to
  272. 14:03assess how quantum or quantum computing
  273. 14:06techniques could be leveraged in in this
  274. 14:08case. Specifically for this problem uh
  275. 14:11we are looking at a joint loss scenario
  276. 14:14for rainfall and for earthquake being
  277. 14:16triggered simultaneously across two
  278. 14:18countries in Carrick. And with this
  279. 14:22problem, because we're dealing with
  280. 14:23earthquakes and we're dealing with with
  281. 14:25rainfall events, there is the issue that
  282. 14:28we do have a very strong data
  283. 14:31constraint. We don't have a lot of
  284. 14:32historical data, maximum 30 to 50 years
  285. 14:35of high quality data. Um the
  286. 14:39dependencies are poorly constrained
  287. 14:41because of uh the geography, the style
  288. 14:43of geography between these two
  289. 14:45countries. And we're really looking at
  290. 14:47sort of very interesting uh novel
  291. 14:51perhaps very far out out of the box blue
  292. 14:54sky ideas where we could leverage new
  293. 14:57technologies. And in this case we're
  294. 14:59looking at quantum to improve the style
  295. 15:02and the way we produce these models in
  296. 15:04such a way that could potentially be
  297. 15:06reproduced across the bank and also
  298. 15:08deliver deliver us some form of quantum
  299. 15:11advantage. whether that be sort of
  300. 15:13improvements in efficiency or
  301. 15:15improvements in sampling or so on and so
  302. 15:17forth.
  303. 15:19So thank you. Thank you very much.
  304. 15:21>> Thank you. Yep. Now to Verdan on more
  305. 15:26technical side. Thank you James.
  306. 15:29>> Okay. Um yeah thanks a lot James for
  307. 15:31that uh quick introduction about what
  308. 15:33the challenge means. Uh so if you anyone
  309. 15:35of the participants today is coming from
  310. 15:37more quantum background or from a more
  311. 15:40technical computer science math
  312. 15:42background uh the introduction is good
  313. 15:44but you must be like wondering where is
  314. 15:46the math give me the equation and I'll
  315. 15:48you know model around it I'll build the
  316. 15:50solution around it. So that is where I
  317. 15:52come in. uh this entire problem that
  318. 15:54James described the joint loss modeling
  319. 15:56problem can be broadly described with
  320. 15:59the equation that is uh presented over
  321. 16:00here right and to take an example and uh
  322. 16:04mind you like today this would be a very
  323. 16:06quick overview of a lot of uh technical
  324. 16:08stuff but obviously as you proceed
  325. 16:10forward with a challenge and you want to
  326. 16:12work hands-on all the slides all the
  327. 16:14equations that are mentioned over here
  328. 16:15along with clear descriptions of what
  329. 16:17can be done what has already been done
  330. 16:19for you to benchmark and uh relate your
  331. 16:21resources with uh is available on the
  332. 16:24web page on the challenge description.
  333. 16:26So in fact the entire code repository
  334. 16:29has been shared by ADB. Uh it is
  335. 16:32available on uh the web page. Uh so yeah
  336. 16:35diving into the equation right what does
  337. 16:37this entire problem mean? Uh let's first
  338. 16:40consider an example problem. We see
  339. 16:42there are two countries. One is Kirius
  340. 16:45the other one is Tajak. And there are
  341. 16:48two kind of disasters that we are
  342. 16:50considering over here. So one is
  343. 16:51earthquakes and the other one is extreme
  344. 16:53rainfall.
  345. 16:55Now each of these uh countries has some
  346. 16:58kind of uh principle of three-year terms
  347. 17:01right of 80 million US which is split
  348. 17:05into 70 million for earthquakes and 10
  349. 17:08millions for rain. Now this equation
  350. 17:11over here kind of uh shapes this entire
  351. 17:13joint loss modeling into a mathematical
  352. 17:16format and and the way I look at it uh
  353. 17:19is going from inside out. So you start
  354. 17:21making sense of it. So let's maybe first
  355. 17:23fix uh the the x term over here that's
  356. 17:26the joint outcome. So a variable x kind
  357. 17:30of defines what kind of parallel unit
  358. 17:33over here uh within a certain period. In
  359. 17:36our example, that could be a coupled
  360. 17:38unit between a country and a disaster.
  361. 17:41So it could be something like kyus
  362. 17:43earthquake or rain in kyus. Uh similarly
  363. 17:47earthquake in Tajjak or rain in Tjak and
  364. 17:50and so on and so forth. Right? So for
  365. 17:53this example, those would be the set
  366. 17:55variables. But if you take up another
  367. 17:58joint loss modeling example or you take
  368. 17:59some different kind of data sets that
  369. 18:01are available publicly, uh those
  370. 18:03variables might change a little bit.
  371. 18:05That's kind of the essence over here.
  372. 18:07Now for each of these variables X uh
  373. 18:10when we take P of X that's basically the
  374. 18:12probability of that disaster happening
  375. 18:14or that peril actually occurring. Uh
  376. 18:18that probability will give you how
  377. 18:20frequent that outcome is for a given
  378. 18:23country. Ideally we would want that to
  379. 18:25be as low as possible. But from uh this
  380. 18:28principle point of view, we understand
  381. 18:30that this probability will help us
  382. 18:33decide how how we want to model our
  383. 18:36economics around that. Right? Very
  384. 18:39similar to the probability, we also have
  385. 18:42f of x over here which basically gives
  386. 18:44you a solid number for any kind of
  387. 18:47payouts that uh eventually turn into the
  388. 18:50quantity that we might care about.
  389. 18:52Right? So for example, it could be the
  390. 18:54total loss. It could be yeah just a yes
  391. 18:57and no answer for whether the countries
  392. 18:59are being paid for the disaster
  393. 19:01happening or something along those
  394. 19:02lines. Right?
  395. 19:04And um let's see what what else is
  396. 19:07there. And then we have the contract
  397. 19:08terms, right? So the big L the sum of CI
  398. 19:11is the total loss portfolio of that
  399. 19:13entire uh bond. Right? Now all of this
  400. 19:18jargon kind of starts making sense if
  401. 19:20you look at from a bird's eye point of
  402. 19:22view when we say that for each country
  403. 19:24we want to take uh different kind of
  404. 19:26probabilities of how frequently a
  405. 19:28disaster might be occurring and based on
  406. 19:31that we need to compute the loss models
  407. 19:34we need to compute the payout terms for
  408. 19:36each of these countries.
  409. 19:38uh a few more uh guidelines or a few
  410. 19:41more I would say rules for modeling this
  411. 19:44is number one uh that the earthquakes
  412. 19:48are actually paid in quarter steps so
  413. 19:50not annually they are accumulated four
  414. 19:53times a year up to the total sublims
  415. 19:58the precipitation that remains as an
  416. 20:00annual thing so all the data all these
  417. 20:02rules all these guidelines are again
  418. 20:04shared on the repository so you can have
  419. 20:07a look and make more sense of it. I know
  420. 20:10just talking about this might be very uh
  421. 20:13technically jargon for some of you and
  422. 20:14that's that's totally okay but this is
  423. 20:16just to give you all of you a reference
  424. 20:18point of what modeling this means like
  425. 20:21now once you have this entire model the
  426. 20:24good thing is that uh the benchmarks are
  427. 20:26already been provided by ADB so you can
  428. 20:28actually compute this small data set you
  429. 20:30can actually do the classical routine
  430. 20:32already and compare it with how ADB has
  431. 20:35done it or what the baseline benchmarks
  432. 20:38are and the interesting thing is that we
  433. 20:41want all of you to go beyond that. So
  434. 20:42not just replicating what is already out
  435. 20:44there but we all want to uh you know
  436. 20:47start building crazy ideas to James
  437. 20:49point where you want to uh have out of
  438. 20:52the box ideas from all of you. Uh we are
  439. 20:55sharing some uh approaches that we might
  440. 20:57feel are uh are suitable enough. There
  441. 21:00have been some past literature on some
  442. 21:02of these approaches using quantum
  443. 21:03amplitude amplification or uh say
  444. 21:07sampling methods or quantum machine
  445. 21:08learning for instance. But uh this is a
  446. 21:12challenge and you are more than welcome
  447. 21:14to use your own resources come up with
  448. 21:16some new approaches that you might want
  449. 21:18uh to try out over here and uh yeah get
  450. 21:22us solution right. Uh the end goal over
  451. 21:25here is kind of two-part. So number one,
  452. 21:27we want to see benchmarks against the
  453. 21:30classical solutions that we're already
  454. 21:31providing you with, right? So any kind
  455. 21:33of solution that you come up with, it
  456. 21:35has to have a onetoone benchmark of uh
  457. 21:38how it compares with the current
  458. 21:39solution. And number two, we want to see
  459. 21:42scalability, right? Because the data set
  460. 21:44that we're providing today and the
  461. 21:45approaches that we're providing today
  462. 21:46might work in a limited data set, but
  463. 21:49what happens when we are actually doing
  464. 21:51it at a production scale? uh what
  465. 21:53happens when we want to use quantum
  466. 21:55potentially for solving real world
  467. 21:57problems and not just you know toy
  468. 21:59challenges. So we do want to see a real
  469. 22:02validation of how scalable a solution is
  470. 22:05and to make that easier we are also
  471. 22:07providing credits through AWS. So Fabian
  472. 22:10shared a little bit on that. Uh we will
  473. 22:12be providing credits to select teams for
  474. 22:14scaling up their work right. So whether
  475. 22:16that means running it on GPU resources
  476. 22:18or running it on real quantum computer
  477. 22:21uh you let us know right uh so that is
  478. 22:24where kind of the challenge uh takes
  479. 22:26shape uh so if you move to the next
  480. 22:28slide uh sea
  481. 22:31>> yep so here are the deliverables and
  482. 22:32I'll pass it on to sa to just go through
  483. 22:35all of that y
  484. 22:37>> uh yeah sure um the registration is open
  485. 22:39as you know on the on the website um we
  486. 22:43have two phases the phase one is more
  487. 22:45about like the proposal if you can uh
  488. 22:47yeah work on uh plan of action what to
  489. 22:51expect and then of course the team will
  490. 22:52evaluate more deeply if you're um if
  491. 22:55you're okay to move to phase two and
  492. 22:56then once you actually been selected um
  493. 23:00on phase two uh you will be actually
  494. 23:01working on uh working in the prototype
  495. 23:04uh you have to provide a technical
  496. 23:05documentation analysis business case and
  497. 23:09then in the end the final report which
  498. 23:11then will be judged actually by a few
  499. 23:13judges um and I will walk you through
  500. 23:15the timeline uh later in the in this
  501. 23:18workshop. Uh just to be mindful of time,
  502. 23:21I think we need to speed up a little
  503. 23:22bit. So I know there are a few more
  504. 23:23slides. Verdon, maybe you can um walk
  505. 23:26through in terms of evaluation itself
  506. 23:29and then we want to have more time for
  507. 23:31any questions that people have.
  508. 23:34>> Great. uh so now we expect a lot of
  509. 23:36submissions to come our way and that is
  510. 23:38always a good problem to have but uh to
  511. 23:40make it very transparent and to make
  512. 23:42sure that there is a clear guidelines on
  513. 23:44what you might expect from a
  514. 23:45participant's perspective
  515. 23:47>> uh this is what the evaluation criteria
  516. 23:48looks like right so the first one would
  517. 23:50be problem and approach uh it should
  518. 23:53have a clear onetoone matching with our
  519. 23:55benchmark so that is it's it's you know
  520. 23:57easier for us to also evaluate u we are
  521. 24:00keeping the challenge pretty open but
  522. 24:02that does not mean that we are straying
  523. 24:04away from the challenge at hand, from
  524. 24:05the problem statement at hand, right?
  525. 24:07And then obviously uh how does your
  526. 24:10approach, how does the solution match
  527. 24:11with the problem? Uh we believe me, we
  528. 24:15we have heard many times a lot of people
  529. 24:17claiming that hey quantum can do x y and
  530. 24:19zed that's all good but uh how much of
  531. 24:22that is really scalable, right? Uh so
  532. 24:25getting that proof of concept ready with
  533. 24:27clear prototypes with clear results uh
  534. 24:30will be the second metric that we'll be
  535. 24:32judging this on. Uh then obviously comes
  536. 24:34the technical trigger where we will be
  537. 24:36you know diving deep and trying to break
  538. 24:38your code and and understanding what
  539. 24:40works what doesn't work. Uh so make sure
  540. 24:42that it's uh technically solid whatever
  541. 24:45you end up submitting. Now beyond that
  542. 24:48in the first phase of submission we also
  543. 24:51want to see proof and and validation of
  544. 24:53whether your team has what it takes to
  545. 24:57continue into the next phase of our
  546. 24:58entire challenge. So we need to
  547. 25:01understand what your work plan looks
  548. 25:02like right this could include a road map
  549. 25:04this could include I don't know like a
  550. 25:06GAN chart with clear milestones clear
  551. 25:09risks what you expect to happen good but
  552. 25:12also you should have a clear view of
  553. 25:14what uh might not work out right and
  554. 25:16what are some uh backup options that you
  555. 25:19might want to consider already uh so
  556. 25:22getting that work plan ready would be
  557. 25:23really helpful for us to evaluate you as
  558. 25:25a team moving into the next phase and
  559. 25:28similar to work plan your resource those
  560. 25:29estimations. So this ties back with my
  561. 25:32earlier point of saying that all of the
  562. 25:34teams or at least the selected teams
  563. 25:36would be getting hardware credits to
  564. 25:38scale up their resources, right? So how
  565. 25:41much hardware credits do you want?
  566. 25:42Right? Do you need GPUs? Do you need
  567. 25:43CPUs? If you want quantum processors,
  568. 25:45what kind of QPUs do you want? Do you
  569. 25:47want to run it on superconducting
  570. 25:49devices? Do you want to run it on uh
  571. 25:51some of the simulators available? Do you
  572. 25:52want to run it on neutral atom devices?
  573. 25:54Right? you you need to uh let us know
  574. 25:56with clear justification of what you
  575. 25:58want and why you want it and why would
  576. 26:02uh giving you this kind of resources
  577. 26:03make your solution successful so that
  578. 26:06our partners like AWS can kindly grant
  579. 26:08you those kind of credits and make the
  580. 26:10scalability possible
  581. 26:12>> perfect
  582. 26:12>> and uh finally your team skills right we
  583. 26:16don't have kind of any uh barrier over
  584. 26:19here anyone and everyone can participate
  585. 26:20but again we do want to look uh at your
  586. 26:23skills what you're bringing to the table
  587. 26:25to make sure that if you're
  588. 26:26participating, we can bet on you to be
  589. 26:29the best team possible.
  590. 26:31>> Thank you.
  591. 26:33And uh maybe James, do you want to
  592. 26:35finish off with the most important slide
  593. 26:38in terms of the timeline so people do it
  594. 26:40on time? Yeah, thank you.
  595. 26:42>> Yep. Thank you. Uh so we have now just
  596. 26:45less than a couple of weeks uh for
  597. 26:47registration submission. Registration is
  598. 26:49via the ADB challenges websites which I
  599. 26:51will share a post uh a link in the uh in
  600. 26:58the chat shortly. I know there's also
  601. 26:59been some questions around the GitHub
  602. 27:01page. We won't be sharing GitHub page.
  603. 27:03We'll be sharing the raw repository with
  604. 27:06you for you to download. This is a zip
  605. 27:07file and then uh from 23rd 30th so the
  606. 27:11towards the end of this month there'll
  607. 27:12be short listing of the selected teams
  608. 27:15for the challenge. So we'll be choosing
  609. 27:16between five and eight teams and the
  610. 27:18expectation is for the following weeks
  611. 27:21uh there will be development up until a
  612. 27:24pitching event that will be hosted in uh
  613. 27:27January late January next year.
  614. 27:30Um there's a prize uh or
  615. 27:36seed funding attached to this of $20,000
  616. 27:39to one selected team. Um and we're
  617. 27:43looking for submissions anywhere from an
  618. 27:45individual to uh five members uh or be
  619. 27:50associative with an organization or
  620. 27:51more. Um and all submissions please make
  621. 27:54sure they are in English. Um thank you
  622. 27:58very much and yeah uh we will be sharing
  623. 28:01this repo uh very very shortly. If you
  624. 28:04have any questions, uh please feel free
  625. 28:06to reach out to myself or Kai and we'll
  626. 28:09be happy to uh make sure that they're
  627. 28:13addressed and uh provide any assistance
  628. 28:16if if required in getting yourselves set
  629. 28:18up and running. So yeah, please go to
  630. 28:22the link um application link is there.
  631. 28:24I'll post something in the chat as well.
  632. 28:26But thank you.
  633. 28:27>> Thank you. Thank you. And I think now
  634. 28:28it's the perfect time to move to any
  635. 28:29questions anyone has. And now let me
  636. 28:32just uh go to the um
  637. 28:36stop sharing. I'll just go to I think
  638. 28:39there are a few questions that were
  639. 28:40asked
  640. 28:42uh in the chat. Uh so yeah you Eugene as
  641. 28:44I said yeah have you're free to see if
  642. 28:47anyone wants to um wants to join you. Um
  643. 28:54and
  644. 28:55um let me see could we share the link?
  645. 28:57Yeah, the link will be shared now um by
  646. 29:01uh James, but the GitHub GitHub page
  647. 29:04will be actually uh shared on the
  648. 29:07website itself, right? I think there's a
  649. 29:08link to that, James, right? Correct me
  650. 29:10if I'm wrong.
  651. 29:12>> Uh we we won't be showing the GitHub
  652. 29:13page. U
  653. 29:14>> just in complications, we'll be sharing
  654. 29:16the get the repo there and the challenge
  655. 29:19page. If you scroll to the bottom of the
  656. 29:21challenge, should be a challenge
  657. 29:22briefing. And uh if you're interested,
  658. 29:24please make sure to register on that
  659. 29:27platform uh on the challenges ADB
  660. 29:30platform. Sign up, create an account,
  661. 29:31and then uh register your interest in
  662. 29:34this challenge.
  663. 29:36>> Yeah, perfect. Thank you. Thank you. So
  664. 29:38yes, we are open for questions. We have
  665. 29:41around 9 minutes left of this webinar.
  666. 29:43So anyone please ask questions.
  667. 29:48Feel free to unmute yourself and jump
  668. 29:50in. Thank you.
  669. 29:55Hi, this is Abijit Basu. Um, and we are
  670. 29:58already working on a Himalayan glacious
  671. 30:00problem of similar nature. Um, we
  672. 30:02haven't um so we are trying to figure
  673. 30:05out what's the ultimate application that
  674. 30:07you're targeting. So quantum mechanics
  675. 30:10problem and statistics is great but
  676. 30:14curious is a is a very typical to a
  677. 30:16particular subject matter whether it's
  678. 30:17parametric insurance or bond. So do you
  679. 30:20guys have a clear line of thought where
  680. 30:22these applications are going to be
  681. 30:24because based on that how you create
  682. 30:26this use this uh quantum mechanics to
  683. 30:31solve a problem could be very different.
  684. 30:35>> Yeah. Uh so if you don't mind I'll I'll
  685. 30:37take this question. So look ultimately I
  686. 30:40think the goal for us is as a financial
  687. 30:42institution is we want to provide better
  688. 30:44products for our customers. Uh that's
  689. 30:47really sort of summary of what our goal
  690. 30:49is at the end of the day. Um so the way
  691. 30:52we look at this is this this is a
  692. 30:54potentially a way for us to improve the
  693. 30:56quality of the models that we produce
  694. 30:58and with that comes some advantages to
  695. 31:01the to the products that we can offer.
  696. 31:03Potentially it could be uh tighter
  697. 31:05spreads uh or better pricing efficiency.
  698. 31:08Um and to us it becomes sort of also a
  699. 31:11question of better risk, better
  700. 31:12controls, um better sort of the
  701. 31:14fiduciary uh understanding uh of the
  702. 31:18relationship between the products we're
  703. 31:19offering and the customers we're
  704. 31:21serving. So
  705. 31:23I mean for us we're we're looking at uh
  706. 31:26insurance and the insurance industry.
  707. 31:28We're specifically looking at insurance
  708. 31:30link securities.
  709. 31:32Um so we lots of information online
  710. 31:34about how insurance link securities are
  711. 31:37structured. Um and also uh if you go to
  712. 31:41Arteimus actually it's a very good
  713. 31:43industry website you'll be able to see
  714. 31:45recent insurance link securities
  715. 31:47offerings. Um and with that you can also
  716. 31:49see some of the documents that might be
  717. 31:51attached to them that describe uh the
  718. 31:53triggers the type of triggers um the
  719. 31:56tail risk all of these different sort of
  720. 31:58parameters.
  721. 32:00Um so that's that's sort of I'd say the
  722. 32:02primary use case that we've attached to
  723. 32:04this. The sort of secondary use case is
  724. 32:06can we improve our internal models at
  725. 32:08the bank which uh have similar or share
  726. 32:11similar attributes to these joint loss
  727. 32:13modeling problems and that could include
  728. 32:15spillover um in bond markets which I
  729. 32:18think is a very good example. Say you
  730. 32:20have a default that triggers in country
  731. 32:22in one country, country country A. Uh
  732. 32:24what is the sort of codependency or the
  733. 32:27the potential joint loss um and the
  734. 32:31spillover impact on countries B, C, D
  735. 32:34and E who might be holding um some bonds
  736. 32:38in country A. So you very quickly have
  737. 32:40these very large um graphs uh problems
  738. 32:43that are systemically quite challenging
  739. 32:45to model um using traditional techniques
  740. 32:48like Monte Carlo sampling. And that's
  741. 32:51where we were looking at sort of is
  742. 32:53there some approach here that is
  743. 32:55potentially more costefficient uh
  744. 32:58potentially uh provides some insight
  745. 33:01some breakthrough using a quantum
  746. 33:03technique that we can't uh achieve using
  747. 33:06just say a high performance computer or
  748. 33:09a locally run model
  749. 33:12>> we'll be leaning
  750. 33:14>> oh sorry please
  751. 33:15>> yeah so followup question so what I'm
  752. 33:18trying to do is so what we have disced
  753. 33:20so far is the mathematical side of the
  754. 33:23statistical side of it, quantum side of
  755. 33:24it. But ultimately this is going to be
  756. 33:26applied in a use uh business
  757. 33:28environment, right? Which is the
  758. 33:30financial risk or financial uh
  759. 33:32underwriting. Now there are two pieces
  760. 33:35to it. One is this models themselves are
  761. 33:37proprietary sometimes like once you
  762. 33:39build the model it has IP attached to
  763. 33:41it, right? So uh what is the IP
  764. 33:43protection that we get of sharing our
  765. 33:45model? one. The second is when you apply
  766. 33:48to the business outcomes there is a
  767. 33:50there is a basically a benefit to the
  768. 33:52organization ADB or otherwise how do you
  769. 33:55comp compensate us for that kind of
  770. 33:58model. So we have already created this
  771. 34:00models for parametric insurance. So when
  772. 34:02you create a model for parametric
  773. 34:03insurance is very different than
  774. 34:04underwriting a bond because the
  775. 34:06parameters are very different right the
  776. 34:08the facilities are very different the
  777. 34:10policies are very different. So if you
  778. 34:12could help me understand the first two
  779. 34:14questions and then kind of the third
  780. 34:15question is how well are you prepared to
  781. 34:18take quantum mechanics into your policy
  782. 34:20framework that ADB has if you have not
  783. 34:23used quantum mechanics and I'm I'm
  784. 34:24assuming you don't have the policy
  785. 34:26framework for that and how do you plan
  786. 34:27to inform it uh because doing the math
  787. 34:30is the simplest part of it getting the
  788. 34:33business side uh with the IP protection
  789. 34:35covered is second piece of it and third
  790. 34:37piece of it is how you support it with
  791. 34:39policy framework that your organization
  792. 34:40is ble to take advantage of it.
  793. 34:42>> Yeah.
  794. 34:43>> Okay. So, the first question I I won't
  795. 34:45answer myself, but I can certainly have
  796. 34:47a uh colleague follow up. Um
  797. 34:52so, if you actually if I don't have your
  798. 34:54email already, if you could just share
  799. 34:56your email with me, we can I can share
  800. 34:57with you sort of some background on
  801. 34:59that. There should be some information
  802. 35:01available on the ADB challenges website
  803. 35:03around IP.
  804. 35:06The second question um look ultimately
  805. 35:08if if the solution is good, if it's high
  806. 35:10quality, if we see a clear advantage,
  807. 35:13there is potentially a pathway where we
  808. 35:16could look to deploy the solution
  809. 35:19inhouse. Um that is not without some
  810. 35:22complexities. Look, it's not just
  811. 35:25something you can just take a new
  812. 35:26service and bolt it into a into a large
  813. 35:29enterprise environment. But um look we
  814. 35:32have the purview that if solutions are
  815. 35:34proved to be useful then there and there
  816. 35:37is an interest across um the
  817. 35:39organization then uh we would seek to
  818. 35:42engage on something longer term uh
  819. 35:45potentially um but again this is not a
  820. 35:48given this is not a guarantee this is
  821. 35:50shouldn't be sort of the basis of an
  822. 35:52expectation of this project. Um this is
  823. 35:54just uh you know one of the many
  824. 35:56potential avenues uh that uh that could
  825. 36:00be open to uh potential uh participants
  826. 36:03of this challenge. Um on your third
  827. 36:07question look we're our our sort of
  828. 36:10journey in quantum is still sort of
  829. 36:11relatively early um and this technology
  830. 36:14is very nent as well. So we're very much
  831. 36:18of the view we have a a very strong
  832. 36:20curiosity to engage and understand uh
  833. 36:23across our organization how quantum
  834. 36:25could be leveraged uh to provide an
  835. 36:27advantage to our operations and services
  836. 36:29that we provide as an institution. But
  837. 36:33uh you know today I'd say we're maybe
  838. 36:37not at the level degree of maturity of
  839. 36:39other technologies such as machine
  840. 36:41learning for example and AI where we
  841. 36:44might have mature and sophisticated
  842. 36:46policies. But I think part of this
  843. 36:48process is to work in harmony with those
  844. 36:51different different uh sort of uh
  845. 36:55directions which is can we find useful
  846. 36:58applications for this technology? Can we
  847. 37:00find interested business units
  848. 37:02internally? Um, and all three of those,
  849. 37:05you know, that alongside internal policy
  850. 37:07would move in lock step together as we
  851. 37:10build confidence, understanding, um, and
  852. 37:12as well as prove out the the usefulness
  853. 37:15of this technology. So that's something
  854. 37:18that we we expect will happen over the
  855. 37:20coming years as we continue to get more
  856. 37:22comfortable with uh with deploying
  857. 37:24quantum in operations and with any new
  858. 37:27technology because this is such a
  859. 37:29critical application I would expect
  860. 37:31there to be lots of questions around
  861. 37:32sort of you know how does this work is
  862. 37:34it an interpretable technique um
  863. 37:37mechanistically you know how does it
  864. 37:39work under the hood and because
  865. 37:41colleagues will be I think with any new
  866. 37:44technology skeptical to understand
  867. 37:46schedule bet technology until they fully
  868. 37:48understand uh the uh the fundamentals.
  869. 37:52>> Yeah, thank you James. Uh just to be
  870. 37:54mindful of time, we have one minute left
  871. 37:56uh for this webinar. Um so if anyone has
  872. 37:59any more more questions, you're more
  873. 38:00than welcome to email us. Um the webinar
  874. 38:04will be shared online on the website and
  875. 38:07also our emails in case and um also we
  876. 38:11are thinking maybe doing a follow-up uh
  877. 38:13webinar to have more questions like this
  878. 38:15answered and more deep dive because I
  879. 38:17know we were quite short in time but
  880. 38:18this is more of a kickoff and uh yes
  881. 38:21we're super super happy that everyone
  882. 38:23was able to join I know due to time
  883. 38:25differences and thank you everyone
  884. 38:27participating. Thank you Fabian, Vardan
  885. 38:30and James for your inputs and yes this
  886. 38:33is just the beginning and we are very
  887. 38:36happy. So yeah any questions just yeah
  888. 38:38let us know I can um yeah take them on
  889. 38:41and yes thank you thank you everyone um
  890. 38:44for your time. Thank you. Speak speak
  891. 38:46later. Thank you.
  892. 38:48>> Thank you.
  893. 38:49>> Thank you. Bye bye. by

About this transcript

This page contains the full transcript of Challenge Briefing and Live Q&A by ADB Digital Sandbox, generated from the public captions YouTube serves with the video. The transcript has 6,056 words across 893 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.