In this Member Workshop with Mphasis, gain a deep understanding of quantum computing and explore its transformative potential in tackling business challenges with quantum AI, optimization and simulation solutions.
Event Information
In the fast-paced world of business, conventional computers encounter challenges when managing extensive and intricate systems. Quantum computing, a cutting-edge field of computation, introduces a ground-breaking approach to complex problem-solving for businesses. Leveraging computationally efficient methods, quantum computing promises transformative breakthroughs across various domains. From optimizing supply chains and designing innovative materials to enhancing operational efficiency and cost savings, this revolutionary technology positions businesses as pioneers of innovation, appealing to a diverse range of industries.
In this webinar, you'll acquire an in-depth comprehension of quantum computing and discover its transformative potential in addressing business challenges through quantum AI, optimization, and simulation solutions.
Key learnings:
- Explore quantum computing and its versatile applications in various industries.
- Recognize the significance of quantum computing in shaping an effective data strategy for your organization.
- Develop a clear roadmap for the adoption of quantum computing, enabling you to harness its potential effectively.
- Explore real-world use cases understanding its practical applications and benefits.
Workshop Video
Thank you for your interest in our event! The event has now passed, but we are pleased to share that the recording is available for you to review. Catch up on all the valuable insights and information shared during the event. We hope you find it informative and useful.
0:00
okay well thank you everyone for attending today's Workshop uh my name is Taylor Kern and I'm the senior coordinator of member engagement here at
0:06
the chal Board of Trade thank you for joining us this afternoon at our value membered emphasis workshop on
0:11
transforming business using Quantum Computing in this workshop with emphasis you will gain a deep understanding of qu
0:18
Quantum Computing and explore its transformative potential in tackling business challenges with Quantum AI
0:24
optimization and simulation solutions before we begin I'd like to acknowledge
0:29
this land that we are meeting on which is home to diverse n First Nations Inuit and mate people though you could be
0:35
joining us or watching from anywhere the board's offices are located on the traditional territory of many indigenous
0:40
Nations they share with us a sense of responsibility intergenerational Equity the well-being of today and tomorrow and
0:47
now without further Ado I would love to pass it over to VC and roit who are going to be today's hosts on your
0:56
Workshop thanks Taylor hello everyone I'm Rohit so uh what I'll do is I'll share my
1:03
screen just give me a
1:09
minute I hope everyone is able to see my screen tayor could you please confirm
1:14
that yeah just got to put it in presentation mode there there you go
1:20
perfect yeah so uh uh hello everyone once again so as uh Taylor mentioned I'm
1:27
going to present Quantum Computing uh uh and its applications for Industries
1:34
today so uh just a brief about me I have around 14 years of experience with eight
1:40
plus years in data science currently I'm leading the quantum Computing initiative at emphasis next slaps and I hold a
1:48
bachelor's degree in electronics in communication and a MERS in Business Administration I have developed a lot of
1:55
AI and Quantum Computing based Solutions and delivered client projects in areas of pfsi Logistics uh among
2:02
others so the topic for today now so this is the agenda that is what I'm
2:07
going to do first I'll just a brief introduction to the team I am part of emphasis nextlabs then we'll jump into
2:14
Quantum Computing what and why where do you apply Quantum Computing then what is
2:20
the state of quantum Computing today and finally what we we at emphasis are doing an area of quantum
2:26
Computing so first of all the introduction to emphasis next slaps so
2:31
uh next slaps is emphasis R&D um for for
2:37
Research into disruptive Technologies and taking it to the market so you focus in areas of cognitive cloud and Quantum
2:44
Computing on the left hand side you'll see a lot of solutions that we have built in uh the traditional AI that is
2:50
uh uh that is like Solutions like mlops Solutions like customer 360° analysis
2:57
process Mining and others these solution Solutions are available and we have used these to solve our customer problems
3:04
then next more recently we are focusing on generative AI in fact we have launched a unit called emphasis. a which
3:11
is our business unit to address AI related problems here we are working on large language driven models we are
3:18
looking at synthetic data generation we are looking at ml governance issues which are coming up across the world
3:25
given the use of decision making through AI under Computing we are looking at uh
3:32
uh algorithmic development and deployment assessment and Consulting service ml optimization and process
3:38
simulation driven Solutions so I'm going to talk about this today since we are an
3:43
R&D driven lab uh we do a lot of thought leadership we have around five awarded patents we have 20 plus a Quantum
3:50
Computing white papers published and since we also address customer use cases
3:55
we have a design thinking driven Consulting model which allows our customers to identify rank prioritize
4:02
and execute Quantum Ai and Quantum use cases uh we have a lot of Partnerships
4:08
with Academy and Industry for example we are partners with AWS assure we are
4:13
premium Premier partners with them for the AI side of things we have a partnership with University of Calgary
4:20
here where we are the anchor business partner for the quantum City initiative they have taken uh other than that we are also
4:26
partners with Alberta machine intelligence Institute in of area of uh solutioning for uh business problems now
4:35
uh generally our output consist of uh customer projects uh industry oriented
4:41
Solutions and Marketplace listings we have around uh we have around 250 plus a
4:48
Quantum Computing Marketplace listings so if you go to AWS generally you will see we are in uh top the top three
4:56
always so you can use these solutions to solve your business problems or if you want to scale your business problems you
5:03
can reach out to us by looking at what is the art of the possible this is a brief introduction on
5:08
next laps now the topic for today Quantum Computing what and why so uh I
5:14
will take a drill down kind of an approach here where I'll just explain what Quantum technology in general is
5:21
and then I'll go into the Computing side of things so bear with me so the idea is that the first Quantum revolution has
5:28
already taken place once we were armed with the knowledge of behavior of subatomic particles and structure of the
5:33
atom a lot of new technologies
5:40
were which basically got translated into integrated circuits which created your computers uh you know atomic clocks for
5:48
GPS uh you know about lasers very very important MRI machines electron
5:54
microscope all these were made possible because uh the physicist identified the
6:00
behavior of subatomic particles now uh more recently we have
6:05
moved into the second quantum Revolution it started around say late uh early 80s
6:11
where somebody uh thought of using quantum mechanics to build technology uh
6:17
a different type of technology so if you look at it uh it is very if I have to make a comparison it is very similar to
6:24
our current digital electronic technology where the idea is to build sensing machines computation machines
6:31
communication networks and encryption systems you know about the internet today you know about our RS protocols
6:37
and all these systems generally fulfill the human need to communicate to compute
6:44
to encrypt messages and to measure things same idea is taken from a
6:50
classical domain to a Quantum domain where we are again trying to build similar things basically sensors
6:56
computers communication networks and encryption systems but the underlying technology that we are going to use is
7:02
quantum physics quantum mechanics rather than the classical uh classical physics
7:07
of today Quantum Computing is a part of this
7:13
revolution the other being the sensing the encryption and communication the difference generally
7:19
lies again in the way uh the computation is done it is it uses certain Concepts
7:26
like superposition entanglement interference and tunneling to solve certain problems with a significant
7:31
performance Advantage we have to keep in mind that a quantum computer cannot solve all problems that you have today
7:38
with the same efficiency or very much efficient than the current computers
7:43
that we have so it is built for specific problems now uh we talked about Quantum
7:50
technology a little I want to talk about Computing now so uh if you see on your
7:55
screen right now I have put uh different typ type of computers that have been
8:01
conceptualized or created over a period of time so if you have to understand
8:06
Computing the the most familiar object that you can look at is the human brain
8:11
the idea is that human brain looks at the real world takes that information and be it images be it sound be it be it
8:20
tab data or whatever it creates a picture uh it creates an abstraction for
8:25
the same which our brain can understand and process the same now keeping this idea of data representation and data
8:32
processing in mind different machines can be able to do the same task so you
8:38
can talk think about searching uh an item in a database or sorting a list or
8:45
building an AI machine all these tasks can be done using different computers we
8:51
have to keep in mind that Computing is a physical process you may write a software program to execute your
8:57
competition but the end of the day there should be a physical substrate underlying physical substrate which does
9:03
the Computing so in Alternative forms of computing in DNA Computing for example you make use of DNA strands to encode
9:10
information and process them you can build a search machine using a DNA computer similarly with chemical
9:17
computers or bages mechanical computers some some similar things can be achieved
9:22
but more recently we have been looking at uh electronic Computing rather than
9:27
the other forms and here we have taken a path since 1950s let's say where we have
9:33
moved from vacuum tubes to transistors to integrated circuits which have now we
9:39
have converted into uh your CPUs your graphic Processing Unit your dsps and
9:44
others and then we have built super computers using the same thing the most
9:50
recent Advent of this is a quantum computer uh it uses a different form of
9:55
data representation and a different form of processing Elements which gives us certain performance advantages for some
10:03
type of problems now the advantage is exponential in nature things that may
10:10
take say a million years on classical computers for certain tasks can be done
10:15
in milliseconds on a quantum computer so we have to keep that in mind for certain problems this Advantage is uh available
10:22
and hence this area is being currently pursued now look coming from qu
10:29
Computing to Quantum Computing these are some key ideas that you have to look at the idea of representation of
10:35
information the information that is available to you in the real world will always in comes in form of images text
10:43
sound tables whatever it is but the different computers represent it in
10:48
different way in case of our digital computers that we use today it is a zero and one representation that is a binary
10:54
representation your images each of those pixels get converted to a number
10:59
that number can be converted to a binary formulation which is then processed in
11:04
case of quantum that representation does not take only zero and one levels but a
11:10
combination of the two this increases the way the uh information density as
11:16
well as the algorithms that process that information their capability for certain type of problems so this Cubit
11:22
representation of information is very important it is uh I generally uh look
11:28
at these kind of representation in form of a language your quantum computer
11:33
understands a different language from a human computer like me versus a digital computer versus a DNA computer so that
11:40
information May remain the same but the representation of information is different which gives it particular
11:46
advantages from uh execution stand of the processing point of view you can look at superposition entanglement and
11:53
interference as things which allow you to do faster processing in certain cases
11:58
now these are are complicated topics in themselves so I won't jump into it but you might have seen interference of
12:06
interference in the Leakes right so that concept if brought down to the atomic
12:11
level allows you to build computers that's the main idea now the second idea is there are
12:17
multiple type of quantum computers that exist some are Universal in nature that
12:22
is you can solve almost any problem on it uh and for example you want to build
12:28
a sorting machine a searching machine uh uh AI machine any sort of machine can be
12:33
built on those these computers are gate based very similar to what we have today that we use today your laptops your
12:39
supercomputers your mobile phones everything but there are other ones as well which makes use of quantum physics
12:46
at the back end but solve a lesser number of problems so Quantum anas is one type which mostly is focused on
12:53
sampling and optimization kind of problems that occur in everyday life
12:59
now there are different camps of quantum Computing uh when I say camps the idea
13:05
is that Quantum Computing is still a new field it is progressing rapidly uh hence
13:11
it we can divide it into two parts one is the useful quantum computers of today they are available today you can make
13:16
use of them and you can solve certain problems these are called noisy intermediate scale Quantum that is
13:22
they're noisy in nature and they have they are smaller in size uh and second is the quad inspired
13:29
part of it these are not quantum computers but classical computers which are inspired by uh Research into Quantum
13:37
uh uh Computing so these gives you some Advantage it is not 1 million years into
13:43
1 milliseconds but say 10 years into four years or that sort of a thing now
13:49
in the future once we solve the problem of noise as well as the scale that we are looking at we can get exponential
13:56
power the 10 the 1 million years into to one uh uh to 1 millisecond these are
14:01
called fall tolerant large scale and this path is progressing very rapidly
14:07
people are able to uh they have made significant Pro progress in error correction as well as stability of
14:14
things so it is not too far away in the future now what if they become available what will change the impact we have to
14:21
know from business standpoint from societal standpoint from other standpoints right it's a disruptive
14:27
technology it will change the world completely think about uh when the classical computers came into existence
14:34
or when internet came into existence sorry this is the level of
14:40
disruption we are looking at in fact more than that uh the gains will be exponential in nature if you can think
14:45
about uh things running in one millisecond you can do drug discovery on
14:50
click of a button which takes years today you can discover new materials which have properties which can take you
14:57
into say you could through which you can build batteries or systems which can take you far away into space which is
15:04
not possible so those kind of Discovery gains are possible and then it it will impact core systems we not looking to
15:11
solve peripheral problems but very core problems like portfolio optimization like last M root optimization Network
15:18
design drug Discovery so very significant technology when you look at the type of use cases we can address
15:24
using this technology I'll go to the next slide so uh where to apply it so before
15:33
we talk about where to apply Quantum Computing what are the current Focus area so not each computer is built to
15:39
solve each problem so if you look at your uh gpus today right uh they are
15:44
used specifically for graphic processing or running AI workloads if you look at neuromorphic Computing they are building
15:52
uh building processes which address a specific task your game Gamers have
15:58
physics uh chips available or audio chips available to process sound or do a
16:04
computation on physics of objects similarly so all these becomes co-processor quantum computer will also
16:10
be a co-processor and right now the areas that we see where most focus is
16:15
given is an area of uh machine learning optimization and uh simulation so
16:22
simulation when I say it has different connotations and different scenarios but I'll say it is a natural process
16:28
simulation where you can look at uh systems where atoms interact to produce results so
16:34
drug Discovery uh uh chemical reactions all these come into this
16:40
area now what is industry doing in this area right so we talked about the focus
16:46
of algorithms and machines and every everything but is industry adopting
16:51
these doing something with these computers today so yes is the answer so I'm I sorry divided this slide into
16:59
three areas Finance Logistics and others in the area of Finance if you look at the slide here uh a field called Quantum
17:07
banking has itself come into existence and you can read about it but uh the use
17:13
cases that are considered and the same as enhanced credit screening was taken by Trade Tech portfolio optimization was
17:19
taken by namura B and BBVA Goldman saxs looked into security pricing and so on
17:25
and so forth one of those case studies from BBV where they used a machine from dwa which
17:32
is a Quantum Anila they used the hybrid solver of it that is a classical qu the machine the
17:39
number of portfolios that was possible of the data was in tune of 10 to the^ 382 now this is a huge number you have
17:46
to find a few portfolios which give you the best risk return profile not easy
17:52
for a classical computer to do the same so around 171 seconds of run time
17:59
uh was taken by building the solution which is a very small time when you look at the possibilities we found uh they
18:05
found portfolios of around 60% return with 15% risk and they achieved a sharp
18:11
ratio of around 12 for certain portfolios these numbers are huge when they used the same classical some
18:18
classical Solution on the same data their range of returns was minus 20 to plus 20% for a risk of 15% so you can
18:26
see that better risk return portfolios can be identified using this using
18:32
programs built on this technology sorry so now I'll go into the
18:39
logistics side of things and I'll take two case studies one is from Toyota systems they wanted to solve the
18:44
automotive part supply chain optimization problem more than 3 million routes were there to their dozens of
18:51
factories and they wanted to F the most efficient routes they use the quantum
18:56
inspired technology here it's called called digital Ana similar to a Quantum Ana and this is provided by fuju there
19:04
was around 30 minutes of runtime and they said that it could cause a potential reduction in cost of 2 to 5%
19:10
which is a big number given the size of the operations other one comes from Los
19:15
Angeles Port of Los Angeles basically where they wanted to optimize operations
19:21
in their third largest terminal P 300 they handle more than 1 million containers per year and 5,000 trucks
19:28
today by building a Quantum Computing driven solution they were able to increase deliveries by 62% per day per
19:36
crane to their dozens of factories and double their throughput and almost half
19:42
their RTG movement other interesting one is performance modeling of jet engine
19:47
designs that was taken up by rolls D they want to achieve Net Zero carbon emissions by
19:53
2050 they looked at a hybrid methodology GPU and qpu and this is happening both given the size of quantum computers
20:00
today they are being used in hybrid mode with uh other processors like gpus which
20:05
allows them to build scalable Solutions now this is a very complex solution that
20:11
they built using uh nvidia's GPU based framework there was around 10 million
20:16
layers and 39 Cubit Quantum circuit it's basically a gate based system that they
20:21
M use of to solve this problem so they are also looking into it and you can look at a a lot of use cases here drug
20:27
development realtime risk modeling Waste Management all those things now where is quantum Computing
20:34
today so before we jump so somebody is trying it out it does not mean that this
20:39
will be the future but if we look at some indicators it's possibly will'll
20:44
make sure that uh you have it available for your uh uh own execution as well as
20:50
people are putting money into it so the first angle I will look at is the strategy and Investments going into Quantum Computing so it has been
20:56
estimated by McKenzie that it will be around $16 billion by 2040 this is whole
21:02
Quantum economy not just Quantum Computing this includes sensors telecommunication networks encryption
21:08
systems everything uh governments more than 10 countries have invested over 1
21:14
billion USD in this area uh 250 Quantum competing startups and uh I have been
21:20
launched and 36 of those were launched in 202 to itself from 2020 the rise rise
21:26
of uh number of Investments the number of companies all has been exponential
21:31
due to the Advan advancement in research sorry there was an startup
21:37
investment of around $5 billion USD and uh uh from 2020 to 22 and seven deals in
21:45
2022 were greater than $100 million so VC's governments everybody's taking
21:50
notice in investing here uh universities are also taking notice where 50 Quantum comput Quantum technology master's
21:56
degree program and 180 universities have launched Quantum technology research groups and uh there has been
22:04
establishment of R&D driven Quantum ecosystem such as Quantum sitting so these things indicate that things are
22:10
moving fast and in the right direction now what is coming out of it if you look at it we are rebuilding the whole Tech
22:16
stack we looking at the hardware the middleware the system software the application software so through this
22:22
research a lot of money is going into Hardware today all these companies that are listed here some already known some
22:28
new have built quantum computers with atom Computing just few days back they
22:33
launched a 1,180 cubit computer uh recently to give a uh to to put some value to that number
22:43
uh 50 Cubit machine if you look at it will surpass a superc computer today 50 Cubit properly built machine where all
22:51
the cubits are connected to each other it can superp pass a super computer but since they are noisy since they are not
22:58
well connected 1,180 is still a leap but still behind what we can do today uh now
23:05
Hardware is being done system and application software is being developed by fisu CQC Nvidia and others and all
23:12
this is made available over Quantum compute as a service providers so Azure has its own service IBM has its own
23:19
service AWS has Zeno and so on so you can see that an ecosystem is coming up
23:24
money is flowing in you can make use of these systems to build Solutions
23:29
today now what we are doing with Quantum Computing at emphasis is an interesting question so we started this area around
23:38
three and a half four years back and our goal uh since it was started in nextlabs
23:44
the goal was to take these this disruptive technology to the market and we did when we did that uh we looked at
23:51
these uh for these seven areas specifically I'll talk about a few so
23:57
our Focus has has already been applied R&D driven so we look at uh the fundamental research that is being done
24:03
in this area and how we can convert that to uh softwares or applications which we
24:09
can uh take to the market we have built in that ve we have built industry Focus
24:14
Solutions we have done Partnerships with Academy and Industry we have written a lot of white papers we have done a lot
24:21
of webinars and we have also created a design thinking driven engagement model which allows customer to adopt Quantum
24:27
computing our industry Solutions will fall in these areas optimization AI simulation
24:34
and finally we're partnering on cyber security cryptography and other things now this is our solution
24:41
portfolio so it is mostly Quantum machine learning and optimization portfolio and if you look at it uh the
24:49
solutions that we have built are in area of logistics supply chain Life Sciences
24:54
uh the airlines all these areas so last my root optimization is one where we
24:59
look at the delivery of packages to the end customers from a Depot from FedEx
25:05
ups and others portfolio optimization we have applied it to uh risk building risk
25:10
return profile for stocks from dow30 S&P 100 NASDAQ 100 brain tum detection we
25:17
have applied Quantum machine learning for image processing similarly we are looking at each feature selection for
25:22
high dimensional ml it is a complicated process today if you want to run High dimensional machine learning model
25:29
sorry because it consumes a lot of compute power and that uh basically
25:36
increases the runtime and computer requirements for the same and you may not get the right quality of result
25:42
Quantum provides a method to do to identify the right features to build a more accurate and more generalizable
25:48
model so a lot of things we have done Airline creu scheduling is another one where we look at how we can uh allocate
25:55
Cru to how we can GR group different flights together and then assign crws to them at minimal cost while maintaining
26:02
all the constraints that coming in it is a very complicated problem and we have built a solution around it as well so what I'll do for now is I have
26:10
more slides but before I keep talking I'll show you two demos at least that we have built uh internally and so that you
26:18
can get uh what we are trying to build using this technology it is very similar to what we build today but just using
26:25
quantum computers at the backend so I'll stop sharing and I'll reshare my
26:53
screen somehow yeah let I hope you can see my
26:58
screen so uh this is a portfolio optimization and asset allocation solution that we have built the idea is
27:07
that if we can uh apply uh Quantum Computing to portfolio optimization
27:13
problem so this is a recorded video because running the solution still will take a little longer if not years uh
27:20
like in a classical computer if I play this so if you see we have looked at three different stocks th 30 stocks uh
27:27
uh your NASDAQ 100 and S&P 100 and I'll show you all three one by one and I'll
27:33
compare it against the classical solution so we have selected the Dow 30 we have taken data from Jan 15 to
27:38
December 19 we did not take recent data because of the because when Co occurred everything changed so we looked at that
27:45
five years of historical data to create the forward-looking risk and return for
27:51
each of the stocks in theor now taking that as the base the statistical return
27:56
and statistic iCal risk for each of these stocks we tried to combine these different stocks together to build
28:03
portfolios which will give you the risk best risk return so I'm showing the first one this
28:11
is just the descriptive page you can review the data and then you can run the portfolio optimization so once we press
28:18
this button the quantum engine is at works this is running on d-wave so this is the
28:24
first result now I'll explain this a bit so if you see this portfolio Frontier which is the most important graph here
28:30
the blue dots are a classical uh solution a classical Monte Carlo enhan solution that we have built where it
28:37
generates different portfolio each dot is a portfolio and the Amber ones here
28:42
are the portfolio generated using a quantum computer now if you look at
28:47
these two you will see some Advantage for similar returns you are for similar
28:53
returns you can find some stock some portfolios with a lower risk now given it's a 30 stock thing with 1
29:00
million invested uh classical computer can handle it in a good way let's see what
29:07
happens when we go to the next solution the next uh uh next indices so right now
29:13
it shows some annual amount this is the yearly return and also we choose NASDAQ 100 and let's see what happens here so
29:19
I'll directly jump into the results now this is interesting if you
29:27
see again the Amber dots are quantum the blue dots are classical for the same portfolio risk
29:34
you find a lot of uh a lot of portfolios with a better return so classical could
29:40
not do it quantum did now this is the second one each of these graphs here
29:46
basically show how much should you invest in each of these stocks we have limited it to 10% at most for each stock
29:52
because there will be Market limits to it this is not exact replication of how Market works but for a comparative
29:58
solution with a classical system this works well enough now if you go to the last one
30:04
this is S&P 100 and here the results again are phenomenal so if you look at it the classical solution basically for
30:12
the same risk same uh risk range the returns are very high and vice versa so
30:18
this at least establishes to certain extent that Quantum uh driven Solutions
30:24
have some capability to surpass what is available class Bally today so this is the first demo I wanted to show
30:30
unfortunately I'll have to stop sharing again because uh Zoom is not allowing me
30:35
to share that share the other demo in the same application so I'll share again just hold on
30:43
please yeah now I hope you can see my screen again now this is the last myot
30:50
optimization demo as I said if you look at companies like FedEx ups and others
30:55
their uh their goal is is to send packages from point A to point B the structure it takes is that there is a
31:02
city station where the packages are collected for delivery or pickup and
31:08
then they are sent over uh in a hub and spoke model so the last part of it which is connected to the customer for uh for
31:16
pickup or delivery is where a lot of cost goes around 53% cost according to
31:21
one of the reports if you are able to optimize the distance traveled by
31:26
different trucks given your Fleet you'll be able to save a lot of cost and in
31:31
that way it's a very complicated solution think about if there are 10
31:37
drivers five time Windows 100 packages or 10 packages it is 10 into 10 into 5
31:44
500 the number of solutions that uh become feasible are 2 to the^ 500 a very
31:50
large number and FedEx and companies like these operate 20,000 packages with thousands of truck at some of the
31:56
location ations they are not able to solve it using any optimization software available today but rather they break
32:03
down the problem into smaller pieces which is suboptimal in nature Quantum does have the capability to handle the
32:08
larger problem in one go and gives you give you the most efficient routes so this is the landing page where we
32:16
start now here what we do basically is we select a new data set here you provide the capacity of the vehicle you
32:22
can give your own Fleet also but we are giving only the vehicle capacity now you schedule your uh pickup and Delivery for
32:30
the day so for today let's say so now you have the depot and the
32:36
customers what is to be delivered how much to be delivered this is the depot this is the location how much to be
32:43
delivered where to be delivered what is the time window in which it is to be delivered so now what we do is we
32:49
basically review that data and same thing is shown on the graph now you run the quantum
32:55
optimization engine it runs and finally if you see it assigns to each customer of vehicle so
33:02
this is after uh running after identifying the number of vehicles required and secondly uh uh secondly
33:10
designing a route after that how much uh which ve which package should go to
33:15
which vehicle basically and on the right hand side in the chart you'll see uh the
33:21
different colors indicating different trucks and P and D saying uh pick up or delivery with customer IDs and vehicle
33:28
combined together now you can review these you can move on and go to the next
33:33
level here basically we showcase uh some uh kpis with respect to our solution so
33:40
if you see total time taken is 19 hours 616 miles covered $185 of fuel cost
33:46
given some mileage and uh uh mileage and cost of fuel uh cost of per unit fuel as
33:55
considerations also did we miss any customers uh was there any overload in
34:01
any of these vehicle all these are important things that we have to consider when building a
34:06
solution now if we go here uh we can go vehicle by vehicle and see the routes that have been designed using this new
34:13
algorithm that we have put into place it gives you the order it gives you the time it gives you the customer
34:19
everything it goes down to vehicle level details distance traveled per vehicle and all some some different kpis
34:27
now since this is a pickup as well as delivery it also shows vehicle load by
34:33
time so when the delivery happened or pickup happened whether anything was overloaded or offloaded all that is shown in this graph so we are building
34:41
the next version of it where we are not just considering the weight constraints of the vehicle but the volume
34:46
constraints we are looking at equitable distribution of work in this case all that has been is being developed this a
34:53
second demo and given we have some time I would like to show one more Dem demo
34:58
here uh I may have to again share my screen so let me just open the demo and
35:03
then I'll proceed forward yeah I hope you can see it so
35:10
this is Airline creu scheduling very important this problem uh basically uh
35:16
if you look at what happened during uh Christmas when the different flights were landed and what knew how to reach
35:23
schedule and all these problems become very important they are large scale problems not easily solvable and some
35:29
sort of uh uh some sort of comput computation compromise is being done
35:35
today to get the solution quantum computers gives you a way to solve these
35:40
problems in their entirety and uh on in this way and we have built this problem this problem is called Airline Cruise
35:46
scheduling it is a two-part problem first you are given a monthly schedule of some Airlines say United Delta or
35:54
WestJet from that monthly schedule say of 500 flights you have to create
35:59
certain groups based on constraints now these constraints are no two flight should overlap there should be a gap of
36:06
1 hour between each flight these sort of constraints are put by labor organization Airlines uh operating team
36:14
Airport operating Team all those now considering those constraints you have to create these groups say you create 10
36:20
groups of 50 flights each easiest solution for now now to each of these groups you have a crew which are at
36:27
certain base locations these crews are to be assigned now again there are constraints when the crew is available
36:33
how many hours they should work all those is considered so the first part where you have to form the groups is
36:38
called Airline crew pairing the second part where you have to uh where you have to assign Crews to the group is called
36:45
Airline crew rostering now again we'll start with a new one here you have to give the
36:51
monthly schedule first so upload your new flight schedule
37:01
now all this is done you add certain constraints that I talked about you can see flying hours of the group The number
37:08
of flights in the group the interflight rest time Duty period all this is to be
37:13
considered so once you do that we are showing some dashboard one so this is a real world data of uh one of the one of
37:20
North American Airlines which had 13 1,3 flights around from 26 different
37:25
airports with 34 million flights per day this dashboard that you see right now we
37:31
trying to establish the complexity of the problem so if you see flight departure
37:38
by day the by hour of the day by flight frequency per day those sort of things
37:44
we have shown you see a lot of overlap between flights from the same airport or
37:49
same arrival and departure airport combination so it's not easy to solve this problem through any method now here
37:55
we are solving the first problem where we are finding groups called pairs from this initial data now if you see we have
38:02
found some pairs for each pair contains certain number of flights and if you see
38:08
each flight leg has been assigned a pair now if you look at each of those pairs
38:13
some some number of flights are assigned now we have compared this between Quantum and classical and Quantum uh is
38:21
working well and we have to keep in mind these are new smaller computers and still they are almost match almost
38:28
either breaking or matching the performance given by classical computers and they're progressing rapidly so some
38:35
comparison by each pair what are the flying RS what are the overall RS all these become
38:42
important this is the first problem that we have solved in the second problem we have to supply the crew data so at three
38:48
different bases the crew are available we upload this
38:54
data now again you have to supply apply certain constraints from the crew so rest period for the crew the flying ass
39:00
of the crew the overall ass the number of crew members per flight all that is to be given and again the second problem
39:06
is solved for each crew we have given the flights that they're going to cover in which pairing they'll be allocated
39:11
all that is given here sorry so again we have done a classical
39:18
Quantum comparison here to Showcase uh what are the results and though this is
39:24
a demo I uh I would encourage you to look at the white papers they are available uh for free on our website
39:30
where we have done a detailed comparison of uh the the different solutions we have run and as we as what we have found
39:38
that as we increase the size of the problem classical computer stopped giving Solutions all together we waited
39:44
for in fact three days no solution quantum computer was done uh say in half
39:49
a day uh basically in fact so every optimization problem has two part one is
39:55
a formulation part part where you encode your objective and constraints programmatically second is the solver
40:01
part in the classical World both are time consuming solver is more time consuming in Quantum case solver is not
40:08
that time consuming it is it sometimes it takes just 7 seconds or something to
40:13
return return the results so this is where I stop with the demos and what I'll do next is I just go
40:20
back to the deck yeah
40:27
sorry my mouse is
40:32
beh yeah so this is where we were now uh we talked about what we have built but
40:38
uh nobody can do Quantum alone in this world it's it's a very complicated field you need uh you need Academy on your
40:44
side you need uh tools and Cloud providers on your site so our ecosystem is built in two two places India and
40:52
Alberta Alberta is our Quantum headquarters in in India we have we are
40:58
uh one of the two industry partners with uh government of India for their Quantum Computing applications lab initiative uh
41:05
we have partnered with I Madras where we have opened our own Quantum Computing research lab uh we are partners with
41:13
krishi which is a startup working in this area again founded by uh a team from I Delhi uh and other universities
41:21
in Alberta we are anchor business partners with University of Calgary uh it is called Quantum for the their
41:28
initiative called Quantum City we are partners with Calgary Economic Development Alberta Government and invest Alberta and we have a shared uh
41:35
ecosystem of providers such as AWS dawave classic Ando now the goal is to
41:42
basically bring uh uh work with our customers in Alberta Canada North
41:47
America and globally to solve uh their Quantum machine learning optimization
41:52
simulation problems using Quantum computing uh this is a special relationship and
41:59
the goal is to create a client focused Quantum Computing ecosystem just hold on sorry and uh here
42:07
we have joined hands together with University of Calgary for the quantum City initiative they are uh the training
42:14
Partners Talent Development R&D partners for this partnership where emphasis
42:19
brings in Industry relevant Innovations and use cases sorry job creation technology
42:27
Partnerships and all uh this collaboration the goal is to do Quantum awareness and training programs for
42:33
industry build a Quantum Readiness engagement model for adoption of quantum Computing access to Industry scientific
42:40
and Technology experts all this now uh since we're addressing
42:46
client needs there has to be an engagement model is placed our goal is to get our customers Quantum ready the
42:52
technology is take Taking pace and a l adoption has consequences so given that
42:59
idea we want our customers to uh start exploring Quantum technology and for the
43:04
same we have built an engagement model so this is the engagement model for Quantum Computing adoption here uh if you look at it we
43:11
have divided this into four parts given the novelty of this technology and the awareness in the market so we start with
43:18
Quantum awareness sessions the idea is to help you create an in-house team or Coe to focus on your Quantum Computing
43:24
goals these take the of a workshop which can be customized based on audience so
43:30
it takes Business technical and strategy point of view now this generally runs
43:35
throughout the engagement we don't want to run it at the very start it depends on when the Right audience is available
43:41
or when they are ready now once other than that what we start with is an assessment Workshop uh the idea is
43:49
basically you don't know where to apply Quantum Computing we come and help you out if you don't if you already have a
43:55
team in place you don't need to do these workshops but the idea is during this design thinking driven uh phase our
44:02
experts work closely with you to identify Rank and prioritize use cases so you identify the use cases you look
44:08
at parameters like whether this use case benefit from Quantum Computing or not if
44:14
it benefits how much benefit you'll get whether it will make uh decent noise in
44:19
your organization to take Quantum Computing forward and many more things do you have the right data in place to
44:24
solve this use case and the and finally we take one use case now the next step
44:30
is a Consulting Workshop this is a in this case we fall we basically address
44:35
two major objectives if you don't have a Quantum Computing service provider on
44:40
your side we help you identify the right one so the idea is for example if you
44:46
want to solve optimization use case at scale today I would say that go with d-waves Quantum Manila if you say no we
44:53
want to do machine learning I will say let's explore some Gateway systems from IBM atom Computing and others so that we
45:01
help you with then we also do some deep dive and find tuning of shortlisted use cases so the use case that we have
45:07
shortlisted in the last phas we looked at those and we basically try to uh Deep
45:13
dive into it what data would you require to solve this use case what kind of kpis you have today how will the system
45:19
integrated how will it work overcloud all those things we address in this workshops and finally when we are done
45:25
with all this we jump into a longer session a longer phase which is a Quantum solution development this we
45:32
work with you to design and implement the solution the goal is not to build a solution and hand it over to you and run
45:38
away the goal is that we build it and we have continuous sessions with your team
45:43
to help you understand how it is being built so that when you pick up the next use case your team is also ready even if
45:50
not to execute but to understand what we are doing and assess the same so sorry so once so now this is the
45:58
engagement model but how do we execute this engagement model so this we have taken a p driven model uh for this B
46:05
driven delivery model the goal is that your use case and it it can be oriented
46:11
towards IP development client delivery capability building any of those each part is a self-sufficient entity which
46:17
contains a Quantum Project Lead a technical lead a domain expert or Quantum developers project managers who
46:23
look at your use case and apply those four phases to it the ones we saw
46:29
earlier these people are they're multi-disciplinary in nature they understand business they understand
46:34
technology and they work together to solve this problem now we have a strategy pod also in place which
46:41
basically is multi-disciplinary emphasis and partners and they look at the
46:46
direction we should take to build Solutions uh address the different strs of quantum Computing which Partnerships
46:54
to take so on and so forth so the this helps you in getting a guidance on where the uh technology is going and we have
47:00
some very strong Partners like University of Calgary where we have who are experts in this field for more than
47:06
30 years now they know in and out of this technology we bring the industry
47:11
angle as well now these are the three use cases we have done till date and we are
47:17
addressing a few more right now the first one is the last phase of qu uh the
47:23
engagement model we were given the use case all we had to do is to apply Quantum Computing to the use case we did
47:30
not do awareness sessions we did not do assessment or Consulting workshops in this we applied a Last Mile route
47:36
optimization solution that I talked about earlier to a last M
47:41
container container delivery with reverse pickup so this for a five Monon project
47:48
uh and this was done for one of the largest uh freight service providers in
47:53
North America uh that is one now uh next one
47:59
is a qu next one is a Quantum Readiness program that we run for one of the largest American healthcare companies
48:04
this was recently concluded uh this again we did for five to six months here
48:10
we did the end to endend engagement model that I talked about awareness session assessment model Consulting
48:15
model as well as the quantum solution development this started with a team of
48:20
zero people who was part of their Quantum team right now they are working with five people in their team
48:26
addressing different areas finally the last one is an interesting one it is ongoing but the
48:32
goal was given uh bodily fluid samples
48:38
can you identify the chemical profile uh that exist within your blood your serum
48:44
your urine and come up with the pathogens that exist in your body so
48:49
that can be addressed using a machine learning solution and we applied Quantum machine learning for it so in this case
48:55
of also the problem was given to us all we had to do was quantum solution
49:00
development now this is my last slide and after that I can take questions this is the analyst response I have we have
49:07
received from our work in this area so from HFS highlight global data it noara
49:13
Gartner all of them have looked at our work and given these comments so this is
49:18
where I'll stop uh over to you VI or Taylor and I
49:24
can address the question awesome thank you Rohit great work uh we do have a couple questions in the chat
49:31
and then once we get to those if anyone would like to unmute themselves and ask any questions please feel free uh the
49:37
first question is from Naveen and they ask does Quantum Computing require specific Hardware or chip yes so Quantum
49:45
Computing itself means a quantum computer a different type of computer it requires a chip based on principles of
49:52
quantum physics and quantum mechanics and these chips are available from different Hardware vendors over Cloud so
49:58
IBM has it Google has it uh Pascal has it atom Computing has it and they have
50:05
uh provided it over Quantum comput as a service providers overcloud IBM Quantum experience AWS bracket Azure Quantum all
50:14
all of them provides hardware and software tools for you to solve the
50:19
problem awesome thank you uh Raphael asks it's it's a bit of a three-parter
50:24
so I'll start if you want to address it in three different parts we can Rafael says can you explain Energy savings when
50:30
using quantum computers to run these Solutions do Quantum Solutions run on
50:36
real time or do they apply to batches and cues okay so I'll address these one by one the first one about Energy
50:43
savings so I do not have an exact answer with this but uh the idea is this that
50:50
right now most of the computers that are many computers that are available today the system has to be cooled down to say
50:57
below zero kind of a thing right so uh given that a lot of energy is expended
51:04
but room temperature Quantum Computing using photonics is also being addressed
51:09
and they will consume very low energy the reason being is these are reversible competition a special thing that you
51:15
have to look at that during competition they do not lose energy so the overall
51:21
cost to run these is going to be low that is my understanding now are these Quantum Solutions are real
51:28
time so uh let me take it uh in form an
51:33
example if you talk about machine learning let's say when you are trading the system uh it's a similar task as what
51:41
you do in classical machine learning only the backend compute has changed but
51:47
when you look at uh running the inference they will be real time you can
51:52
basically uh uh you can provide your batch data or streaming data and they
51:58
can predict what you want to predict so it applies to batches it applies to Q it depends on the type of problem you're
52:04
trying to address rather than the technology you can look at the classical technology how IT addresses these
52:10
problems the exact same thing will be done by a quantum computer there's no difference only thing is backend machine
52:16
has changed awesome thank you roit uh just a couple more for you before we um before
52:23
we head out of here the first one is Will quantum computers once mature replace classical
52:28
computers uh okay so no uh they will not so as I said earlier also that today
52:35
also there are different computers that exist and each has their own specific task uh quantum computer will also have
52:43
have some task which it will excel at for example it is a very important
52:48
technology for fields of chemistry and Material Science to find new materials
52:54
or drugs or do drug discoveries not easy on a classical computer quantum
52:59
computers can give a big Advantage there for example they're not going to be used for web search classical computers do
53:06
that well enough so no they are not going to replace it but they'll work with classical computers similar to how
53:12
qpu uh your uh C gpus physics processing unit and all that work together to say
53:18
build a gaming system today amazing and the last question we have here I think this is really
53:24
important too is what should be the timeline for adoptions of for adoption of quantum Computing so uh nobody can
53:31
say for certain uh numbers the answers are different somebody says that proper fault tolerant large scale Quantum
53:38
Computing will available only in a decade but that does not mean the comput
53:43
quantum computers that are coming up slowly in or will come up in a course of
53:49
one or two years will not change the way uh your operations can be realized so
53:55
your operations still you are not looking at exponential Advantage but some Advantage which gives you cost savings so from that perspective
54:03
adoption should be faster another reason is it is not an easy technology to master it requires an organization wide
54:10
adoption given two these two things it takes time it is not easy to adopt it
54:15
immediately it's not like for example when we switched from CPU to GPU for
54:20
certain tasks still it was not that difficult the underlying technology is same the principles of programming are
54:26
same many of the things are same and uh we were able to address it here the pro whole prog programming Paradigm itself
54:32
is different you're looking to use superposition interference and entanglement very physics based concept
54:40
to write programs but if you can understand them well enough and you can write certain Solutions the advantages
54:47
are humongous and people do understand it your whole organization do not have to switch to Quantum you need a small
54:53
team to begin with which understands Quantum and then grow that team slowly
54:58
hence we are working in this getting Quantum ready model where we start with a small team and then subsequently grow
55:05
them up so that you can uh get the advantage in the longer run it will be applied in certain scenarios again as I
55:10
said it won't be applied in each scenario so you need a particular team size to get the
55:16
benefits awesome uh somebody has a question here with their HR please feel free to unmute
55:21
yourself uh thank you for the presentation this was really great I wanted to ask about the
55:28
commercialization of uh Quantum Computing products in the SAS space uh
55:33
just like how we're seeing the peration of AI products and machine learning products and SAS um how long how many
55:40
years do you think before we start to see off-the-shelf products that allow us to be able to solve very specific forms
55:47
of quantum problems so you can so SAS so if you talk about SAS products can come
55:54
out anytime but the idea is the only issue with Quantum today is solving problem at scale so even if I give you a
56:01
SAS space solution today and we have already have a SAS space solution deployed an Azure for anomal detection
56:07
you can use it the problem is when it comes to scale uh it may not be it may
56:14
not satisfy the requirement you have the scalability is a bigger problem than the specification of it so people are trying
56:20
to solve the same but still as I said uh few days back only 10 days back this
56:26
1,00 Cubit computer has been launched as I said when it became 50 Cubit it already surpassed the capability of a
56:33
classical computer had it been fall tolerant 1,100 is a big number if they
56:38
achieve fall tolerance tomorrow and connectivity you see that the scale has
56:43
changed completely and you can address some large problem classfication will not take time because everything that
56:49
you do today on a classical computer the way uh the apis work the way the solutions are deployed nothing changes
56:56
right now also the all the solutions that you saw that uh today we made API calls to a quantum computer to get the
57:03
result the programming language used was python nothing changes from that perspective only back backend machines
57:10
are changing so SAS Solutions can come up any time but will they be useful or not is the bigger question at this point
57:17
in time awesome thank you Rohit last call
57:22
for any questions before we wrap up for the day
57:28
okay well thank you all so much for attending today um Rohit and VC will be able to send you all the slides he has
57:34
your they both have your contact information and this recording will be uploaded in a couple of days on the
57:39
charm Board of Trades website uh YouTube website so stay tuned for that and again thank you all for attending today's
57:45
session and enjoy the rest of your
Workshop Agenda
- 1:00 PM - Virtual Check-in
- 1:05 PM - Workshop Programming
- 1:50 PM - Q&A
Speakers
Rohit Kumar Patel, AVP & Lead – Data Science & Quantum Computing, Mphasis
About our member:
Mphasis applies next-generation technology to help enterprises transform businesses globally. Customer centricity is foundational to Mphasis and is reflected in the Mphasis’ Front2Back™ Transformation approach. Front2Back™ uses the exponential power of cloud and cognitive to provide hyper-personalized digital experience to clients and their end customers.