WEBVTT

1
00:00:21.149 --> 00:00:24.600
Hello.

2
00:00:30.899 --> 00:00:34.320
Hello.

3
00:00:35.549 --> 00:00:39.179
Okay.

4
00:00:39.179 --> 00:00:47.640
Silence.

5
00:00:54.630 --> 00:01:00.960
Okay.

6
00:01:05.760 --> 00:01:09.810
Hello.

7
00:01:09.810 --> 00:01:14.010
Oh, I can't hear you.

8
00:01:14.010 --> 00:01:19.769
Hello.

9
00:01:21.150 --> 00:01:28.349
Okay, let me re, log in. For some reason, I can't hear you.

10
00:01:32.459 --> 00:01:39.900
Okay, yeah.

11
00:01:39.900 --> 00:01:44.040
Can you hear me now? Oh, yeah. Now I can hear you. Yeah, thank you.

12
00:01:44.040 --> 00:01:47.640
Yes, it can. Can you hear me.

13
00:01:47.640 --> 00:01:50.700
Not, yes, I can hear you. Great.

14
00:01:50.700 --> 00:01:57.629
Yeah all right so then I don't have to logging in again. Yes.

15
00:01:57.629 --> 00:02:02.760
So, let's see how many people are here.

16
00:02:02.760 --> 00:02:07.739
Okay.

17
00:02:13.110 --> 00:02:18.870
Let's see if anyone.

18
00:02:20.969 --> 00:02:24.120
Well, I would go great. Okay so.

19
00:02:24.120 --> 00:02:28.889
We are, we are on so.

20
00:02:28.889 --> 00:02:36.719
Good afternoon class so this is class 24, Monday, November 23rd and.

21
00:02:36.719 --> 00:02:50.754
We are lucky to have with us professor Wang will talk to us and I put a link to her to our research page on the class blog that I'll show you after and just some announcements 1st for final

22
00:02:50.784 --> 00:02:51.775
presentations.

23
00:02:51.775 --> 00:02:57.025
Let me just do a share here for a minute and.

24
00:02:58.710 --> 00:03:03.870
Okay, so.

25
00:03:03.870 --> 00:03:11.189
Come on here. Okay, so there's the link here fresher wings page.

26
00:03:11.189 --> 00:03:21.389
And presentations remember that the last 3 class days 4 people asked about specific dates. They have them.

27
00:03:22.650 --> 00:03:37.224
And I just assigned filled out the dates with with people to fill out the same number each time. If people have massive objections. And, I mean, some of you are forming teams, then, let me know, we'll coagulate.

28
00:03:37.585 --> 00:03:43.675
And this little joke thing before the talk quantum physics explained to behavior. Okay. So.

29
00:03:45.270 --> 00:03:49.289
Now, against professor, Wang.

30
00:03:49.289 --> 00:03:53.939
And.

31
00:03:55.110 --> 00:04:02.969
Okay, so I can start to share my PowerPoint slides, but I do hear us John. I called.

32
00:04:02.969 --> 00:04:10.650
I'm sorry, I hear I call well, I will mute. It sees everyone.

33
00:04:10.650 --> 00:04:15.990
I will mute myself once you start talking.

34
00:04:25.795 --> 00:04:35.004
Okay, well, thank you Renault for the nice introduction, and also post my personal wide patient link on your blog.

35
00:04:35.425 --> 00:04:45.084
Unfortunately, on my web page, I didn't have any information related to quantum computing. So this is, um, like a quantum computing is a new topic.

36
00:04:45.084 --> 00:04:52.584
Um, so we are also learning exploring a tough also my web page I haven't updated for years.

37
00:04:52.889 --> 00:05:07.858
Minimum 2 years so okay. Maybe I'll do it this winter winter break. Okay so, here, this is actually a presentation I gave earlier this year in August to group all for our professor. So so.

38
00:05:08.153 --> 00:05:18.413
So, here, I'll just recycle the slides. So, the goal is to introduce you a alternative way to build a quantum case, or, you know, quantum computers.

39
00:05:18.803 --> 00:05:29.483
So, by taking a professor, our runoffs class, you probably random Franklin's class. You probably learned a IBM and Google other tool like, um.

40
00:05:30.238 --> 00:05:38.668
Big, I like, um, um, how maybe we can see as a competitor so the 2 places so they have the, um.

41
00:05:39.084 --> 00:05:48.894
Established quantum computing, and the demonstrates they demonstrated up to 53, right in Tango like, um, you know, Quip it's a calculation.

42
00:05:49.043 --> 00:06:01.374
So, um, the 1 thing about the way they, um, you know, achieve it from physics point of view, the user Joseph production to build, um, like a, the basic circular blocker.

43
00:06:01.374 --> 00:06:13.553
So, this is a nice way to do it. But the 1 limitation is, if you do it in Joseph Junction, it have to operate in 2 to 3 cabin. So you have to are.

44
00:06:14.278 --> 00:06:16.613
Make a temperature,

45
00:06:16.613 --> 00:06:16.944
so,

46
00:06:17.303 --> 00:06:17.603
uh,

47
00:06:17.603 --> 00:06:17.934
it's,

48
00:06:17.934 --> 00:06:18.233
um,

49
00:06:18.473 --> 00:06:18.923
um,

50
00:06:18.954 --> 00:06:19.343
you know,

51
00:06:19.343 --> 00:06:28.884
the associated equipment is a massive or so it's hard to really transform a quantum computer based John,

52
00:06:29.244 --> 00:06:30.233
Joseph junction.

53
00:06:30.233 --> 00:06:33.204
Um, you know, personal. Usually, it's hard to imagine at home.

54
00:06:33.473 --> 00:06:48.204
We said, um, you know, refrigerator, the actually, the refrigerator requires diffusion refrigerator is a special kind to be able to pump down to a temperature either to 3 cabin. Okay.

55
00:06:48.204 --> 00:07:02.723
That's extremely low temperature. So here, we presented our alternative way to build a quantum computer is actually use the full transcript. So, um, there's a couple advantages using, um, like, uh, optics to do the quantum computing. So, why is it?

56
00:07:02.723 --> 00:07:08.093
Room temperature? So, you don't have to worry about 2 is more, um.

57
00:07:08.963 --> 00:07:20.963
Noisy robust, because the photons are confined in, like, a week guys waiting fiber. So so they are more robust. It shows the experimental noise.

58
00:07:21.053 --> 00:07:32.454
So that's actually to the teacher, um, for, um, like, um, for tanika based quantum computer, but it doesn't mean the quantum computer doesn't have any challenges it does. Okay.

59
00:07:32.454 --> 00:07:33.204
For example,

60
00:07:33.204 --> 00:07:33.894
1 way to do,

61
00:07:33.894 --> 00:07:48.444
it is generate the entangled 4 times the will use a different semiconductor material and then capital the tangled the full town to the compute computing unit by the copying parties can be a challenge because it

62
00:07:48.444 --> 00:07:51.923
requires extremely low loss at the interface okay,

63
00:07:51.923 --> 00:07:54.564
so this is a pilot beta of engineering,

64
00:07:54.834 --> 00:07:58.103
practical challenges to implement a large scale.

65
00:07:58.283 --> 00:08:04.553
But nevertheless, they have already demonstrated the useful way to build a quantum computer.

66
00:08:04.764 --> 00:08:16.673
And the many researchers nowadays in award is doing some of the, um, like defense agencies like air force is doing it. There's a couple for startup companies.

67
00:08:17.694 --> 00:08:26.274
In Silicon Valley, they are actually actively aggressively pursuing this technology try to push it to the marketplace.

68
00:08:26.694 --> 00:08:41.033
So I was fortunate to United, um, like, uh, 1 of the startup companies that to talk at a conference. So he layout to the roadmap is about 5 years.

69
00:08:41.274 --> 00:08:55.494
So, he anticipated, basically, 5 years, he think his company's product launch on to the launch into the marketplace of beta for panel computing. So we'll see. Okay so I'm just a slice pretty simple.

70
00:08:55.524 --> 00:09:08.634
I would just go over some basic things. So, you probably know many of the days. So also, if your customer feel free to stop me to ask her, okay, let me put it in full screen. Okay.

71
00:09:09.269 --> 00:09:20.099
Okay, so let's start with, um, classical pizza and the classical logic gator. So, in classical, because we know it's a 0T and 1.

72
00:09:20.099 --> 00:09:23.908
And, uh, for classical Gates, so are we.

73
00:09:23.908 --> 00:09:27.298
We have 1, not Gates.

74
00:09:27.298 --> 00:09:35.938
Or Gates, so, and also a number of other Gates, so you're probably familiar with data from your circuit, the class. Um, so, um.

75
00:09:35.938 --> 00:09:41.548
If we recall how we build this, a classical pizza in a.

76
00:09:41.548 --> 00:09:54.653
Semiconductor circuit we so you probably remember in your Micro electrons class where your analog circular class or digital circuit class. So that's actually you learn later.

77
00:09:54.653 --> 00:09:58.403
So the classical piece, I realized the most fat.

78
00:09:58.918 --> 00:10:11.783
Like, uh, you know, like a circuit transistor building block so your transistor may spit out a way out the voltages 1 okay that can be re, present a pizza 1. okay.

79
00:10:11.783 --> 00:10:21.053
Then, if your voltage is the end Apple, low voltage, high, low energy, and have a low voltage, maybe 0T. Okay. That's can represent a classical.

80
00:10:21.418 --> 00:10:27.658
Classical B. to 0T. Okay. And then you program code 8 0, 1 0T you know.

81
00:10:27.658 --> 00:10:40.619
Different coding by change by changing the, uh, the voltage of the transistor. So you can generate the design, the algorithm in the.

82
00:10:41.183 --> 00:10:54.323
Transistors and the build this data. So when you build this case, you may have a number of transistor put together, they realize this functions. Okay. So I'm now going through the details. Because I assume you, you're pretty familiar with the, for example.

83
00:10:54.323 --> 00:11:04.974
Now, get into, if you have input of 0T, I'll put it shouldn't be 1. okay. Input 1 just update a flip beta so you're going to be 0T. Okay so now, let's see.

84
00:11:06.293 --> 00:11:19.193
How the quantum beta I realized, so in quantum awards, we also have 2, 2 States, but there is a representative director. So so you probably read if you're really literature.

85
00:11:19.193 --> 00:11:21.293
So you're going to see these 2.

86
00:11:23.428 --> 00:11:38.399
Like, comes up to being used interchangeably so when we see a Stater basically means it's whacked here. Okay so actor can be a 0 1. okay. Um, the cubic.

87
00:11:38.399 --> 00:11:48.178
Is constructed by the Super superposition of the 2 States. Okay. So, um, the QB defy.

88
00:11:48.178 --> 00:12:02.634
Is our 5 times it's like 0T so we call the base is 0T and 1 called the basics that they are circling on to each other. Okay. So our 0T and the beta 1. okay. So, by the alpha beta is now random.

89
00:12:02.693 --> 00:12:16.884
So they have to satisfy this relation. So the plus monitors beta equal to want to are for beta itself will can be complex numbers. So, on you may also see a different notations.

90
00:12:17.364 --> 00:12:26.634
Wow. For the very commonly used in rotations. You express the Stacy like, accountant. So the 0T okay S1 0T. Okay.

91
00:12:26.663 --> 00:12:41.364
In column factor and the 1 0T and so he's a reflector from mathematical point of view. This is the base, right? So if you have more okay, so then you can see 3 dimensional you can have 10001 0. 0.

92
00:12:41.394 --> 00:12:51.744
0. 1, so that's like, spend the space you can also do for. Okay. So so, um, alright, so here is a single cube. All right the 5.

93
00:12:55.854 --> 00:13:08.634
The quantum state of Phi, then can be expressed by the combination of this. So what's the difference of base and a bit a bit? So you are sure in the end.

94
00:13:08.813 --> 00:13:16.224
Okay, so you either get 0T or get 1? Okay in the content better. You got a combination of this.

95
00:13:16.678 --> 00:13:21.869
Okay, so if you can get any value in between so it's a match you reach.

96
00:13:21.869 --> 00:13:26.038
In our output, so this is also.

97
00:13:27.293 --> 00:13:39.173
So, the, the, the probe side of data okay so basically, quantum computing you can consider this is a way of analog computing. So that's has like, a infant has a number of.

98
00:13:39.984 --> 00:13:44.484
Okay because this is the linear combination of these 2 bases.

99
00:13:44.724 --> 00:13:58.823
Okay, so that's make quantum computing a mirror powerful when you're handover complex more very tedious some mathematical problems right? So it's became very powerful doing that.

100
00:13:58.823 --> 00:14:11.933
It bothers also on the con side, because it's I dunno computing. So, sometimes you gather value, you don't awareness what you really get, or there's actually noise error.

101
00:14:12.269 --> 00:14:22.974
Like, somehow a couple into the system, so this is so important for quantum computers to have the error correction code.

102
00:14:23.063 --> 00:14:31.673
So I'm sure, like professor, um, like a rundown frankly I'm going to, uh, you know, has covered it in the class right?

103
00:14:32.908 --> 00:14:45.803
Assuming you covered it right? Okay. Yeah, I see. I notice you're just correcting in detail. We've alluded to it, but oh, okay. All right.

104
00:14:45.803 --> 00:14:54.354
So, but anyway, so, I assume students will know, like, if you look at quantum computing, the arrow crushing code is a big part of it okay.

105
00:14:54.563 --> 00:15:05.933
To, um, you know, otherwise how, how a also 1 way to, like a major how result. So they use that, like, we'll do coffee to migrate so.

106
00:15:07.134 --> 00:15:21.173
All right, so, let's move on to next 1. okay. From physical points. We're on. Okay. So, since it is a quantum computer, if you ever study the quantum mechanics, so you will know in quantum mechanics.

107
00:15:21.563 --> 00:15:28.344
So we all the computations and realized that by operate. hersa. Okay. So.

108
00:15:28.708 --> 00:15:42.234
If we connected to the digital quantum computing, so we call okay, we try to study. What is the quantum gator quantum gator? Indeed. Mathematically that adjuster operator quantum operator. Okay.

109
00:15:42.443 --> 00:15:52.583
So, um, more specifically we use the unitary operator to achieve the, to build the basic quantum gator. Okay, so here are we.

110
00:15:52.918 --> 00:15:57.808
Do you know what is a, you and you turn a matrix unitary operator.

111
00:15:59.063 --> 00:16:13.822
We discussed it. Yeah, it does. Okay. Great. Great. Okay. So, quantum gator always the 2 important to output. Okay this can construct a single cube. It. Okay. So the quantum good acting on a single cubital.

112
00:16:14.303 --> 00:16:28.913
Is a 2 by 2 unit matrix so basically, if you want to do an operation, so this is how it achieved. Okay so this is the basic definition of your internal matrix. I'm glad a professor frankly has covered in the class. So this.

113
00:16:29.219 --> 00:16:37.318
Luckily can be very easily realized the photo for so.

114
00:16:37.764 --> 00:16:52.283
All right, a condom case with inputs and outputs, so we'll make it a represented by a matrix of a degree, or for 2 to the power of N. so, for example, or to QB, the gator can be implemented as a matrix degree of 2.

115
00:16:52.313 --> 00:17:04.854
so then become 2, so then you will require a 4 by 4 unit matrix or later I'm going to show you a format for unitary matrix that can be expanded to a network or for.

116
00:17:05.189 --> 00:17:17.368
2 by 2, inter Matrix, uh, as a matter of fact, um, it can be proved rigorously that any arbitrary number or 4 unit matrix and by end all be expanded.

117
00:17:17.368 --> 00:17:26.519
Okay, build upon the, basically your intro and matrix at to back to. Okay, so that's give us a way to construct.

118
00:17:26.519 --> 00:17:30.479
The quantum data use for panic approach. Okay.

119
00:17:30.479 --> 00:17:35.578
Alright, so here it is nice. Move on to the next life.

120
00:17:36.989 --> 00:17:45.898
All right, let's look at the example see how quantum operator quantum quantum gets apply.

121
00:17:46.644 --> 00:17:59.394
On the basic quantum states. Okay. So, um, party matrix. Okay party matrix is a unit. Matrix is also permission matrix. Okay. So it's a here.

122
00:17:59.423 --> 00:18:12.923
We just consider the single copious data. So, if you you, perhaps that too. So, the pounding Matrix, the X Y, and Z. so this is their standard expression matrix. Okay.

123
00:18:12.923 --> 00:18:19.284
Now we consider party matrix what we call the poly operator, or you can hire to Polly.

124
00:18:19.618 --> 00:18:24.388
Like, a data, so the data, a pie.

125
00:18:24.388 --> 00:18:38.699
I'm a single qbtr single cobia consists of 2 bases. 0T and 1. okay. So the, um, the, the party X. okay. 0, 1 0T this is a gate operator apply on the.

126
00:18:38.699 --> 00:18:53.124
Like, for example, on the 0T. Okay so then this is a 1 0T this the expression this is 1 in the next example is there a pipe? 1 is a 1 0T. Okay. So here you, when your Pi on 0T. Okay.

127
00:18:53.364 --> 00:19:08.183
So what do you got is you flip to 1 in the here and when your Pi to the quantum cubic to 1, you flip to 0T. Okay you can also pull for you can have a Super position for 1.

128
00:19:08.213 --> 00:19:12.233
0T. Okay, so you'll find out the poly X or prater.

129
00:19:12.598 --> 00:19:26.189
Effectively change, um, the like, our portal. So therefore, Polly, um, so to the updater so therefore, X, our operator is often called a, not.

130
00:19:26.189 --> 00:19:39.388
Data or not operator so it's a big flip operator. Okay, so, let's see next let's look at a party. Why in the policy operators are here I show the party. Why? Um.

131
00:19:39.864 --> 00:19:54.413
The basic operation. Okay. And it was interesting. I want to show you is like a policy. So, this is the policy matrix apply, um, like, uh, opry's status on 0T and the state of 1. so, what you'll find out is.

132
00:19:55.019 --> 00:20:02.663
Well, I'll please stay 0 0T. Okay when they are pretty on stage 1, because a negative 1. okay.

133
00:20:02.663 --> 00:20:17.544
So, here, are we only use the, um, like, um, you know, consider the quantum standard is along the on the basis, but it can be any arbitrary, like a direction. So, this is when we have returning arbitrary.

134
00:20:19.673 --> 00:20:33.443
Condoms data is a basic superposition of war 0T state and the 1 data. Okay. So when you have to see operator Pi interrupt, so you'll find out 1 of them. Okay. The 0T doesn't change. Okay.

135
00:20:33.443 --> 00:20:44.663
The, uh, the state 1 is always become inactive. So, it's basically to say, I'll pray to achieve the feast sleep data. Okay so this is the so here we, um.

136
00:20:44.999 --> 00:20:53.308
You've seen them poly X and the Isaac example to connect um.

137
00:20:53.308 --> 00:20:58.229
The quantum gater, the quantum digital gator with the classical.

138
00:20:58.229 --> 00:21:12.503
A digital gator, so this is how we realize the quantum theory they'll get into like, a party made using that. Basically you return matrix and the special example is a panel matrix. Okay.

139
00:21:13.679 --> 00:21:18.598
All right, so the basic content, certainly the diagram.

140
00:21:18.598 --> 00:21:29.699
So probably go to our single arbitrary qbtr. So, here you have the arbitrary quantums data. So, this is in okay. Passing through the.

141
00:21:29.699 --> 00:21:44.094
A operator, so this is the audio garter. Okay. So that's why we are not data. So this is a, um, you know, can be a again. So arbitrary quantum states are failing passing through why this is.

142
00:21:45.173 --> 00:21:57.084
So this is and see, we hide our phase 3 upon quantum gator. Okay. Here I also list another 1 commonly used in quantum.

143
00:21:57.388 --> 00:22:01.229
Um, data okay, quantum operator is Carter.

144
00:22:01.229 --> 00:22:14.243
Uh, okay, so this is the, this is the outer so the matrix, the form of the matrix is has the following.

145
00:22:14.273 --> 00:22:18.203
Okay so this is also quite a commonly used parameter.

146
00:22:18.628 --> 00:22:22.949
Okay, so this is a.

147
00:22:22.949 --> 00:22:37.193
Basically, we just finished the very basic introduction how the quantum gator was constructed for a single quantum like, uh, for 1 qbtr. Okay. For Tokyo.

148
00:22:37.193 --> 00:22:50.003
But now, covered here. But you probably can find it in many our reference box. Okay. So, next I want to show you how we're going to construct the single quantum.

149
00:22:50.459 --> 00:22:56.159
Um, gator are in optical circuit. All right so here.

150
00:22:56.159 --> 00:22:59.278
This is like a, um.

151
00:22:59.278 --> 00:23:13.794
Backup 2 connected. So this is important. This you can consider is a web wall, optimal fiber. So okay, so the opt. This is too important. Here is the splitter typically uses 3 DB, a splitter.

152
00:23:13.794 --> 00:23:26.183
Last means the optical power been split into 2 arm 2 arms evenly in both these arms. So we put a face shifter. Okay. We can control the shift from 0T to pay by.

153
00:23:27.358 --> 00:23:36.328
A plane, maybe a voltage in May or maybe later to control the light to pass ensure that fits shift. Okay so this is the output.

154
00:23:36.413 --> 00:23:46.374
This is our portal. Okay. So if you recall the classes, you're maybe phase 1 or physics, 2 classroom you'll probably learn the Michaelson.

155
00:23:46.584 --> 00:24:00.894
Inter so, basically, this is the MicroCenter parameter more integrated. The form. Okay. Or you can see is in a 2 dimensional. Okay. Form. Okay.

156
00:24:00.894 --> 00:24:14.814
Instead of a free spacer. Right? So this is again this is important. We can actually mathematically express all 1 and auto output as a function for you. I won I took passing through this.

157
00:24:15.358 --> 00:24:20.368
Transform matrix of okay, these transfer matrix later and we can prove what.

158
00:24:20.368 --> 00:24:28.019
Is just a unitary matrix. Okay, so I'll show you the math. 1st, so the trend for matrix of this.

159
00:24:28.403 --> 00:24:42.834
Circle the construction why is this by default okay so this is the construction for the circuit. That is 7 5 and the 5th shift herself. So this entire transfer matrix can be expressly new.

160
00:24:43.193 --> 00:24:51.443
Okay so we can combine the forms. We found out all 1 old tool. Is equal to you times.

161
00:24:52.078 --> 00:25:01.769
I want to input the matrix and this can be approved. Okay mathematical approved. This is indeed a unitary matrix. So this.

162
00:25:01.769 --> 00:25:15.959
Whole layout structure will became the unitary Matrix, the basic building blocks of the Atlantic space quantum computing. Okay. So here, I would like to show you this.

163
00:25:15.959 --> 00:25:20.489
Okay, so this is actually a circular way actually built a.

164
00:25:20.489 --> 00:25:35.453
Okay, in the semiconductor substrate. Okay. There's also some, um, you see this metal pad this manual pad is the heating elements are used to the hating element until we can control see that in the file. Okay.

165
00:25:35.453 --> 00:25:39.534
So this is some more detailed picture how these things are constructed.

166
00:25:40.318 --> 00:25:49.828
Okay all right so this is some measurements. Okay so measurements that we performed our escape this detail. Okay. Um, so here.

167
00:25:49.828 --> 00:26:03.653
I'm like, I would like to show you how the format for unity matrix is constructed from 2 by 2 unit matrix. So, here I mention here, this is a basic opinion barcode 2 by 2 neutral matrix here.

168
00:26:05.183 --> 00:26:12.173
When we say, if you have a degree of 2. okay, 2 cubic you need a 4 by 4 union matrix.

169
00:26:12.173 --> 00:26:24.054
So if you go to 3 qbr you need to to the 3, 8 by 8, a unit or a matrix or so all these things is a build upon a network of 2 by 2 matrix.

170
00:26:24.054 --> 00:26:36.144
If you look at the detail here, this is the T1, the 1st, unitary Matrix, the 2nd unit matrix. So, here we build so by put them in our right order. Okay.

171
00:26:36.173 --> 00:26:47.364
We can build a 4 by 4 inventory matrix and also, this is now the only way. So, basically, if you want to build a 4 by 4 matrix so there's multiple.

172
00:26:48.923 --> 00:27:00.624
Ways that you can construct to your offer right? For from a 2 by 2 unit matrix. And here I'd show 2 examples. So you can actually, um, like, um, you know, proof.

173
00:27:00.653 --> 00:27:09.263
It's it can be so easy, you have to have the save and find the fit shifter correctly to achieve the, um, disseminating the function.

174
00:27:09.443 --> 00:27:24.084
So, I won't go into the detail, but there's a procedure mathematically can follow to set up correctly to generate these 4 by 4 union industry matrix from 2 by 2. okay. This is again. Some, like, actually circular layout.

175
00:27:24.624 --> 00:27:26.003
We, we put down.

176
00:27:26.608 --> 00:27:38.848
Like, you know, the design cater, so this is we have 2 fabrication already, so this is actually the new fabrication we'll send out some. Okay, so here.

177
00:27:38.848 --> 00:27:44.189
How does it race through this? So the matrix operation are universal.

178
00:27:44.189 --> 00:27:56.094
A university unitary for the devices, so different Matrix, the conversation approach. So there's estimation there are some multiple decompensation approaches so you can have different the construction.

179
00:27:56.094 --> 00:28:00.233
So, the goal is to have the minimum number all form.

180
00:28:00.568 --> 00:28:14.273
The basic of building block 2 by 2. okay. So then if we reduce the number of basically 2 by 2 information, who can reduce the whole dimension of this chip layout. So, the good news about this. Okay.

181
00:28:14.273 --> 00:28:18.023
1, extremely attractive feature of this. So we can always.

182
00:28:18.568 --> 00:28:23.398
On okay, so here we can always.

183
00:28:23.398 --> 00:28:33.023
Go back to your changes, sit on the fly, so only change the seat of fire. We change the unit to Matrix property. So what that means means that we're programming.

184
00:28:33.503 --> 00:28:38.604
So basically we can easily change by changes citizen file and then program.

185
00:28:38.939 --> 00:28:42.534
The entire quantum gator circuit,

186
00:28:42.534 --> 00:28:43.284
so okay,

187
00:28:43.284 --> 00:28:51.713
so that's why this is very it became very attractive for okay if you want to know more detail,

188
00:28:51.834 --> 00:29:00.443
how the 2 by 2 Unitrin matrix can be built to for I forward even larger array of a union to Matrix.

189
00:29:00.804 --> 00:29:13.554
So here's a link okay. To the more detailed information. Okay, so I'll skip this 1. this is more about the physics of the device here. I'll give you an example actually.

190
00:29:14.578 --> 00:29:20.818
Of all 4 researchers from Australia from British, from China.

191
00:29:20.818 --> 00:29:25.739
They, they work together as team constructed this for tanika.

192
00:29:25.739 --> 00:29:30.239
Um, quantum computing, like a cheaper, the whole cheap 1.

193
00:29:30.239 --> 00:29:35.423
Is powder of for less than 10 millimeter reserved. I'm pretty smart. Okay.

194
00:29:35.634 --> 00:29:49.854
So they have like, a, they demonstrated to Cuba this circuit that can be expanded to into so, and so on. So, here they just show their 1st, prototype of 2 qbtr.

195
00:29:50.183 --> 00:30:00.683
So, the optical wave coming from here, okay going through these 2 passes is siblings a lighter. Okay so they use a, what do we call for Wave? Mixing.

196
00:30:01.199 --> 00:30:15.054
Uh, which is, um, take advantage of the nonlinear property of Silicon okay. To generate a full comparison. Okay. Basically, you're generating entanglement that with the full transfer. Okay, so, then they create a 2 qbtr okay.

197
00:30:15.054 --> 00:30:24.173
From Tokyo button. So, here, the 1st part is the generator call clocked in tanglement in tanglement.

198
00:30:24.683 --> 00:30:36.233
The 2 is the 2nd, apparently, is actually preparation of the initial single qbtr state and then next is implemented the single copious operator.

199
00:30:37.044 --> 00:30:49.884
If you look carefully here, you identify a number or for 2 by 2 unit or a matrix that basically the max out her into parameter April. Okay. So this achieves the gate function with operator function.

200
00:30:50.153 --> 00:31:03.054
So, after that, there's like a linear combination stage and then last part is the change in the measurement basis for basically prepare for measurement. So you see here, this.

201
00:31:03.358 --> 00:31:10.409
This might online this is the collection so you see this connector and so this is eventually.

202
00:31:10.409 --> 00:31:18.449
Connected to a eternal controller board or control panel so allow you to program it.

203
00:31:18.449 --> 00:31:29.903
To program them, like the safe and the fire, the data condition so you can achieve the design, like quantum computing function. Okay.

204
00:31:29.903 --> 00:31:38.544
So, yeah, here I listed the chief size is actually 7.1 millimeter lens and the width is 1.2 millimeter. So it's actually a lesson.

205
00:31:39.989 --> 00:31:49.078
Like, 1 centimeter is a pretty small. Okay. All right. So, um, in this on this shape, would they utilized.

206
00:31:49.284 --> 00:32:02.604
Um, like a, for pump. Okay so this is more optics is all scape. So they use the 50 8218. we've got the crossers and 14 optical.

207
00:32:04.709 --> 00:32:08.578
Um, greeting cup versa. So it's, um.

208
00:32:09.473 --> 00:32:22.344
Can be pretty complicated, but again, the good news is the whole thing is operating room temperature. Okay. So it's a pretty noise bastion 2 against the environment environmental fluctuation.

209
00:32:22.703 --> 00:32:26.634
So, this is a paper from this group. Okay. He's the leading answer.

210
00:32:26.969 --> 00:32:36.148
Okay, so that's pretty much my very quick and the brief introduction of for how to build a.

211
00:32:36.148 --> 00:32:42.868
A quantum computer for how many? But chips. Okay. So the basic.

212
00:32:42.868 --> 00:32:52.048
Quantum computing, like our discussions come from this David maker may. Okay so this is a pretty.

213
00:32:52.048 --> 00:32:56.759
You know, quantum computing award is relative to old, but still like, uh.

214
00:32:56.759 --> 00:33:00.028
Um, fairly useful bulka yeah.

215
00:33:03.449 --> 00:33:11.038
Okay, I guess that's it. Any questions. Yeah, thank you. Very much. Shop floor is open.

216
00:33:14.608 --> 00:33:20.788
Well, I'll start with a question, so yes, you can get out you can put a gateway between any pair of Cubics in this obviously.

217
00:33:21.929 --> 00:33:30.838
Yes, yeah, this is the basic 2 by 2 unitary matrix. So this is a 4 by 4. this is basically the operator.

218
00:33:30.838 --> 00:33:38.909
Yeah, if you want to expand the, the operator, okay you just, you know, just connect them. Yeah.

219
00:33:38.909 --> 00:33:42.058
Have different stages.

220
00:33:44.548 --> 00:33:51.628
Is there any, like, I'm thinking some of the problems, like the J. J. have with D, coherence time and so on is that a problem here?

221
00:33:51.628 --> 00:33:55.409
Well.

222
00:33:55.409 --> 00:34:02.098
Like, in the mechanics award, the same we very most is a coherent time.

223
00:34:02.098 --> 00:34:06.118
So, Michael, I said to them.

224
00:34:06.118 --> 00:34:20.724
The circuit, so, yeah, when we 1st, do we want the, like, the coherent time to be as nice as possible? That's how long we can maintain a likely entanglement before. It's just a lasting noise. So that's actually 1 of the biggest problem.

225
00:34:20.844 --> 00:34:26.333
1 of the problem actually in using for tonics it for like a quantum computing.

226
00:34:28.679 --> 00:34:35.458
Okay, is there any obvious limit to how big? Oh, Congress got a question.

227
00:34:35.458 --> 00:34:40.918
Okay, so, let me see what limits the number of cubic to.

228
00:34:41.844 --> 00:34:42.893
Expand on this,

229
00:34:44.244 --> 00:34:48.563
plus is 1 because every stage there's a little bit allows,

230
00:34:50.634 --> 00:34:54.114
maybe every so this is a so basically,

231
00:34:54.114 --> 00:35:03.173
every of these may have maybe point 1 dB of loss accumulatively you could have relatively large loss.

232
00:35:03.594 --> 00:35:17.934
But I spoke to those people who are really experts have been working on this for 10 years seems so far they are now Super concerned about loss. Okay. They are more concerned about loss of shape.

233
00:35:17.994 --> 00:35:20.364
So when you connect, for example, the optical source.

234
00:35:20.849 --> 00:35:35.693
To the table. Okay so that's actually 1 part, uh, you know, there's may easily encounter a couple of dB of velocity. Okay. 3 D base means your or cut your power by 50%. So that's most of their concern here is to cubital.

235
00:35:35.693 --> 00:35:49.043
Well, if you want to build a for qbtr. Okay, so here's 3. here's 3. DB, Lawson, you add them together is going to be 60 days, or if you're making this out to you make it as a tank feel better.

236
00:35:49.043 --> 00:36:01.704
So the estimation is the company loss between will be shouldn't be control point 1. DB. So to maintain a high fidelity of the table by right now.

237
00:36:02.579 --> 00:36:09.478
On technology, we can readily achieve these so 1 DB, loss at the interface. So then how we can get from.

238
00:36:09.478 --> 00:36:19.074
Perhaps the 1 thing to point when the interface allows this a challenge is a long term challenging in terms of community. We haven't really figured out a good way to do it. Yet.

239
00:36:20.184 --> 00:36:30.594
Another limitation is you see the cheap site at some point 1 on 1.9 millimeter. So, you can't imagine when this this is only a 4 by 4.

240
00:36:31.469 --> 00:36:41.123
Okay, Eva, we make it larger like, you know, now to Q3 here, but it's a 10 qbtr this size of this dimension of this shape.

241
00:36:41.123 --> 00:36:54.653
We're going to expand exponentially and at some point it's going to exceed your, um, you know, your wafer account offer. So that's can be also a like a, you know, payment the current technology.

242
00:36:54.653 --> 00:36:56.934
So then people may consider repeaters.

243
00:36:59.159 --> 00:37:02.248
This the engineering problem, engineering problem.

244
00:37:02.248 --> 00:37:06.958
Okay, I hope I answer a question.

245
00:37:07.554 --> 00:37:21.713
So limits. Okay. Yeah noise is a big problem and the coherence. Okay. So when he lost his care coherence, that's another, uh, like, uh, can be another problem. Okay from John How's the coherent time modified?

246
00:37:22.079 --> 00:37:32.693
And the factor, the meeting, the full time, coherent time, the current time, sometimes it is come from the imperfection of application.

247
00:37:33.054 --> 00:37:47.844
So, for example, we want to have smooth wall, but sometimes we don't have the smos while we have well, maybe maybe a task or some imperfection during the following. Our process will end up a notch.

248
00:37:48.353 --> 00:37:56.304
Okay so your mobile slightly change okay. Compared to others. So, they don't have this. Okay so all these things will.

249
00:37:56.639 --> 00:38:02.369
Um, somewhat a D, coherent, uh, the full times when you go through to pass.

250
00:38:02.369 --> 00:38:05.369
Okay.

251
00:38:05.369 --> 00:38:10.318
Uh, let me say another question franchise also. How to Cupid.

252
00:38:10.318 --> 00:38:17.909
A case realize so that's a good question. Yes. You'll actually answer yourself. Yes, it's called the control. The data.

253
00:38:17.909 --> 00:38:30.690
The eyes control data, your, um, your basis basically had changed that I remember I, I show that it's either 0. 1, right? So, if it's become too cobit, then you will have 10001 0.

254
00:38:30.690 --> 00:38:38.250
1, 1, I see the basic bases. Okay. Yes. You're everything going to be expanded. Yeah.

255
00:38:38.250 --> 00:38:42.329
Okay, so from, um.

256
00:38:42.925 --> 00:38:55.885
How faster can find the say to be manipulated on the date. Okay. That's also a fantastic question. This face. Okay. Depends. On what technology we are using.

257
00:38:55.885 --> 00:39:06.264
So, here we use the technology is a user Peter. The reason we use hazer is because of hater you just gentilly hit Apple 1.

258
00:39:06.599 --> 00:39:10.710
Side of the arm. Okay. Introduce the list.

259
00:39:10.710 --> 00:39:20.579
Optical loss to this whole, like, Matrix. So using heater technology is operator roughly a.

260
00:39:20.579 --> 00:39:28.320
I would say maybe 100 megahertz is now we're faster. Um, so basically, if you want to program it.

261
00:39:28.320 --> 00:39:35.125
Okay, it can be relatively slow by the, but this is so far not a big concern to the community.

262
00:39:35.364 --> 00:39:44.934
So, basically, you know, the pipe fits shifter, like a file and the scenario is achieved by these micro wires through external control.

263
00:39:45.985 --> 00:39:56.664
But since you're when we have the quantum computing, we used to deal with those hard a mess. Mathematical problems is take a long time for computing.

264
00:39:56.875 --> 00:40:09.985
You want to program it to certain data then settle settled with data then computing to carry out. Then later on, you maybe want to work out different the problem and change the parameters. Okay. Basically change your data.

265
00:40:10.014 --> 00:40:16.764
Maybe instead of 1 together now, gate, you want to get into some other data. So then you can change the programming. So.

266
00:40:17.574 --> 00:40:31.135
And if we use a different technology, for example, use the voltage to control the, the seat, and the fire, that can be in very fast, it can be as faster as they say, speed wise is a 10 gigahertz easy.

267
00:40:31.164 --> 00:40:32.905
And that means is roughly about.

268
00:40:33.599 --> 00:40:42.179
Like, a 100 people, 2nd, like a change in time. So right now, using this is on the order or form I would say.

269
00:40:42.179 --> 00:40:45.360
Summer millisecond was submitted a 2nd.

270
00:40:45.360 --> 00:40:48.420
To change it to reprogram.

271
00:40:51.510 --> 00:41:02.280
Okay, how do we do a measurement again?

272
00:41:02.280 --> 00:41:10.949
Yeah, the measurement is also now trivial. Yeah, well, for the.

273
00:41:10.949 --> 00:41:17.010
Yeah, it's, uh, I'm, I'm not too sure how I'm.

274
00:41:17.010 --> 00:41:29.579
Okay, so free entanglement to confirm that we use about, like, a testing into here. So they yeah, it's, um, it's quite complicated.

275
00:41:30.659 --> 00:41:35.699
So, I don't have the detail to provide here, but I, what I know is.

276
00:41:35.699 --> 00:41:40.619
Even for the measurement parameter, uh, it's, uh, can be quite complicated.

277
00:41:42.420 --> 00:41:55.619
Okay, so so this is very promising, but there's still some engineering involved, right? But sorry computing. That's a fair way to say. Yes. Yeah. So.

278
00:41:57.300 --> 00:42:10.855
Now, all the other technologies also seems as right now is in a quantum computing world, and nobody knows which technology go into, whether that's why they try different things. Yeah, we tried different approach also.

279
00:42:10.855 --> 00:42:24.355
People try microwave, right? They, they, they use like, a, to microwave to generate entanglement. So that's a totally different way people's that says more promising. So familiar with that.

280
00:42:24.355 --> 00:42:35.454
1, Karen, I think he's seeing Perdue University. They are very like, because they have longterm working on.

281
00:42:38.130 --> 00:42:50.815
Maybe a oscillator quantum quantum, like, uh, like atomic o'clock or so they are they have been working with for a long term.

282
00:42:51.054 --> 00:43:05.215
So, yes, they, they use a fairly different approach to achieve data. So this 1, you've seen the full time. tanglement is actually relatively straightforward. You just generate in Tango the last, your optical source in the past.

283
00:43:05.215 --> 00:43:08.275
So there's a unitary matrix then do the calculation.

284
00:43:11.875 --> 00:43:24.114
Let me see another question from the chief it seems like there are open, ended input and output. Do they do anything? It's not really open and data so.

285
00:43:24.809 --> 00:43:35.579
This is a passive cheaper. Okay. So we have to send the single knowing so optical signal. So this will come from the left. And so this is basically your laser.

286
00:43:35.579 --> 00:43:48.385
These are important for this particular chamber the optimal source to couple the lighting the light is is a visible lighter, but actually the computing wavelengths the is actually a 1550. okay.

287
00:43:48.385 --> 00:43:58.735
Telecom communication is a different this is achieved by the forward mixing this, like a clinical professor. Frankly, just ask them, unfortunately.

288
00:44:02.099 --> 00:44:13.469
Not, I didn't care for her good answer, but yes, there is some circuit after this. Actually, I should say it's for hannika equipment to do the measurement.

289
00:44:14.909 --> 00:44:29.635
Yes, this is a big part of the yeah. Didn't show on the trip, but so there's some testing equipment associated with data to basically figure out these about this is supposedly. Okay, so for a real quantum computer, these are pretty much at the end of it.

290
00:44:30.114 --> 00:44:44.875
Okay. Just basically how you're going to interpret data. So I assume here, you need to have a 40 attacked her because we don't typically directly measure full time. So we're going to have a detector here by measuring because everything you actually have to convert to intensity.

291
00:44:44.994 --> 00:44:51.355
Okay. We're usually intensity to translate. What do you exactly your quantum state?

292
00:44:51.750 --> 00:44:59.730
Quantum States are so that's the physical end of the physical layer of the quantum computing.

293
00:45:00.750 --> 00:45:04.650
Hey, thank you. Anyone else have any questions.

294
00:45:07.110 --> 00:45:21.025
No, well, thank you very much. This is very great. So okay, thank you. Thank progressive frankly. Nice me to share some of these interesting quantum computing.

295
00:45:21.025 --> 00:45:33.954
We see you're on your class. I'm very impressed by all the questions they ask. I'm like, I just last thing. Okay. So, I'm going to teach offer a class called electronics.

296
00:45:34.500 --> 00:45:46.409
Intro to upload electronics technology in spring semester. So any students that you are interested in, learning more about so you're welcome to sound for the class. I know today's the last day.

297
00:45:46.409 --> 00:45:54.119
Of phase 1. okay. Okay thanks again. Yeah. Give me a blurb about that and I'll put it on the blog right away.

298
00:45:54.119 --> 00:46:05.730
Okay, I'll send you the course syllabus. Okay. Yeah. Send an email with the way. Okay. And this is a link or something. I'll update the blog later so that we, the students can see that in writing. So.

299
00:46:05.730 --> 00:46:09.030
Okay, sounds good. Thank you so much. All right.

300
00:46:09.030 --> 00:46:15.929
Okay, so I will just log out and let you continue your class. Okay Thank you.

301
00:46:15.929 --> 00:46:21.389
Yep bye so.

302
00:46:24.960 --> 00:46:31.500
So class we have a lot of expertise at our in quantum computing.

303
00:46:31.500 --> 00:46:36.989
So, if I can go back to the.

304
00:46:44.010 --> 00:46:50.250
Oops, here we go at.

305
00:46:52.530 --> 00:46:56.909
Okay.

306
00:46:56.909 --> 00:47:04.110
Um, well, the thing, so I put a really simple homework on due.

307
00:47:04.110 --> 00:47:08.369
In a long time, just do write a.

308
00:47:08.369 --> 00:47:15.239
Progress report due on on your project, and just to motivate you to start doing something before it's due.

309
00:47:15.239 --> 00:47:25.710
Okay, other things are there is this good keynote talk on trapped ion, quantum computing.

310
00:47:25.710 --> 00:47:30.090
You tell me, we can look at it started of it now for a little while.

311
00:47:30.090 --> 00:47:41.550
Or we can end the class early, and you can look at it on your own. I would encourage you to look at it. He's a good speaker. This tells us more about a another technology and actually.

312
00:47:41.550 --> 00:47:46.199
The floor is.

313
00:47:46.199 --> 00:47:51.480
Open, would you like to.

314
00:47:51.480 --> 00:47:56.699
Would you like to see this now or would you like to end the class early and.

315
00:47:56.699 --> 00:48:00.659
You know, work on your own stuff so.

316
00:48:01.710 --> 00:48:08.429
Silence.

317
00:48:10.019 --> 00:48:14.369
Silence.

318
00:48:19.230 --> 00:48:26.969
Okay.

319
00:48:26.969 --> 00:48:30.989
Let's see.

320
00:48:30.989 --> 00:48:38.519
Started it early then and learn some more the reason I'm building the course again from.

321
00:48:38.519 --> 00:48:48.480
You know, stuff I've found around the web is they bring in videos I bring in visuals and they give you, I think, a better vectoring experience than.

322
00:48:48.480 --> 00:48:54.449
Then I could give you on these various topics because they've spent a lot of time and then they know the stuff better. Obviously.

323
00:48:54.449 --> 00:48:58.440
So that.

324
00:48:58.440 --> 00:49:04.559
I see.

325
00:49:04.559 --> 00:49:08.519
You start this.

326
00:49:08.519 --> 00:49:12.000
Have this cast of speakers later in the day.

327
00:49:12.000 --> 00:49:15.059
And I will actually, I realize, I'll probably connect to all of them.

328
00:49:15.059 --> 00:49:18.329
In my in my presentation here.

329
00:49:18.329 --> 00:49:21.480
And I added a couple of words to the title funnel competing with Adams.

330
00:49:21.480 --> 00:49:27.719
Optics, and if I were to cut to the chase, I would say that the cubic part, the quantum part's done.

331
00:49:27.719 --> 00:49:32.369
Those are the atoms there are atomic clocks and I say they're perfect, meaning their.

332
00:49:32.369 --> 00:49:35.940
Much better than we need them to be all of the challenges.

333
00:49:35.940 --> 00:49:40.739
Remain in the classical controllers that that in our technology.

334
00:49:40.739 --> 00:49:46.289
Means optics, so this is a picture of a semiconductor, a chunk of Silicon.

335
00:49:46.289 --> 00:49:50.309
Later by goal, that's part of the controller. These are just electrodes.

336
00:49:50.309 --> 00:49:53.849
That allow individual atomic ions to hover.

337
00:49:53.849 --> 00:49:59.159
Float above this trap and they're glowing because we're shining C. W laser.

338
00:49:59.159 --> 00:50:03.750
On the atoms and they're flourishing and the control of them has other lasers that.

339
00:50:03.750 --> 00:50:08.550
Execute quantum Gates and I'll dive a little bit into that technology.

340
00:50:08.550 --> 00:50:13.260
But 1st, since this is a talk on quantum computing, I, but I would start.

341
00:50:13.260 --> 00:50:17.190
On very broad towns, defining terms and so forth.

342
00:50:17.190 --> 00:50:22.889
And this may sound a little overhype and I don't want to I want to stress it that much, but.

343
00:50:22.889 --> 00:50:26.369
The high performance computing community is.

344
00:50:26.369 --> 00:50:29.400
Running scared a little bit now, because we all know that.

345
00:50:29.400 --> 00:50:34.980
More as law will not continue forever, at least how it has over the last many decades because the individual.

346
00:50:34.980 --> 00:50:41.340
Transistor elements are getting too small and we can see, even in the last few years, it's starting to saturate.

347
00:50:41.340 --> 00:50:48.239
There are many other forms of computing that are starting to make the light of day. Norm are for computing and so forth.

348
00:50:48.239 --> 00:50:52.349
Quantum computing is 1 of them and it's not exactly clear that this.

349
00:50:52.349 --> 00:50:56.880
You know what the, what the coverage of quantum computing will be.

350
00:50:56.880 --> 00:51:01.829
But I put this plot up here because in the days when Moore's law started.

351
00:51:01.829 --> 00:51:06.389
We're early days of transistor development, solid state transition.

352
00:51:06.389 --> 00:51:09.719
And 1, Richard Feynman was.

353
00:51:09.719 --> 00:51:12.929
Intrigued by the fact that.

354
00:51:12.929 --> 00:51:16.139
Kidding elements to be built from solid state.

355
00:51:16.139 --> 00:51:21.059
And he imagined the day when they could be struck down so small.

356
00:51:21.059 --> 00:51:28.500
That each transistors, a few molecules or few atoms, and that's where we'll be. And if we follow Moore's law, now, that's where we'll be in the next.

357
00:51:28.500 --> 00:51:33.690
Couple decades, but what finance said in this lecture, and in 1959.

358
00:51:33.690 --> 00:51:37.650
Incredibly visionary he said, when we get the circuits of individual atoms.

359
00:51:37.650 --> 00:51:42.539
There will be completely new opportunities for design and I think at that time, it wasn't.

360
00:51:42.539 --> 00:51:47.130
Knowing what those opportunities would be, I think, by now over the last few decades.

361
00:51:47.130 --> 00:51:52.769
We understand that opportunity is quantum computing and the reasons there's new opportunities of course, is that.

362
00:51:52.769 --> 00:51:56.219
When you get down to the simple forms of matter with very few degrees of freedom.

363
00:51:56.219 --> 00:52:06.030
The laws of physics change, and those laws of physics, quantum physics may allow us to compute in new ways. And that's the opportunity.

364
00:52:06.030 --> 00:52:09.570
So very briefly, most of you.

365
00:52:09.570 --> 00:52:13.349
This is probably a review from you, but quantum.

366
00:52:13.349 --> 00:52:16.500
Are very much like classical bits, zeros ones, except.

367
00:52:16.500 --> 00:52:20.340
Because of the quantum superposition principle.

368
00:52:20.340 --> 00:52:26.699
Information can be stored in parallel and even at a single bit level. So quantify the superposition of 0T and 1.

369
00:52:26.699 --> 00:52:36.269
And it's very hard to draw that because we live in a classical world here and this is a classical projector but that's a single electron of course, in 2 States at the same time.

370
00:52:36.269 --> 00:52:40.679
And it has weightings of the 2 States and what's.

371
00:52:42.030 --> 00:52:47.639
There are lots of ways to couch. The strange opportunities of quantum 1 of them. Is that.

372
00:52:47.639 --> 00:52:52.019
When you measure a quantum state, you have to use probabilities. Probabilities have to.

373
00:52:52.019 --> 00:52:59.550
The info, they come out of scenario in other series of nature probabilities are used because of our ignorance. But in quantum mechanics, you have to use it.

374
00:52:59.550 --> 00:53:09.449
So that's 1, strange feature of quantum mechanics. And so what happens of course, is that the state collapses to 1 of 2, definite States, depending on the waitings of the original superposition.

375
00:53:09.449 --> 00:53:13.440
Now, it gets interesting when you put many cubes together.

376
00:53:13.440 --> 00:53:17.010
Even 2, there's something called quantum entanglement.

377
00:53:17.010 --> 00:53:21.360
And in this case, this is a particular entangled state of 2 quantum beds.

378
00:53:21.360 --> 00:53:24.659
And you can see when you measure such a system, you'll see.

379
00:53:24.659 --> 00:53:27.900
You'll see 1 of the 2 States with perfect correlations.

380
00:53:27.900 --> 00:53:35.610
Even though individually, they're totally random. Okay. So the concept of entanglement is is in a sense.

381
00:53:35.610 --> 00:53:39.840
Having connections or wires without.

382
00:53:39.840 --> 00:53:46.260
Due to the structure of quantum mechanics and that, to me is the essence of quantum computing we're using connections that aren't there.

383
00:53:46.260 --> 00:53:50.820
I should say that with quotes because there's lots of subtleties involved in that.

384
00:53:50.820 --> 00:53:55.679
So, the story of quantum computing, and where it came from and.

385
00:53:55.679 --> 00:54:00.510
And how you could do something interesting is it's a very subtle 1 and I'm not going be able to touch on.

386
00:54:00.510 --> 00:54:04.769
A whole lot of it. I like to think of it as a good news. Bad news. Good story.

387
00:54:04.769 --> 00:54:12.090
And the 1st piece, the good news might be apparent from this. The fact is that we get exponential growth in storage when we add Cubans together.

388
00:54:12.090 --> 00:54:15.719
3, KPIs there, 8 States, you can store.

389
00:54:15.719 --> 00:54:21.150
With lots more canvas, you get this exponential growth, every additional cube it gives you twice.

390
00:54:21.150 --> 00:54:24.150
The states space to store information, if you want.

391
00:54:24.150 --> 00:54:28.590
And, for instance, if you wanted to compute the function, you could do it in parallel.

392
00:54:28.590 --> 00:54:34.170
With the same device with only 1 input, but the input has a superposition of all the numbers. Okay.

393
00:54:34.170 --> 00:54:38.130
And when because it's exponential growth, it doesn't take.

394
00:54:38.130 --> 00:54:42.869
Not too much scaling to get something you could not think of doing without quantum.

395
00:54:42.869 --> 00:54:46.320
But just a few 100 inhibits you're representing.

396
00:54:46.320 --> 00:54:50.190
A number of configurations, it's more than the number of fundamental particles in the universe.

397
00:54:50.190 --> 00:54:53.909
So, even with Moore's law, even if more elaborate continue forever.

398
00:54:53.909 --> 00:54:57.329
You still would not be able to approach what you could do.

399
00:54:57.329 --> 00:55:01.650
Potentially with a quantum computer with just a few 100. that's the great good news.

400
00:55:01.650 --> 00:55:05.070
It's almost too good to be true. And it is when you think about.

401
00:55:05.070 --> 00:55:09.750
The computer you have to measure it of course. And when you measure it, that's bad. News.

402
00:55:09.750 --> 00:55:14.670
And the more complex, the superposition, the less information you get out.

403
00:55:14.670 --> 00:55:21.119
Right. So you have 2 to 300 inputs and outputs and you make a measurement and you'll get 1 answer and you don't know what it was.

404
00:55:21.119 --> 00:55:26.130
Random that seems very bad. I'm sure finally realize this and 59.

405
00:55:26.130 --> 00:55:32.010
And didn't know how to put it together, but in fact, 20 or 30 years ago, there is a final piece of good news that was.

406
00:55:32.010 --> 00:55:35.250
I'd say how to find more by David than anybody else.

407
00:55:35.250 --> 00:55:46.559
And he showed that there are cases where you can have interferences between these amplitude. It's very hard to draw. But if you imagine all these inputs and superposition, these red dots are called quantum Gates.

408
00:55:46.559 --> 00:55:51.329
They allow very specific interferences to occur entanglement to be formed and so forth.

409
00:55:51.329 --> 00:55:55.559
And what happens here is you have this exponentially rich input state.

410
00:55:55.559 --> 00:56:00.239
It sort of it's forced down to 1 output or a few outputs.

411
00:56:00.239 --> 00:56:03.360
And that you can measure, you can measure in the.

412
00:56:03.360 --> 00:56:10.889
There's very little entropy in that output state. And the point there is that when you make that measurement, in some cases, it can depend on all of these.

413
00:56:10.889 --> 00:56:15.539
Exponentially many inputs. Okay, so that's the basic idea of quantum computing and.

414
00:56:15.539 --> 00:56:19.500
I'm sweeping a lot under the rug because not every problem can be stated this way.

415
00:56:19.500 --> 00:56:22.559
What type of Gates do you need to do something? Interesting?

416
00:56:22.559 --> 00:56:27.239
Well, there are there are a few applications out there. I would say that.

417
00:56:27.239 --> 00:56:32.519
The, the killer app quantum computing is, you might now is factoring large numbers.

418
00:56:32.519 --> 00:56:37.590
This is pointed out by Peter shore about 2025 years ago.

419
00:56:38.670 --> 00:56:41.969
And he basically showed how you could build a quantum circuit.

420
00:56:41.969 --> 00:56:46.800
With a polynomial number of gates that would allow you to factor a big number and.

421
00:56:46.800 --> 00:56:52.409
What's remarkable is that classically the best known algorithms still scale exponentially with the size of the input.

422
00:56:52.409 --> 00:56:56.880
And so they're very, very broadly what he.

423
00:56:56.880 --> 00:57:00.480
Showed us that you could take a quantum way function with lots of.

424
00:57:00.480 --> 00:57:06.059
Lots of states this would this would have like, 6 or 7 cube whatever log base. 2 of 39 is.

425
00:57:06.059 --> 00:57:11.219
Maybe you need some more workspace and execute quantum Gates so that the way function does this.

426
00:57:11.219 --> 00:57:14.670
It's sort of coherently evolves.

427
00:57:14.670 --> 00:57:18.119
For the 2 factors of the number, you're trying to factor and then when you make a measurement.

428
00:57:18.119 --> 00:57:21.960
You can get information on all those inputs.

429
00:57:21.960 --> 00:57:28.139
Factoring and I would say the problem with factoring is that 2 factor interesting number.

430
00:57:28.139 --> 00:57:34.320
Requires thousands of bits you probably need air correction because you need you need.

431
00:57:34.320 --> 00:57:42.059
Billions of operations, and this is sort of like an analog computer that noise will add. So you have to do air correction. So you probably need trillions of operations.

432
00:57:42.059 --> 00:57:48.030
And millions of kibble so this problem is still far in the distance. It was back in 95.

433
00:57:48.030 --> 00:57:52.139
When i1st got in this field and it is now, it feels like it's the same distance away.

434
00:57:52.139 --> 00:57:59.070
But the good news is, there are other problems that are, I would say, even more interesting and more widespread than factoring.

435
00:57:59.070 --> 00:58:03.690
That have the flavor of having an output that globally depends on inputs.

436
00:58:03.690 --> 00:58:07.230
So, optimization is it's a little bit of a catch all.

437
00:58:07.230 --> 00:58:13.110
But when you optimize a very complex function, there's only 1 optimal answer or very few numbers of answers.

438
00:58:13.110 --> 00:58:19.050
And they can depend on all the inputs. This is a simple 1 if we have a function of just 2 variables, we can.

439
00:58:19.050 --> 00:58:22.860
Clearly see the minimum, but if you have a function of a 1000 variables.

440
00:58:22.860 --> 00:58:28.469
The configuration of those variables that results in the minimum or the maximum minimizing some cost function.

441
00:58:28.469 --> 00:58:31.590
That's a really hard problem and there are all kinds of comments are.

442
00:58:31.590 --> 00:58:36.030
Optimization problems from from finance from physics.

443
00:58:36.030 --> 00:58:39.179
From a chemistry that.

444
00:58:39.179 --> 00:58:42.300
Look, very interesting for quantum computer to solve.

445
00:58:42.300 --> 00:58:45.389
Not that said, we don't yet have a proof.

446
00:58:45.389 --> 00:58:49.739
That any 1 of these applications can be exactly solved on a quantum computer.

447
00:58:49.739 --> 00:58:53.940
There are headsets that quantum computers could be a heuristic that.

448
00:58:53.940 --> 00:58:57.960
It gives you a better answer than any classical approach could be.

449
00:58:57.960 --> 00:59:03.780
Traveling sales, which is my favorite, because it's very easy to depict. What's the shortest path between a bunch of cities?

450
00:59:03.780 --> 00:59:08.280
And the problem here is there's too many configurations once you get a few 100 cities.

451
00:59:08.280 --> 00:59:11.940
So, that type of a logistics problem is, I think where the field was headed.

452
00:59:11.940 --> 00:59:15.119
It's good because this is more broad than just background.

453
00:59:16.260 --> 00:59:25.230
So I want to talk about hardware because it will get some exotic performance computing and therefore demands pretty exotic hardware.

454
00:59:25.230 --> 00:59:30.119
To do the job. So, my colleague Bill Phillips is fine, was saying that.

455
00:59:30.119 --> 00:59:33.690
A quantum computer differs more from a classical computer.

456
00:59:33.690 --> 00:59:39.719
Then that classical computer differs from an advocates and what he means here is that these 2 machines follow the same.

457
00:59:39.719 --> 00:59:44.039
Abstract theory given by Alan Turing, they're both touring machines.

458
00:59:44.039 --> 00:59:47.519
1 is electronic with transistors it's very fast and.

459
00:59:47.519 --> 00:59:51.809
And you can have trillions of transistors here, but there's the same rules.

460
00:59:51.809 --> 00:59:54.960
Is totally different set of rules.

461
00:59:54.960 --> 01:00:02.159
And so, I guess, I would say corroborate to the statement, is that why would we ever expect on a computer to look anything like the classroom.

462
01:00:02.159 --> 01:00:05.489
It's that different and so that's why we have to.

463
01:00:05.489 --> 01:00:14.760
Maybe step back and look at anything that supports quantum coherence, anything that we use in quantum research to think about building devices from them.

464
01:00:14.760 --> 01:00:23.460
So, the last 5 or 8 years have been very exciting in this field, primarily, because we're starting to see systems engineering.

465
01:00:23.460 --> 01:00:27.719
Quantum systems, engineering, building a device that a 3rd party can use.

466
01:00:27.719 --> 01:00:33.119
Without knowing every detail about what's the side we need an application.

467
01:00:33.119 --> 01:00:39.840
But maybe we need a device 1st, that somebody can use that might not be an expert in quantum physics to use it.

468
01:00:39.840 --> 01:00:42.960
So superconducting bits is 1.

469
01:00:42.960 --> 01:00:46.530
1 path forward where superconducting currents.

470
01:00:46.530 --> 01:00:52.440
They sat there, they flow without this patient, so it's a perfectly interesting bit to think about.

471
01:00:52.440 --> 01:01:04.889
The scale problem, it's very hard because every cube is a little different there manmade and I would say the leading candidate is maybe even more exotic or less looking like a classical computers, not South state at all individual atoms.

472
01:01:04.889 --> 01:01:11.039
And this is what I want to talk about and the opportunities with atoms are many fold 1 of them is that.

473
01:01:11.039 --> 01:01:16.440
Atomic cubital. All right they're identical. They're not manmade. They're giving to us.

474
01:01:16.440 --> 01:01:25.380
These atoms are a ton of clocks, and they have the same features that atomic clocks have in terms of their longevity, their coherence and so forth.

475
01:01:25.380 --> 01:01:29.519
And, like I said, at the outset of the challenge is how to control them, how to make Gates.

476
01:01:29.519 --> 01:01:32.610
How to hold on to them and isolate them from the, from the environment.

477
01:01:32.610 --> 01:01:36.900
So, I want to talk a little bit about that and I'll note also that.

478
01:01:36.900 --> 01:01:44.940
Good industry is getting involved in this scale in a big way mostly with these 2 technologies that are a few others as well. And these are.

479
01:01:44.940 --> 01:01:51.539
Very well healed companies that are especially computer companies like IBM, Intel.

480
01:01:51.539 --> 01:01:58.079
Microsoft and Google that are looking to next new models of quantum computing and a host of start ups. Of course.

481
01:01:58.079 --> 01:02:01.349
The government here in the U. S. I've listed a few of the big investments.

482
01:02:01.349 --> 01:02:05.880
Some of them are in intelligence agencies, Department of Defense Department of energy and so forth.

483
01:02:06.929 --> 01:02:10.980
So now turning out to this technology of individual atoms.

484
01:02:10.980 --> 01:02:21.090
Dive into a little bit of atomic physics we store our KPIs within individual atoms. Very much. Like 2 levels are stored in the atomic clock.

485
01:02:21.090 --> 01:02:24.869
So, pick a good 2 little system that is insensitive to.

486
01:02:24.869 --> 01:02:28.199
External like nuclear, spend something like this.

487
01:02:28.199 --> 01:02:34.829
And that's a potential Cuba. Now, atomic guy is very interesting because they interact very strongly through their cool long force.

488
01:02:34.829 --> 01:02:38.280
I'll tell you a little bit how we harness that, but.

489
01:02:38.280 --> 01:02:44.400
Again, because they have very low idling errors, almost 0T and they're perfectly replicable.

490
01:02:44.400 --> 01:02:50.309
The atomic systems are scalable fundamentally scalable in a way that no solid state system.

491
01:02:50.309 --> 01:02:54.690
Evergreen now.

492
01:02:54.690 --> 01:02:59.789
Center saying that to control these atoms we need optical.

493
01:02:59.789 --> 01:03:03.750
Sources we need lasers to do this and this is.

494
01:03:03.750 --> 01:03:09.449
100% of the challenge, the technology I'm dealing with, so to give you a little idea here and also.

495
01:03:09.449 --> 01:03:16.980
Here's a little bit of history in this picture. The atomic cubic that I've been investing in for a long time is atomic utopian.

496
01:03:16.980 --> 01:03:21.119
Looks a lot like hydrogen actually spin 1, half nucleus in 1, outer electron.

497
01:03:21.119 --> 01:03:26.250
So, their hyperphile levels, and we store our cubic in the so called clock States of this Ion.

498
01:03:26.250 --> 01:03:29.969
And we drive coherent transitions with with the pulse laser.

499
01:03:29.969 --> 01:03:36.840
And the history there is, I come from the field of cold, atomic, physical lasers with.

500
01:03:36.840 --> 01:03:40.230
11 digits of stability and wavelength.

501
01:03:40.230 --> 01:03:43.500
So, I have my whole life until.

502
01:03:43.500 --> 01:03:47.880
When I went to Michigan in 2000, I was very scared of pulse. lazers.

503
01:03:47.880 --> 01:03:55.889
And, you know, it's nice to nice to speak in the same session that Donna and Gerard are. And I'll add that Phil, my colleague in Michigan.

504
01:03:55.889 --> 01:04:03.269
They got me less afraid of pulse, lasers and I have to say that this this particular band of pulse, laser 355 nanometers. It's.

505
01:04:03.269 --> 01:04:09.360
I'm sorry, it's not 7 seconds 10 seconds. It's too too broad bandwidth for us in 2 seconds.

506
01:04:09.360 --> 01:04:17.610
But this laser is very much I can see W laser. It's so well engineered because it's used for lift atrophy. It's an industrial standard. Coherent makes several 100 of these a year.

507
01:04:17.610 --> 01:04:25.289
And even though I love playing around with lasers, I used to do that a lot. I'm happy that this laser only has 1 knob and it's the.

508
01:04:25.289 --> 01:04:31.139
And that makes this a system, this laser runs for years and years without thinking about.

509
01:04:31.139 --> 01:04:34.139
That's very important and it has all the coherence we need.

510
01:04:34.139 --> 01:04:37.590
It actually is sort of a frequency, common narrow band frequency comb.

511
01:04:37.590 --> 01:04:41.519
That we, I won't talk too much more about it, but we were able to.

512
01:04:41.519 --> 01:04:46.440
Apparently drive this candidate and apply forces with this kind of a laser beam and yeah, it's pulse list.

513
01:04:47.519 --> 01:04:53.489
Not super high power, but okay, so I want to show a little animation on how we.

514
01:04:53.489 --> 01:04:59.460
Build a small register of kibbutz and untangle that evolving this type of a laser and.

515
01:04:59.460 --> 01:05:05.309
So the idea starts with a bunch of laser beams. This is a multi channel talk about that later. If you want.

516
01:05:05.309 --> 01:05:08.309
And these laser beams are all fed across this chip.

517
01:05:08.309 --> 01:05:12.900
Now, we 1st, initialize all these qbrs with the CW laser. That does optical pumping.

518
01:05:12.900 --> 01:05:17.940
And then we execute a quantum circuit so this is space and time we're going forward in time.

519
01:05:17.940 --> 01:05:25.650
Now, these, these lasers, these strict 5 nanometer lasers are basically optimal tweezers and we're going to hear from on the video on the physics of that.

520
01:05:25.650 --> 01:05:30.119
We're driving forces between these atoms and that's how we can couple. The information.

521
01:05:30.119 --> 01:05:35.250
And execute quantum Gates, the say, small quantum circuit that execute some.

522
01:05:35.250 --> 01:05:40.230
Some computational task at the end of the day we send another laser for measurement.

523
01:05:40.230 --> 01:05:45.900
And it's a very simple measurement if the Adam glows and 1, it's dark it's in 0T. So that's it.

524
01:05:45.900 --> 01:05:50.309
A lot of optics here in fact, that's, as I said, many times now.

525
01:05:50.309 --> 01:05:54.989
That's a 100% of our challenge is the optimal controller around these perfect Cubans.

526
01:05:56.159 --> 01:05:59.340
So this is an example of an algorithm.

527
01:05:59.340 --> 01:06:04.619
A very small 1 I won't get into details too much and involves 3. 4. this is 10.

528
01:06:04.619 --> 01:06:07.679
And it has on our 50 or so operations.

529
01:06:07.679 --> 01:06:14.730
These are head of our Gates these are icing Gates controlled Gates. I won't get into details, but this is a quantum computation.

530
01:06:14.730 --> 01:06:24.360
It actually it's an algorithm. It's called the hidden shift algorithm. It's, it's an Oracle algorithm. It's sort of useless, but it's a benchmark demonstration. You have 10 cubes you might think of.

531
01:06:24.360 --> 01:06:27.869
You know, measuring business of your system running this.

532
01:06:27.869 --> 01:06:31.320
Algorithm it says if you're given 2 functions.

533
01:06:31.320 --> 01:06:34.769
And they're identical, except the inputs are shifted by some unknown amount.

534
01:06:34.769 --> 01:06:38.340
What is that unknown amount? How many queries you have to make.

535
01:06:38.340 --> 01:06:43.349
Of these functions to find the unknown amount. Well, classically you need a lot. You have to evaluate many.

536
01:06:43.349 --> 01:06:46.469
Inputs to figure this out, but quantum, we can do it with just 1 query.

537
01:06:46.469 --> 01:06:54.030
You can see both the interesting thing and the useless thing. I mean, nobody's going to give you a quantum computer and tell you what's the hidden shift here?

538
01:06:54.030 --> 01:06:59.670
But it's interesting because it does give you in a sense something an opportunity could not be classic.

539
01:06:59.670 --> 01:07:09.690
And so this is, this is the initialization, the Gates, and then the measurement at the end and here we programmed all 1024, hidden shifts.

540
01:07:09.690 --> 01:07:18.269
2 to the 10th is 1024 on the vertical axis and this is, this is the measurement. If things were perfect, we already had a 1 on the diagonal here.

541
01:07:18.269 --> 01:07:21.510
But you can see there is some noise it's very structured noise.

542
01:07:21.510 --> 01:07:30.119
You look closely, you can even see little circles here and these lines. We understand these noise. They have to do with cross talk and imperfect laser noise and so forth.

543
01:07:30.119 --> 01:07:33.929
But we understand the entire hair budget in this algorithm and we're.

544
01:07:33.929 --> 01:07:37.139
You're working on building more, and I should say this comes from I and Q.

545
01:07:37.139 --> 01:07:41.099
Start out that has 3 full stack quantum computer systems right now.

546
01:07:41.099 --> 01:07:49.710
Now, back to the University of Maryland, we have several systems that we're also building out. This is a little bit of more of a physics experiment.

547
01:07:49.710 --> 01:07:53.670
But with lots and lots of candidates, and what we've studied there is.

548
01:07:53.670 --> 01:07:57.570
Not a quantum computing circuit, but what we might call it quantum simulation.

549
01:07:57.570 --> 01:08:01.139
And in this case, we're able to apply Hamilton in.

550
01:08:01.139 --> 01:08:09.059
That is an icing model between all pairs of Spence and it's a long range icing model and it's anti magnetic. So it's frustrated. There's lots of entanglement.

551
01:08:09.059 --> 01:08:13.230
The ground state of that of that, have a tongue in and we have a transverse field.

552
01:08:13.230 --> 01:08:17.609
So this is a very simple textbook model of something that it's a phase transition.

553
01:08:17.609 --> 01:08:21.390
In this case, we've been studying something called the dynamic or phase transition.

554
01:08:21.390 --> 01:08:26.939
And it has to do with the competition between B and J, if the spends are prepared along X.

555
01:08:26.939 --> 01:08:40.470
And the icing company is very big, they're already along the X direction. So nothing very exciting happens but if it's much bigger than Jay, then they process around the CX so that they stay around X order around extra today processor on Z.

556
01:08:40.470 --> 01:08:44.850
By varying B and Jay, we can sort of see different dynamics.

557
01:08:44.850 --> 01:08:55.470
Again, I'm sorry, I have to kind of rush through this, but this is the total the brightness of the chain. You can see when B is small, not much happens to the initial diarization.

558
01:08:55.470 --> 01:08:59.729
But when B is bigger than Jay, they sort of go to 0T. They process around 8.

559
01:08:59.729 --> 01:09:03.390
And I would call that a phase transition. It's a very smooth.

560
01:09:03.390 --> 01:09:07.109
It's a very smooth transition between these and I think this is with only.

561
01:09:07.109 --> 01:09:15.689
16 spans or something we can now look at the correlations. We measure every possible correlation to these Spence. This is something that you can't do in a real solid.

562
01:09:15.689 --> 01:09:21.149
And now, as we vary the number of spends every day, we seek sort of a dip emerging, but it's still.

563
01:09:21.149 --> 01:09:26.010
Even have to 53 spends where there are 2 to the 53 configurations.

564
01:09:26.010 --> 01:09:29.399
You know, we not see a real pronounced tip, but when we look at.

565
01:09:29.399 --> 01:09:34.529
The body correlation function in this case, we're looking at the size of the biggest domain.

566
01:09:34.529 --> 01:09:37.829
In the system, as we very be over day, we see a much more pronounced.

567
01:09:37.829 --> 01:09:42.449
Came here, and I will say that we couldn't compute where that kink was going to be.

568
01:09:42.449 --> 01:09:45.689
The system is too complicated, so.

569
01:09:45.689 --> 01:09:54.689
I'm not claiming this is a real model of a real material, or it's even interesting, but it's something you could not compute using classical physics. So, to me, that's kind of interesting.

570
01:09:55.890 --> 01:10:00.329
So, how are we going to scale and I would say in the atomic physics realm of quantum computing.

571
01:10:00.329 --> 01:10:04.319
It's often said, well, with with transistors, we can.

572
01:10:04.319 --> 01:10:08.789
I believe the amount of chat, even the Super doctors we can think of how to.

573
01:10:08.789 --> 01:10:17.130
Print thousands and thousands on a chip, but with Adams, you're showing me data with 10 Cubans or 50 cubic. How are you going to really scale? Well.

574
01:10:17.130 --> 01:10:23.130
The scaling in a sense, we have the fundamental ingredients because when we add an Atom, we know it's the same.

575
01:10:23.130 --> 01:10:26.850
It's all a question of controlling it from the trap itself.

576
01:10:26.850 --> 01:10:34.590
We have to make the track big enough to hold these atoms and 1 way to scale beyond 10 or 20 bits is to just add Adams.

577
01:10:34.590 --> 01:10:38.279
And you might say, well, now you need a 1M laser veins. Well, not really.

578
01:10:38.279 --> 01:10:43.470
You can imagine having a fixed set of laser beams and shoveling the atoms around the beans.

579
01:10:43.470 --> 01:10:53.520
Very much like a tape moves across the head. There's some overhead there you have to move the atoms around, but that's actually a very mature technology at shuttling atoms back and forth.

580
01:10:53.520 --> 01:10:59.579
I'd say Dave violin and his group in this folder. I was there in the nineties. We really pushed this.

581
01:10:59.579 --> 01:11:06.899
This vision of shuttling around corners and so forth and it missed. They're doing some wonderful experiments moving atoms between different. Sounds even.

582
01:11:06.899 --> 01:11:11.670
The technology that needs to approve there is the ability to make.

583
01:11:11.670 --> 01:11:17.609
Chip traps that are that are more agile. They allow us to do many more things that.

584
01:11:17.609 --> 01:11:21.000
None of this is quantum. Is this ion traffic? Just.

585
01:11:21.000 --> 01:11:24.869
It's a chunk of a bunch of conductors that hold voltages.

586
01:11:24.869 --> 01:11:31.890
We have to control those voltages now when we want to scale beyond hundreds of Cubics and this is where it gets really wild.

587
01:11:31.890 --> 01:11:35.939
We're going to almost certainly have to adopt a.

588
01:11:35.939 --> 01:11:40.439
Architecture just like we do with multi fascicle processors.

589
01:11:40.439 --> 01:11:46.560
And the great thing about atomic Cuban, they're optically active, we can map the cubic from the.

590
01:11:46.560 --> 01:11:51.569
Clock state, the spend to 2 properties of a propagating fulltime.

591
01:11:51.569 --> 01:11:55.890
Polarization or or or color, or maybe a delay.

592
01:11:55.890 --> 01:12:03.390
Main degrees of freedom there, and we can think about inventing a very high level architecture where individual modules of traps.

593
01:12:03.390 --> 01:12:06.539
With debts, connections inside these red boxes.

594
01:12:06.539 --> 01:12:12.960
Are augmented by by tonic connections and everything over here is classical.

595
01:12:12.960 --> 01:12:17.430
This is just a non blocking and my end cross connect switch.

596
01:12:17.430 --> 01:12:24.960
You can build this. This is a camera these are being splitters. And for instance, when when 2 pixels on this CCD fire.

597
01:12:24.960 --> 01:12:28.289
That means, depending on the configuration of the switch box.

598
01:12:28.289 --> 01:12:33.359
That cubital is entangled with that, so it's a module architecture that allows you to scale.

599
01:12:33.359 --> 01:12:42.479
Now, what you sacrifice here is kind of activity, just like, in the classical case, you can't easily do a gate between this transistor and that transition. You have to move things around.

600
01:12:42.479 --> 01:12:46.920
That's certainly true, but it allows you to build the system up and if it's.

601
01:12:46.920 --> 01:12:51.119
Then you can really thinking the industrial pride treat stamp out individual modules.

602
01:12:51.119 --> 01:12:54.840
So, there's lots of work that has to be done and I think we look forward to.

603
01:12:54.840 --> 01:13:00.630
Having lots of chip foot tonic wave guys that allow us to move this information back and forth.

604
01:13:00.630 --> 01:13:06.449
So, the real problem in the scale up and many people visit our labs and ammo and they'll see this.

605
01:13:06.449 --> 01:13:09.720
And I'll just say, okay, nice visiting you.

606
01:13:09.720 --> 01:13:16.649
So this table actually has, it has 2 iron traps, but our 1st experiments were involving.

607
01:13:16.649 --> 01:13:21.090
2 atoms, they were about a meter apart and we were the information.

608
01:13:21.090 --> 01:13:27.420
Between 1 Atom, and the other, we could actually say we telephone it and Adam, but it's of course, it's just the information.

609
01:13:27.420 --> 01:13:40.020
1, Cuban, and it requires a great deal of control. A lot of these are that have to be stabilized and locked to 9 or 10 digits. There is a pulse laser on the table on this table. It's way over here. You can't even see it.

610
01:13:40.020 --> 01:13:43.649
And all the electronics, you can recognize these top good controllers.

611
01:13:43.649 --> 01:13:49.500
We love topic they and other laser companies are starting to their lasers either.

612
01:13:49.500 --> 01:13:55.260
And on that note, and this is with very little exaggeration, everything you see here.

613
01:13:55.260 --> 01:14:00.060
Is in that box right now this is a new project that we started a few years ago.

614
01:14:00.060 --> 01:14:08.340
Largely funded by our, and they funded us to build a a small quantum computer system to do air to do an air correcting code. But.

615
01:14:08.340 --> 01:14:11.550
And afforded us the opportunity to really design top down.

616
01:14:11.550 --> 01:14:14.939
And so we've contracted with many outfits to help us.

617
01:14:14.939 --> 01:14:18.569
They make this happen, including coherent. It makes the lasers.

618
01:14:18.569 --> 01:14:27.720
Harris, I didn't talk much about this 32 channel. Post optic was a beautiful device that Harris makes the semiconductor industry, and we told them, you know.

619
01:14:27.720 --> 01:14:32.670
It's beautiful device, but you need to be making it for atoms. Not not for.

620
01:14:32.670 --> 01:14:38.760
Great. And so we're partnering with them on this. The top guy mentioned sense as well and making.

621
01:14:38.760 --> 01:14:44.340
Making individual lasers that can be locked all put in a drawer and fiber delivered.

622
01:14:44.340 --> 01:14:48.779
And finally, Sandia makes makes the the.

623
01:14:48.779 --> 01:14:52.079
The Silicon chip trap I showed before, so.

624
01:14:52.079 --> 01:14:56.850
This is starting to look like a product and that's of course the story behind Q.

625
01:14:56.850 --> 01:14:59.850
We wanted to add professional engineering.

626
01:14:59.850 --> 01:15:02.939
And not university engineering, which is a little bit of a.

627
01:15:02.939 --> 01:15:11.609
And I think to do engineering, so well, that a 3rd party could use a device. It's very hard to do that in a university setting.

628
01:15:11.609 --> 01:15:16.260
So, there's the box, you can sort of see, there's lots of cables, this kind of ugly.

629
01:15:16.260 --> 01:15:21.390
These are all fiber pump these are that are controlling the.

630
01:15:21.390 --> 01:15:26.189
The channels and the, the electrodes on the trap.

631
01:15:26.189 --> 01:15:30.869
And there's a very agile software controller that runs all of this stuff.

632
01:15:30.869 --> 01:15:33.989
The actions right here, this is where all the improvements going to happen.

633
01:15:33.989 --> 01:15:38.670
We look forward to to making that even smaller inside there.

634
01:15:38.670 --> 01:15:42.720
Is the chip I showed earlier the Sandia chip there are many ways to make this chip.

635
01:15:42.720 --> 01:15:53.670
I didn't mention that the atoms are about a 10th of a millimeter above the surface, and that's why it looks like a bow tie. We shine lasers across that surface. They focus tightly without hitting aside.

636
01:15:53.670 --> 01:16:03.449
So that's what's going on there. Now, this, this plot I borrowed from my friend heartbeat, Nevin at Google he leads to Google quantum effort.

637
01:16:03.449 --> 01:16:08.460
And it's sometimes called the quantum supremacy chart. And what we have here is the number of Cubics.

638
01:16:08.460 --> 01:16:12.779
And the number of operations circuit depth, how many can you do?

639
01:16:12.779 --> 01:16:18.840
It's pretty clear if you have some number of candidates, you probably to make use of that system you want more and more operations.

640
01:16:18.840 --> 01:16:26.579
You know, be up into the right and in fact, if you're not sufficiently up or not as efficient and not sufficiently to the right.

641
01:16:26.579 --> 01:16:31.380
You're in this pink zone where it's not so interesting. So we can classically simulate what goes on.

642
01:16:31.380 --> 01:16:37.649
If you only have, if you have 50 Cubics, you better be able to do several 100 operations on them.

643
01:16:37.649 --> 01:16:41.489
If you have a 1M canvas, but you can only do 1 or 2 operations. Not interesting.

644
01:16:41.489 --> 01:16:45.119
If you have 3 of this, you can do a 1B operations. Not interesting.

645
01:16:45.119 --> 01:16:48.539
You have to do both and with trap time technology.

646
01:16:48.539 --> 01:16:52.020
It's easy to speculate in the future, but.

647
01:16:52.020 --> 01:16:56.399
The black data and the gray data. This is existing research.

648
01:16:56.399 --> 01:17:01.560
And we look forward in the coming years, the pass through this so called supremacy boundary.

649
01:17:01.560 --> 01:17:04.680
Whether there is going to be something interesting there or not.

650
01:17:04.680 --> 01:17:11.430
Remains the same and when we, when we get sufficiently high and cubic number and circuit depth, we have to do air correction.

651
01:17:11.430 --> 01:17:14.640
And with trap high and technology, we don't worry about that yet because.

652
01:17:14.640 --> 01:17:18.899
Passive errors are so good. We don't need to worry about air correction.

653
01:17:18.899 --> 01:17:24.510
We really think we can get in this land of treasure with several 100 Cuba, several 10000 operations.

654
01:17:24.510 --> 01:17:28.859
With passive control and what's a little bit of a.

655
01:17:28.859 --> 01:17:33.630
Disappoint I think what's happening in solid state. Quantum computing is they're moving up into the left.

656
01:17:33.630 --> 01:17:37.949
Maybe not a surprise in a solid system when you make the system big.

657
01:17:37.949 --> 01:17:45.029
There are many more types of errors that can come in. There's crosstalk hairs. And the differences between cubex gets magnified.

658
01:17:45.029 --> 01:17:52.229
When you have a big system and the idle errands, when you leave a cell, say, keep it alone, it dies all by itself because it's tightly coupled to its environment.

659
01:17:52.229 --> 01:17:55.979
Solid the great thing in atomic physics is that.

660
01:17:55.979 --> 01:18:01.590
These Adams, the background is a real vacuum. Some people may say, well, we're back to back into well, you know.

661
01:18:01.590 --> 01:18:07.979
Biomechanics may demand that, because there's very little. I mean, the vacuum is perfect. There's nothing there.

662
01:18:07.979 --> 01:18:12.420
And furthermore the optical controllers of these atoms allow us to reconfigure.

663
01:18:12.420 --> 01:18:17.939
Program and they allow us great flexibility and applying new types of value because there's no wires.

664
01:18:17.939 --> 01:18:21.239
Between these Cubans, so.

665
01:18:21.239 --> 01:18:27.930
With that I wanted to back out and talk a little bit about the big challenge ahead in this field.

666
01:18:27.930 --> 01:18:38.100
Which is to do quantum engineering and this is since I work inside the way I get roped into working with the government on something called the national initiative and I.

667
01:18:38.100 --> 01:18:41.340
For this audience, especially the optical society.

668
01:18:41.340 --> 01:18:44.819
In the National photography initiative played a huge role.

669
01:18:44.819 --> 01:18:49.470
And and in Congress actually passing this national car initiative act.

670
01:18:49.470 --> 01:18:58.380
And I show a picture here, Mike grammar, and I was sort of leading stakeholders. We formed this group across the country to lobby and Laura Colton.

671
01:18:58.380 --> 01:19:01.649
We used to be the federal liaison and USA, and now, David Lang here.

672
01:19:01.649 --> 01:19:09.359
They've been tireless in, in getting us connected to offices in Congress and this mounted to the, and act.

673
01:19:09.359 --> 01:19:13.770
December broad bipartisan support the president signed it immediately. We have.

674
01:19:13.770 --> 01:19:16.979
There are people from our community in the White House now that are.

675
01:19:16.979 --> 01:19:21.359
Advising, and and and also the defense and intelligence agencies.

676
01:19:21.359 --> 01:19:24.720
How to coordinate? So.

677
01:19:24.720 --> 01:19:29.399
The, especially for this crowd, the, the, the goal here is to get up.

678
01:19:29.399 --> 01:19:35.460
Optics engineers to start thinking about quantum so, with that, I guess that's all I had to say I wanted to.

679
01:19:35.460 --> 01:19:38.789
But this joke up here, this is the cover of a tech review.

680
01:19:38.789 --> 01:19:43.739
From last year that you could say quantum computing somewhere in here, there's a lot of hype in the field.

681
01:19:43.739 --> 01:19:49.350
But there's a lot of promise and so very quickly. My group at the University of Maryland.

682
01:19:49.350 --> 01:19:55.470
Funded by many, many sources in Washington and the company I, in queue, which.

683
01:19:55.470 --> 01:19:59.189
I'm sort of stepping back from now they're about 35 employees.

684
01:19:59.189 --> 01:20:02.699
And if you're interested in January and wants to do.

685
01:20:02.699 --> 01:20:05.850
Quantum, please give us a call we're going to probably triple in size.

686
01:20:05.850 --> 01:20:12.149
Over the next couple of years, and we're hiring aggressively. Thank you.

687
01:20:16.920 --> 01:20:21.359
Silence.

688
01:20:21.359 --> 01:20:26.819
Silence.

689
01:20:26.819 --> 01:20:29.850
Okay.

690
01:20:34.020 --> 01:20:40.140
So so you see you next week, have a good Thanksgiving.

691
01:20:40.140 --> 01:20:45.090
And we'll have another guest lecture and then maybe some random things.

692
01:20:45.090 --> 01:20:48.630
If you have any questions, then.

693
01:20:49.739 --> 01:20:53.069
Passed away, other than that.

694
01:20:53.069 --> 01:20:56.609
See, you in a week bye.

695
01:21:06.390 --> 01:21:10.439
Thank you have a happy Thanksgiving. Everyone.

696
01:21:20.489 --> 01:21:25.260
Silence.

697
01:21:34.529 --> 01:21:38.640
Silence.