Movatterモバイル変換

Quantum ComputingLecture 1: Basic IntroductionMountain View CA, July 28, 2020Slides: http://slideshare.net/LaBlogga“The laws of physics present no barrier to reducing thesize of computers until bits are the size of atoms”— Richard P. Feynman (1985)Melanie Swan

28 July 2020Quantum ComputingTheoretical Model of Quantum Reality Quantum reality is information-theoretic and computable Lecture 1: Quantum Computing basics (hardware) Lecture 2: Advanced concepts (control software betweenmacroscale reality and quantum microstates) Lecture 3: Application (B/CI neuronanorobot network)1

28 July 2020Quantum Computing Agenda What is a Qubit? How are Qubits made? Qubit methods technical deep dive Quantum Programming Applications The Future: Quantum Photonics Conclusion2Quantum Computing1. Basic Introduction

28 July 2020Quantum ComputingFeynman: Universal Quantum Computer3Sources: Feynman, R.P. (1985). Quantum Mechanical Computers. Foundations of Physics. 16(6):507-31.Feynman, R.P. (1982). Simulating physics with computers. International Journal of Theor. Physics. 21(6):467-88. “The laws of physics present no barrier to reducingthe size of computers until bits are the size of atomsand quantum behavior holds sway” (1985) Vision: build a “universal quantum simulator” in thestructure of nature (1982) Simulate field theories with lattice works of spins

28 July 2020Quantum Computing 4(abstract)Computational infrastructure is more powerfulwhen it is in the same shape as the underlying3D structure of physical reality(concrete)Quantum Computing Tipping Points: universal quantum computing chips exotic superconducting materials deployment quantum optics: global quantum photonictelecommunications networksThesis

28 July 2020Quantum ComputingQuantum Scale5QCD: Quantum Chromodynamics “Quantum” = anything at the scale ofatomic and subatomic particles Theme: ability to manipulate physicalreality at increasingly smaller scalesSubatomic particlesMatter particles: fermions (quarks)Force particles: bosons (gluons)Scale Entities Physical Theory1 1 x101 m Humans Newtonian mechanics2 1 x10-9 m Atoms, ions,photonsQuantum mechanics(nanotechnology)3 1 x10-15 m Subatomic particles QCD/gauge theories4 1 x10-35 m Planck length Planck scaleAtoms Quantum objects:atoms, ions, photons

28 July 2020Quantum ComputingQuantum: many exponential speed-ups1. Bit (0 or 1)2. Qubit (0 and 1 in superposition)3. Qudit (more than 2 values in superposition) Microchip generates two entangled qudits each with 10states, for 100 dimensions total, for more than sixentangled qubits could generate (Imany, 2019 )4. Optics (time and frequency multiplexing) Existing telecommunications infrastructure Global network not standalone computers in labs5. Optics (superposition of inputs and gates)6ClassicalComputingQuantumComputingSource: Imany et al. (2019). High-dimensional optical quantum logic in large operational spaces. npj Quantum Information. 5(59):1-10.

28 July 2020Quantum Computing 7What is Quantum Computing?Quantum Computing is using quantum-mechanicalproperties (SEI: superposition, entanglement, andinterference) to perform computation with2n scaling (e.g. 9-qubit system tests 512 states (29)

28 July 2020Quantum ComputingQuantum smartphone ship date? Technology is notoriously difficult to predict I think there is a world market for maybe five computers- Thomas J. Watson, CEO, IBM, 19438Source: Strohmeyer, R. (2008). The 7 Worst Tech Predictions of All Time. PCWorld.D-Wave Systems10-feet tall, $15mCurrent: Ytterbium-171 isotopes at 1Kelvin (-458°F)Actual room-temperaturesuperconductor: ??

28 July 2020Quantum ComputingQuantum Computing impact Why is it important? Immanent as substantial new computing paradigm Immediate: upgrade to new global cryptography standards Ongoing: substantial step-ups in processing power When is it coming? Maybe within 10 years, early commercial systems shipping now Do all problems become solvable? No, one-tier improvement in problem solving complexity How can I try it? 1-minute per month free cloud access D-Wave Systems, IBM Program and test algorithms9

28 July 2020Quantum ComputingComputational Complexity and Quantum Computing10 Computational complexity: amount (time and space) ofcomputing resources required to solve a problem QC: one-tier improvement in computational complexity Canonical Traveling Salesperson Problem: check twice as manycities in half the time using a quantum computer Solve the next tier of designated problem difficulty with thecurrent tier’s computational resource (in time and space) NP becomes solvable in P, EXP becomes solvable in NP Example: factoring large numbers becomes time-reasonableP: polynomial time (e.g. solvable in human-reasonable amount of time); NP: non-polynomial (not solvable in human-reasonableamount of time); EXP: exponential (requires exponential time/space to solve)ComputationalComplexity

28 July 2020Quantum ComputingGoogle: Quantum Advantage (October 23, 2019) First quantum computer to solve a problemclassical computers cannot solve timely 53-qubit Sycamore chip (one damaged qubit) Task: random circuit sampling (provable randomness) Sampling versus one answer (i.e. Shor’s factoring, Grover’s search) Google: Sycamore repeats a random circuit sampling processa million times in 200 seconds (stores circuits in RAM) Claim: the most powerful classical computer (supercomputer)would take 10,000 years to do the same task IBM counterclaim: no, the calculation could be performed in2.5 days (write circuits to hard disk and then sample) Write circuits to 250 petabytes of hard disk (Summit Oak RidgeNational Lab supercomputer) and check with vector matrixmultiplication11Source: Arute et al. Quantum supremacy using a programmable superconducting processor. Nature. 574:505-11, andhttps://www.scottaaronson.com/blog/?p=4372

28 July 2020Quantum Computing A qubit (quantum bit) is the basic unit ofquantum information, the quantum versionof the classical binary bit13What is a Qubit?Bit always existsin a single binarystate (0 or 1)Qubit exists in a state of superposition, atevery location with some probability, untilcollapsed into a measurement (0 or 1)Classical Bit Quantum Bit (Qubit)Source: https://www.newsweek.com/quantum-computing-research-computer-flagship-eu-452167

28 July 2020Quantum ComputingQudit (quantum information digit) Qudits: quantum information digits that can exist inmore than two states A qubit exists in a superposition of 0 and 1 before beingcollapsed to a measurement at the end of the computation A qutrit exists in the 0, 1, and 2 states until collapsed formeasurement (triplet is useful for quantum error correction) 7 and 10 qudits tested 4 optical qudits achieved the processing power of 20 qubits Motivation: generalize known quantum computingtechniques to higher level systems14Sources: Qudits: Fernando Parisio; Michael Kues. “It from Bit” Wheeler, J.A. (1990). Information, Physics, Quantum: The Search forLinks. In Proc. 3rd Int. Symp. Foundations of Quantum Mechanics, Tokyo, 1989, pp.354-368.Qutrit stabilizercode on a torusIt from Bit -> It from Qubit -> It from QuditThe Wheeler Progression

28 July 2020Quantum Computing Any stable two-level quantum-mechanicalsystem might be used as a qubit If can obtain 0s and 1s usable in computation16How are Qubits made?

28 July 2020Quantum Computing1. Superconductors2. Photonics3. Trapped ions17Source: Economist, Architecture Race for Quantum Computers, 20 June 2015.Top 3 Qubit Generation Methods (2015)

28 July 2020Quantum ComputingTop 3 Qubit Generation Methods (2020)181. Superconductors Commercial systems (on-premises and cloud-based) IBM & Rigetti: controllable gate model superconductors(~19 qubits) for all computational problems D-Wave Systems: less-controllable quantum annealingmachines (2048 qubits) for optimization problems2. Photonics3. Trapped ionsShippingResearch

28 July 2020Quantum ComputingD-Wave SystemsQuantum Annealing Solve optimization problems as low energy landscape Setup: qubits exist across the landscape in superpositions of0/1 (quantum wave function) Like a fog blanketing the problem space Annealing cycle: runs and the fog layer condenses to onepoint as the global minimum of the landscape Qubit spins flip back and forth until settling into the lowest-energy state of the system Readout: lowest-energy state is optimal answer Spin glass analogy (flexible spins funnel to lowest energy) Holographic annealing Use AdS/CFT correspondence to map boundary-bulk energyoperators to readout solution in one fewer dimensions19Image Source: Qolynes et al (2014) Frustration in biomolecules

28 July 2020Quantum ComputingCommercial Status by Platform20Source: Synthesized from QCWareOrganization Qubit Method # Qubits Status1 IBM (Almaden CA) Superconducting (gate model) 19 (50) Available2 D-Wave Systems (Vancouver BC) Superconducting (quantum annealing) 2048 Available3 Rigetti Computing (Berkeley CA) Superconducting (gate model) 19 Available4 Google (Mountain View CA) Superconducting (gate model) 53 (72) Built, unreleased5 Intel/Delft (Netherlands) Superconducting 49 Built, unreleased6 Quantum Circuits (New Haven CT) Superconducting Unknown Research7 IonQ (College Park MD) Trapped Ions 23 Built, unreleased8 Alpine Quantum Tech (Innsbruck) Trapped Ions Unknown Research9 Microsoft (Santa Barbara CA) Majorana Fermions Unknown Research10 Nokia Bell Labs (Princeton NJ) FQH State Unknown Research11 Xanadu Photonics (Toronto ON) Photonics Unknown Research12 PsiQuantum (Palo Alto CA) Photonics Unknown Research Tipping point: universal quantum computing chips

28 July 2020Quantum ComputingPhysical Qubit Generation Method #1Superconducting Circuits22Source: http://news.mit.edu/2014/cheaper-superconducting-computer-chips-1017 Idea: extend semiconductor product line Use existing global fab infrastructure Produce superconducting chips Superconductors: materials with zeroelectrical resistance when cooled below acertain critical temperature More than half of the periodic table elements Electrons travel unimpeded (no energy dissipation) 20% of electricity is lost due to resistance At critical temperature, two electrons (usuallyrepelling) form a weak bond (a Cooper pair) that cantunnel through metal with no resistanceSuperconducting circuitSuperconducting chip

28 July 2020Quantum ComputingKey enabling technology: Materials advance“Room-temperature” Superconductors23 Implication: cool with liquid nitrogen not helium “Desktop” computing without bulky cryogenic equipment Initial superconducting materials (1986): copper oxides Bismuth strontium calcium and yttrium barium copper oxide New wider range of materials (2008) Metal-based compounds of iron, aluminum, copper, niobium Experimental high-pressure materials (2015) Hydrogen sulfide and lanthanum superhydrideSuperconducting Material Critical Temperature Discovery1 Ordinary superconducting materials Below 30 K -303 °C 19112 High-temperature superconducting materials 138 K -135 °C 19863 Room-temperature superconducting materials 203 K -70 °C 20154 High room-temperature superconductingmaterials260 K -13 °C 2019

28 July 2020Quantum ComputingSuperconducting Circuits24 Josephson junction: nonlinear superconductinginductors create qubit energy levels The nonlinearity of the Josephson inductance breaksthe degeneracy of the energy level spacings, allowingsystem to be restricted to only the 2-qubit states Josephson junctions needed to produce qubits,otherwise superconducting loop is just a circuit Linear inductors in a traditional circuit are replaced withthe Josephson junction, a nonlinear element thatproduces energy levels with different spacings fromeach other that can be used as a qubit Superconducting loop is a SQUID (superconductingquantum interference device) magnetometer (adevice for measuring magnetic fields)Josephson: NobelPrize in Physics(1973) for workpredicting thetunneling behaviorof superconductingCooper pairs

28 July 2020Quantum ComputingSuperconducting Circuits: Rigetti25 Single Josephson junction qubiton a sapphire substrate Electrical circuit with oscillatingcurrent forms the qubits and is andcontrolled by electromagnetic fields Substrate embedded in a copperwaveguide cavity Waveguide coupled to qubittransitions to perform computation Chip: Alternating fixed andtunable transmon qubits 19Q (one qubit not tunable)Source: Otterbach, J.S., et al. (2017). Unsupervised machine learning on a hybrid quantum computer. arXiv: 1712.05771v1

28 July 2020Quantum ComputingSuperconducting Circuits: Google26 Qubits are electrical oscillators constructedfrom aluminum (niobium is also used) Superconducting at 1 K (−272°C) The oscillator qubits store small amounts ofelectrical energy Oscillator in the 0 state has zero energy Oscillator in the 1 state has a single quantum of energy Oscillator resonance frequency 6 gigahertz (300 millikelvin) Sets the energy differential between the 0 and 1 states Low enough frequency to build with off-the-shelfcomponents High enough frequency so ambient thermal energy doesnot scramble the oscillation and introduce errorsSuperconductingmicrowave circuit

28 July 2020Quantum ComputingPhysical Qubit Generation Method #2Quantum Photonics27Image Source: PSI QuantumPhoton movement Quantum-mechanical objects Atom, ion, photon Optical circuits do not require error correction Global communications networks built onphotonic transfer Quantum photonics (general) Single photons represent qubits Realized in computing chips or in free space Compute with entangled states of multiplephotons (photonic clusters) Single photons are sent through the chip or freespace for the computation and then measuredwith photon detectors at the other endQuantum photonic processorQuantum photonic waferQuantum photonic array

28 July 2020Quantum ComputingContinuous Qubit Optical Interfaces28Source: Vuckovic, J. (2020). Stanford Optimized Quantum Photonics; Sapra, N.V., Yang, K.Y., Vercruysse, D., et al. (2020). On-chipintegrated laser-driven particle accelerator. Science. 367(6473):79-83 All-optical platform from the beginning Homogeneous qubits with optical interfaces Method: exploit color center defects (Fabre effect) Color centers in diamond (silicon and tin vacancy) Color centers in silicon carbide (manufacture siliconvacancy in 4H poly tech type (thin film)) Exploit energy level differentials due to missingatoms in the lattice structure The wavelength between two color centers depends onwhich atom in the lattice is missing and can be used forcomputation

28 July 2020Quantum ComputingQuantum Photonics29 Diamond center defects method Introduce impurities to diamond crystal lattice Implant ion to create nitrogen vacancy Nitrogen vacancy produces the Farbe center(color center), a defect in a crystal latticeoccupied by an unpaired electron The unpaired electron creates an effectivespin which can be manipulated as a qubit Quantum state can be initialized, manipulated,and measured at room temperature Uses the same physics and math as forJosephson junctions in microwave chips But, coherence time limited to spin timeRelated work:Accelerator-on-a-chip(Stanford Nanoscale andQuantum Photonics Lab)Source: Vuckovic, J. (2020). Stanford Optimized Quantum Photonics; Sapra, N.V., Yang, K.Y., Vercruysse, D., et al. (2020). On-chipintegrated laser-driven particle accelerator. Science. 367(6473):79-83

28 July 2020Quantum ComputingPhysical Qubit Generation Method #3Trapped Ions30Source: Images: IonQ, College Park MD Silicon chips store Ytterbium ions inelectromagnetic traps Manipulate in computation with lasers andelectromagnetic fields Ions (atoms stripped of electrons) Easier to compute with positively ornegatively charged ions Ytterbium ions do not need supercooling,have a long coherence time, and requireless error correction

28 July 2020Quantum ComputingTrapped Ions31Source: IonQ, College Park MD1. Silicon chip with 100 electrodes confinesand controls ions in an ultrahigh-vacuum Electrodes underneath the ions apply electricalpotentials to hold the charged particles togetherin a linear array2. Lasers initialize the qubits, entangle themthrough coupling, and produce quantumlogic gates to execute the computation3. At the end of the computation, anotherlaser causes ions to fluoresce if they arein a certain qubit state Fluorescence collected to measure each qubitand compute the result of the computationIons trapped in arrayTrapped-ion quantumprocessor

28 July 2020Quantum ComputingPhysical Qubit Generation Method #4Topological Qubits: Majorana Fermions32 Topological qubits Qubits made from particles on topologicalsuperconductors and electrically controlled incomputation based on movement trajectories Majorana fermions (particle + anti-particle pairs) Novel quantum phases arising in condensedmatter with Cooper pairing states (i.e. quantumcomputable states) on superconductor edges Majorana fermions move in trajectoriesresembling a multi-stranded braid Use braid wave functions as quantum logic gates

28 July 2020Quantum ComputingDiVincenzo Criteria for Universal Computing34 Quantum computing standards for gate array computing 1: demonstrate a reliable system for making qubits 2-5: perform accurate computation Qubit formation (criterion #1)1. A scalable system of well-characterized qubits Qubit control for computation (criteria #2-5)2. Qubits that can be initialized with fidelity (to the zero state)3. Qubits with long-enough coherence time for calculation4. A universal set of quantum gates5. Capability to measure any specific qubit in the ending resultSource: DiVincenzo, D.P. (2000). The physical implementation of quantum computation. Fortschrit. Phys. 48(9–11):771–83.

28 July 2020Quantum ComputingHardware for Qubit Generation and Control35Source: Synthesized from QCWareQubit Type Qubit formation(DiVincenzo criterion #1)Qubit control for computation(DiVincenzo criteria #2-5)1 SuperconductingcircuitsElectrical circuit with oscillating current Electromagnetic fields and microwavepulses2 Photonic circuits Single photons (or squeezed states) insilicon waveguidesMarshalled cluster state of multi-dimensional entangled qubits3 Diamond centerdefectsDefect has an effective spin; the two-levels of the spin define a qubitMicrowave fields and lasers4 Trapped ions Ion (atom stripped of one electron) Ions stored in electromagnetic trapsand manipulated with lasers5 Majorana fermions Topological superconductors Electrically-controlled along non-abelian “braiding” path6 Neutral atoms Electronic states of atoms trapped bylaser-formed optical latticeControlled by lasers7 Quantum dots Electron spins in a semiconductornanostructureMicrowave pulses Race to build first universal gate quantum computer Easy to generate qubits, difficult to compute with fidelity

28 July 2020Quantum ComputingQuantum Programming Standard gates Hadamard gate: acts on one qubit toput it in a superposition CNOT gate: acts on two qubits to flip one Toffoli gate: acts on three or more qubits to implement the sixBoolean operators (AND, conditional AND, OR, conditional OR,exclusive OR, and NOT) Computing paradigms Classical computing relies on electrical conductivity Boolean algebra (true/false, and/or) to manipulate bits Quantum computing relies on quantum mechanics Linear algebra to manipulate matrices of complex numbers (i.e. theamplitudes of possible states)36

28 July 2020Quantum ComputingStandardized Tools37 Bernstein-Vazirani algorithm (1997) “Hello, World!” of quantum: extract specific bits from a string Variational quantum eigensolver (VQE) (Peruzzo,2014) Find the eigenvalues of a matrix; An eigensolver is a programdesigned to calculate solutions to 3D problems Quantum approximate optimization algorithm (QAOA)(Farhi, 2014) Combinatorial optimization problems (Traveling SalesmanProblem, find a “good” solution (acceptable answer) inpolynomial time (a reasonable amount of time); max-cutpartition function, solve as energy landscape minimization

28 July 2020Quantum ComputingGoal: Standard Gate Array Computing38 2n scaling: 9-qubit system (29) represents 512 statesSource: D-Wave Systems, A Machine of a Different Kind, Quantum Computing, 2019

28 July 2020Quantum ComputingQuantum Computing Roadmap39 Long-term: Universal quantum computing Universal computation devices using fault-tolerantquantum information processors Error correction required (system noise overwhelmscoherent wave activity of qubit particles) Available now: NISQ devices (noisy intermediate-scale quantum) Error correction not required Applications in optimization, simulation, machinelearning, and cryptographySource: Preskill, J. (2018). Quantum Computing in the NISQ era and beyond. Quantum. 2(79):1-20.

28 July 2020Quantum Computing 40Sources: Preskill, J. (2018). Quantum computing in the NISQ era and beyond. Quantum 2(79):1–20.https://amitray.com/roadmap-for-1000-qubits-fault-tolerant-quantum-computers/Quantum Computing Roadmap Long-term applications Shor’s factoring algorithm (could break currentcryptography standard (RSA)) Grover’s search algorithm (faster search through largedata sets)

28 July 2020Quantum ComputingQuantum Computing Applications Cryptography and security Certifiable randomness (entropy) and quantum statistics Quantum machine learning Optimization, nitrogen sequestration Simulation Exotic materials, molecular dynamics, drug discovery Emulation and pathology resolution Quantum computing for the brain42

28 July 2020Quantum ComputingQuantum Cryptography Quantum computing implicated in eventually being ableto break existing cryptographic standards (2048-bit RSA) 2019 US National Academies of Sciences report “unlikely within 10 years” however methods improving Solution: US NIST developing next-generation standards Lattice cryptography (complex 3D arrangements of atoms) Instead of the difficulty of factoring large numbers (RSA 2048) orany other number theory-based methods (e.g. discrete log) Overall mathematical shift to group theory (lattices) from numbertheory (factoring)43Source: Grumbling, E. & Horowitz, M. (2019). Quantum Computing: Progress and Prospects. Washington, DC: US National Academies ofSciences, p 157.

28 July 2020Quantum Computing 44NIST Next-generation Cryptography NIST: 26 of 69 algorithms advance topost-quantum crypto semifinal (Jan 2019) Public-key encryption (17) Digital signature schemes (9) Approaches: lattice-based,code-based, multivariate Lattice-based: target the Learning withErrors (LWE) problem with module or ringformulation (MLWE or RLWE) Code-based: error-correcting codes (LowDensity Parity Check (LDPC) codes) Multivariate: field equations (hidden fieldsand small fields) and algebraic equationsSource: NISTIR 8240: Status Report on the First Round of the NIST Post-Quantum Cryptography Standardization Process, January2019, https://doi.org/10.6028/NIST.IR.8240.

28 July 2020Quantum ComputingNature’s Quantum Security Features45Source: Swan et al. (2020). Quantum Computing. London: World Scientific. Reality is information-theoretic Computational complexity class of quantuminformation (BQP/QSZK) has access to Security features naturally built into quantummechanical domainsPrinciple Security Feature1 No-cloning theorem Cannot copy quantum information2 No-measurement principle Cannot measure quantum information without damaging it(eavesdropping is immediately detectable)3 Quantum statistics Provable randomness: distributions could only have been quantum-generated (implications for quantum cryptography)4 Quantum error correction Error correction via ancilla (larger state of entangled qubits)5 BQP/QSZK computationalcomplexityQuantum information domains compute quickly enough to performtheir own computational verification (zero-knowledge proofs)

28 July 2020Quantum ComputingKiller App: Quantum Machine Learning46 Machine Learning and Quantum Computing Statistical methods with probabilistic outputSigmoidal Function 3D Hilbert Space0101Machine Learning Quantum ComputingSource: Image: machine learning object recognition: Anandkumar (2014). Tensor models for machine learning.

28 July 2020Quantum ComputingQuantum Optical Neural Networks47 Classical neural network architecture Hidden layers are rectified linear units (ReLUs) and the output neuronuses a sigmoid activation function to map the output into the range (0, 1) Quantum optical neural network architecture Inputs are single photon Fock states. The single-site nonlinearities aregiven a Kerr-type interaction applying a phase quadratic in the number ofphotons. Readout is given by photon-number-resolving detectors, whichmeasure the photon number at each output modeSource. Steinbrecher, G.R. et al., (2019). Quantum Optical Neural Networks. npj Quantum Information. 5(60):1-9.Quantum Optical Neural NetworkClassical Neural Network

28 July 2020Quantum ComputingQuantum Optimization Use Cases (D-Wave)48Sources: Hetzner, C. (2019). VW, Canadian tech company D-Wave team on quantum computing. Automotive News Canada; Fassler, J.(2018). Hey Amazon, Kroger’s new delivery partner operates almost entirely on robots. New Food Economy. UK online grocer Ocado’s automated warehouses: 1100 robots, 250,00 items, 3m instructions Optimization algorithm coordinates hundreds of robotspassing within 5 millimeters of each other at speeds of 4meters per second, fulfilling 65,000 orders per week Volkswagen: 418-taxi network in Beijing Optimize travel time, implement resulting trafficmanagement system in Lisbon and beyond AdTech for web browser promotion placement

28 July 2020Quantum ComputingSimulating Chemistry Molecular dynamics Can simulate millions of atoms at present Need quantum computing to capture quantum-mechanical interactions between electrons49Sources: Univ. of Illinois at Chicago/Argonne National Lab/Univ. of Southern California; Kandala et al. (2017). Nature. 549, 2427-qubit superconductingcircuit (false color) tosimulate a berylliumhydride molecule (IBM)

28 July 2020Quantum ComputingChips: CPU -> GPU -> TPU -> QPU GPU (graphics processing unit) 3D graphics cards for fast matrix multiplication TPU (tensor processing unit) (Edge TPU 2018) Flow through matrix multiplications without having tostore interim values in memory QPU (quantum processing unit) (Sycamore 2020) Solve problems quadratically or polynomially fasterexploiting SEI Properties (superposition, entanglement,interference)50TPU processing cluster andSycamore quantumsuperconducting chipTipping point:universal quantumcomputing chips

28 July 2020Quantum ComputingQuantum Optical Networks52 Quantum photonics could be at the center offuture global communications networks just asoptical networking is today Many ways to make qubits for computing onstandalone machines For a larger architecture of networkedmachines, electrical signals must beconverted to optical signals Photonics: core global telecoms networktechnology Future: “Cisco for optical routers” Quantum photonics: next-gen telecomsImage Source: Walther, P. (2018). Photonic Quantum Computing. Vienna Center for Quantum Science and Technology.Integrated quantumoptical switch

28 July 2020Quantum ComputingScalable Global Quantum Networks53 Quantum internet Security, privacy, scalability Quantum key distribution, quantum routers, quantum repeaters,quantum simulators, quantum components, quantum memory1. Homogeneous scalable qubits (standalone machines)2. Efficient optical interfaces (networked machines) Quantum transducers convert microwave to optical Optical interfaces (optical superconducting) Microwave interfaces (Josephson junction superconducting)Source: Vuckovic, J. (2020). Stanford Optimized Quantum Photonics; Sapra, N.V., Yang, K.Y., Vercruysse, D., et al. (2020). On-chipintegrated laser-driven particle accelerator. Science. 367(6473):79-83

28 July 2020Quantum ComputingScalable Global Quantum Networks54 Two methods currently in development Microwave superconducting platform interfaced to opticalnetworks with electrical-optical interconnects Optical platform with continuous qubit optical interfaces Photonic Integrated Circuits (PICs)Source: Vuckovic, J. (2020). Stanford Optimized Quantum Photonics; Sapra, N.V., Yang, K.Y., Vercruysse, D., et al. (2020). On-chipintegrated laser-driven particle accelerator. Science. 367(6473):79-83Quantum PhotonicProcessorsAll-optical PlatformSuperconductingProcessorsOptical-ElectricalInterconnectsDriver: quantum computing Drivers: 5G, data center, 100GbEAll OpticalOptical-ElectricalChipsGlobalCommsNetworks

28 July 2020Quantum ComputingQuantum Photonic Spacetime Multiplexing55 Standard quantum computing speed-up Space accelerated by testing states of 3D space Superposition of inputs Optical quantum computing speed-up Time accelerated by testing permutations of gate order Superposition of gate order and inputsSources: Procopio et al. 2015. Experimental Superposition of Orders of Quantum Gates. Nature Communications. 6(7913):1-6;Walther, P. (2018). Photonic Quantum Computing. Vienna Center for Quantum Science and Technology.Quantum Photonic Gate SuperpositionParallel to time-spacemanipulation in globalfiberopticcommunicationsTDM/WDM: time-division wave-divisionmultiplexing

28 July 2020Quantum ComputingRisks and Limitations Implementation stalls Qubits are more sensitive to environmental noise than bits Error correction stalls Unable to move past contemporary 50-70 qubit machines tomillion-qubit machines Materials discovery stalls Cannot find actual room-temperature superconductors Limitations of underlying physical theories Quantum mechanics Need beyond-probability methods that emphasize spectra,entanglement, entropy (irreversibility), and field flux Technology cycle is too early57

28 July 2020Quantum Computing 58Conclusion High-dimensionality is a central theme inscience and technology development Not just 3D but higher-dimensionality Nature’s built-in quantum security features No cloning, no measurement, zero-knowledgeproofs, quantum statistics & error correction Killer app: quantum machine learning Statistical methods with probabilistic output Apps in general: Cryptography, superconducting materialssimulation, quantum computing for the brain The future: quantum optical networks

28 July 2020Quantum Computing 59(abstract)Computational infrastructure is more powerfulwhen it is in the same shape as the underlying3D structure of physical reality(concrete)Quantum Computing Tipping Points: universal quantum computing chips exotic superconducting materials deployment quantum optics: global quantum photonictelecommunications networksThesis

Quantum ComputingLecture 1: Basic IntroductionMountain View CA, July 28, 2020Slides: http://slideshare.net/LaBloggaThank you!Questions?Melanie Swan

Movatterモバイル変換

Change Language

Quantum Computing Lecture 1: Basic Concepts

Embed presentation

Recommended

More Related Content

What's hot

Similar to Quantum Computing Lecture 1: Basic Concepts

More from Melanie Swan

Recently uploaded

In this document

Quantum Computing Lecture 1: Basic Concepts