SlideShare a Scribd company logo

Quantum Computing with Amazon Braket

•
•
Chris Fregly
Chris Fregly

Quantum Computing with Amazon Braket In this talk, I describe some fundamental principles of quantum computing including qu-bits, superposition, and entanglement. I will demonstrate how to perform secure quantum computing tasks across many Quantum Processing Units (QPUs) using Amazon Braket, IAM, and S3. AI and Machine Learning, Quantum Computing, Amazon Braket, QPU

Related slideshows

Taking Quantum Computing for a Spin: What is Imaginary and What is Real?
Taking Quantum Computing for a Spin: What is Imaginary and What is Real?Taking Quantum Computing for a Spin: What is Imaginary and What is Real?
Taking Quantum Computing for a Spin: What is Imaginary and What is Real?
 
Quantum computing.ppt
Quantum computing.pptQuantum computing.ppt
Quantum computing.ppt
 
Quantum computing meghaditya
Quantum computing meghadityaQuantum computing meghaditya
Quantum computing meghaditya
 
1 of 35
Download to read offline
Chris Fregly Developer Advocate AI and Machine Learning @AWS Quantum Computing with Amazon Braket
Me: Chris Fregly Developer Advocate AI and Machine Learning @ AWS (Based in San Francisco) Co-Author of the O'Reilly Book, "Data Science on AWS.” https://www.datascienceonaws.com Founder of the Advanced Kubeflow Meetup (Global) https://meetup.com/Advanced-Kubeflow github.com/data-science-on-aws @cfregly linkedin.com/in/cfregly
Data Science on AWS – Book and Workshop Outline https://www.datascienceonaws.com/
Agenda • What is Quantum Computing? • Use Cases for Quantum Computing • QPUs, Qubits, Superposition, Entanglement, Teleportation • Programming QPUs • DEMO: Quantum Computing with Amazon Braket
What is Quantum Computing? • “Quantum computers are no longer theoretical devices.” -- Programming Quantum Computers by Johnson, Harrigan, Gimeno-Segovia • Based on Quantum Mechanics (1930’s) • Quantum Mechanics is “Nature’s Operating System” • Einstein Called Quantum Mechanics ”Spooky” “No reasonable definition of reality could be expected to permit quantum mechanics.”
Why Quantum Now? • Enormous Long-term Potential • Learn & Build Expertise Now • Develop New Algos and IP • Start the Quantum Flywheel! • Quantum Today = Deep Learning 10-15 Years Ago Growth Users Hardware capabilities Use cases and algorithms Hardware developers Betteruser experience Better understanding of QC use cases
Quantum Computing Eraserrorrate(quality) number qubits (quantity) Eras: 1. Classically Simulatable 2. Noisy Intermediate-Scale Quantum (NISQ) ç Today 3. Error-Corrected Quantum Computing ç 5-10 Years? >50, <100k Qubits: Today’s Moderate-Useful Apps Quantum Supremacy (we just crossed over) “Supremacy”: We can no longer track what a quantum computer is doing. (Huge milestone achieved in 2019.) more qubits betterqubits Quantum Advantage (not yet crossed, find useful apps)
Amazon Braket puts quantum computing in the hands of every developer and scientist Managed development environments High-performing circuit simulators Hybrid workloads on secure, on-demand quantum hardware
Amazon Braket provides secure, on-demand access to different quantum computing technologies
Use Cases for Quantum Computing • Cryptography • Factor 2048 digit RSA key: quadrillion years (digital), 3 mins (quantum) • Classic Machine Learning Algos are Limited on Quantum • Linear Regression, PCA, SVM, Fast Fourier Transform, System of Eqns • Exponential Speed-ups for Certain Classes of Algorithms • ie. From O(n) to O(log n) • Financial, Chemical, Material Science Simulations
Quantum Software Industry https://www.creativedestructionlab.com/streams/quantum
Customer: Fidelity • Financial and Market Simulations • Creating Synthetic Securities • Options Pricing https://www.fcatalyst.com/projects/aug2020/amazon_braket https://www.fcatalyst.com/projects/aug2020/amazon_braket
Customer: Volkswagen Group • Automotive
Customer: Amgen • Biotech and Drug Discovery
Customer: Qu & Co • Chemical Simulations and Material Science
QPUs • Quantum Processing Unit (QPU) • Hybrid Quantum, Co-Processors Similar to GPUs • Short Bursts of QPU Computations, Coordinated by CPU • QPUs Defined by Number of Qubits (Quantum bits) • Today’s Quantum Computers have <100 Qubits • Today, Some Qubits Needed for Redundancy & Error Correction
Qubits • Qubit: Quantum Bits, Qunibble: 4 Qubits, Qubyte: 8 Qubits • Qubits are Ultimately 0’s and 1’s Just Like Digital Bits (?!) • There is only 1 photon per “shot” • Value of Qubit is Probabilistic Until READ (73% 0, 27% 1) • READ Forces Qubit Value to 0 or 1; Destroys Quantumness • Probabilistic State is called Superposition 1 Shot
Superposition • ie. Rolling Dice are in Superposition • Value of Dice is Probabilistic Until READ • Humans READ Dice When Dice Stop Rolling • Superposition Represented in Circle Notation • Probability of Reading a 1 (vs. 0) from Superposition No Superposition Superposition Normalized to 50% 0, 50% 1
Circle Notation • 2^n “Circles”; n=# of Qubits • 32 Qubits: 4,294,967,296 Circles • 300 Qubits: More Circles Than Atoms in the Universe!
Qubits Needed for Use Cases
Superposition Amplitude • Measured in Magnitude (Radius) + Relative Phase (Angle) • Phase Manipulation is Critical in Quantum Programming • Put Convention Data into Superposition with HAD Gate Quantum Random Number Generator
Entanglement • “Bell Pair”, John Bell, Irish Physicist • READ One Qubit, Another Qubit Reveals the Same Value • ie. 2 Dice Always Show Same Value…Across the Universe • Einstein Called This “Spooky Action at a Distance” • Extends to Multiple Qubit Groups (ie. Qu-nibbles, Qu-bytes)
Teleportation • “Quantum Ethernet Cable” • Requires 2 (or More) Entangled Qubits in a Bell Pair • Transport State of Qubit A to Qubit B; Destroys Qubit A • Currently, Only Transports Short Distances within a QPU • Today, Longest Teleporation is 1,400km (870mi) • Note: Not Transporting Matter – Only Information
Programming QPUs: Quantum Computing Simulators • Limited Qubit States Due to Digital RAM Constraints • 30 Qubits == 16GB, 40 Qubits = 16TB, 50 Qubits = 16PB
Build: Managed Jupyter Lab Environments Fully managed infrastructure Pre-built environments Learning resources Build and test algorithms
Test: Hosted Circuit Simulators Serverless execution Optimized performance Individual circuits and hybrid jobs Schrödinger simulator Tensor network simulator …
Run: Managed Execution of Hybrid Quantum Algorithms Users Local IDE Fully managed execution Pay only for what you use Change backend with a single line of code Amazon Braket Classical compute Simulator Amazon S3 Amazon CloudWatch QPU
Programming QPUs – Random Number Generator circuit = Circuit() a = circuit.h(0) # Put data into Qubit Superposition b = a.cnot(0, 1) # Entangle Qubits task = device.run(b, s3_prefix, shots=100000) task.result().measurement_counts Counter({'11': 50084, '00': 49916})
Programming QPUs - Teleportation • Step 1: WRITE Conventional Data from CPU into QPU • Step 2: Put Data into Superposition using HAD Operation • Step 3: Entangle Qubits using CNOT Operation • Step 4: Perform PHASE Ops on Qubits in Superposition • Step 5: READ Qubits from QPU back into CPU
© 2020, Amazon Web Services, Inc. or its affiliates. All rights reserved. DEMO: Amazon Braket
Looking Ahead • Quantum Research Improves Conventional Algos • ie. Recommendation Systems • Error-Correcting Qubits • Enables Deeper Circuits Across More Qubits • Quantum Now = Deep Learning 10-15 Years Ago • Limited by Hardware Availability … Until Now!
AWS Center for Quantum Computing (Opens 2021) Near-term applications Error correction & programming models Quantum hardware and technologies
Amazon Quantum Solutions Lab (Available Now) Quantum Computing Machine learning HPC Build and optimize deep learning approaches on AWS Build and optimize classical and quantum-inspired approaches on AWS Target use case Build quantum computing prototypes Develop new algorithms and approaches Benchmarking Results Collaboration Education Solutions
Technology and consulting partners provide software and service solutions on Amazon Braket
Thank you! © 2020, Amazon Web Services, Inc. or its affiliates. All rights reserved. Chris Fregly github.com/data-science-on-aws @cfregly linkedin.com/in/cfregly https://www.datascienceonaws.com https://aws.amazon.com/quantum-solutions-lab/

More Related Content

Quantum Computing with Amazon Braket

  • 1. Chris Fregly Developer Advocate AI and Machine Learning @AWS Quantum Computing with Amazon Braket
  • 2. Me: Chris Fregly Developer Advocate AI and Machine Learning @ AWS (Based in San Francisco) Co-Author of the O'Reilly Book, "Data Science on AWS.” https://www.datascienceonaws.com Founder of the Advanced Kubeflow Meetup (Global) https://meetup.com/Advanced-Kubeflow github.com/data-science-on-aws @cfregly linkedin.com/in/cfregly
  • 3. Data Science on AWS – Book and Workshop Outline https://www.datascienceonaws.com/
  • 4. Agenda • What is Quantum Computing? • Use Cases for Quantum Computing • QPUs, Qubits, Superposition, Entanglement, Teleportation • Programming QPUs • DEMO: Quantum Computing with Amazon Braket
  • 5. What is Quantum Computing? • “Quantum computers are no longer theoretical devices.” -- Programming Quantum Computers by Johnson, Harrigan, Gimeno-Segovia • Based on Quantum Mechanics (1930’s) • Quantum Mechanics is “Nature’s Operating System” • Einstein Called Quantum Mechanics ”Spooky” “No reasonable definition of reality could be expected to permit quantum mechanics.”
  • 6. Why Quantum Now? • Enormous Long-term Potential • Learn & Build Expertise Now • Develop New Algos and IP • Start the Quantum Flywheel! • Quantum Today = Deep Learning 10-15 Years Ago Growth Users Hardware capabilities Use cases and algorithms Hardware developers Betteruser experience Better understanding of QC use cases
  • 7. Quantum Computing Eraserrorrate(quality) number qubits (quantity) Eras: 1. Classically Simulatable 2. Noisy Intermediate-Scale Quantum (NISQ) ç Today 3. Error-Corrected Quantum Computing ç 5-10 Years? >50, <100k Qubits: Today’s Moderate-Useful Apps Quantum Supremacy (we just crossed over) “Supremacy”: We can no longer track what a quantum computer is doing. (Huge milestone achieved in 2019.) more qubits betterqubits Quantum Advantage (not yet crossed, find useful apps)
  • 8. Amazon Braket puts quantum computing in the hands of every developer and scientist Managed development environments High-performing circuit simulators Hybrid workloads on secure, on-demand quantum hardware
  • 9. Amazon Braket provides secure, on-demand access to different quantum computing technologies
  • 10. Use Cases for Quantum Computing • Cryptography • Factor 2048 digit RSA key: quadrillion years (digital), 3 mins (quantum) • Classic Machine Learning Algos are Limited on Quantum • Linear Regression, PCA, SVM, Fast Fourier Transform, System of Eqns • Exponential Speed-ups for Certain Classes of Algorithms • ie. From O(n) to O(log n) • Financial, Chemical, Material Science Simulations
  • 12. Customer: Fidelity • Financial and Market Simulations • Creating Synthetic Securities • Options Pricing https://www.fcatalyst.com/projects/aug2020/amazon_braket https://www.fcatalyst.com/projects/aug2020/amazon_braket
  • 14. Customer: Amgen • Biotech and Drug Discovery
  • 15. Customer: Qu & Co • Chemical Simulations and Material Science
  • 16. QPUs • Quantum Processing Unit (QPU) • Hybrid Quantum, Co-Processors Similar to GPUs • Short Bursts of QPU Computations, Coordinated by CPU • QPUs Defined by Number of Qubits (Quantum bits) • Today’s Quantum Computers have <100 Qubits • Today, Some Qubits Needed for Redundancy & Error Correction
  • 17. Qubits • Qubit: Quantum Bits, Qunibble: 4 Qubits, Qubyte: 8 Qubits • Qubits are Ultimately 0’s and 1’s Just Like Digital Bits (?!) • There is only 1 photon per “shot” • Value of Qubit is Probabilistic Until READ (73% 0, 27% 1) • READ Forces Qubit Value to 0 or 1; Destroys Quantumness • Probabilistic State is called Superposition 1 Shot
  • 18. Superposition • ie. Rolling Dice are in Superposition • Value of Dice is Probabilistic Until READ • Humans READ Dice When Dice Stop Rolling • Superposition Represented in Circle Notation • Probability of Reading a 1 (vs. 0) from Superposition No Superposition Superposition Normalized to 50% 0, 50% 1
  • 19. Circle Notation • 2^n “Circles”; n=# of Qubits • 32 Qubits: 4,294,967,296 Circles • 300 Qubits: More Circles Than Atoms in the Universe!
  • 20. Qubits Needed for Use Cases
  • 21. Superposition Amplitude • Measured in Magnitude (Radius) + Relative Phase (Angle) • Phase Manipulation is Critical in Quantum Programming • Put Convention Data into Superposition with HAD Gate Quantum Random Number Generator
  • 22. Entanglement • “Bell Pair”, John Bell, Irish Physicist • READ One Qubit, Another Qubit Reveals the Same Value • ie. 2 Dice Always Show Same Value…Across the Universe • Einstein Called This “Spooky Action at a Distance” • Extends to Multiple Qubit Groups (ie. Qu-nibbles, Qu-bytes)
  • 23. Teleportation • “Quantum Ethernet Cable” • Requires 2 (or More) Entangled Qubits in a Bell Pair • Transport State of Qubit A to Qubit B; Destroys Qubit A • Currently, Only Transports Short Distances within a QPU • Today, Longest Teleporation is 1,400km (870mi) • Note: Not Transporting Matter – Only Information
  • 24. Programming QPUs: Quantum Computing Simulators • Limited Qubit States Due to Digital RAM Constraints • 30 Qubits == 16GB, 40 Qubits = 16TB, 50 Qubits = 16PB
  • 25. Build: Managed Jupyter Lab Environments Fully managed infrastructure Pre-built environments Learning resources Build and test algorithms
  • 26. Test: Hosted Circuit Simulators Serverless execution Optimized performance Individual circuits and hybrid jobs Schrödinger simulator Tensor network simulator …
  • 27. Run: Managed Execution of Hybrid Quantum Algorithms Users Local IDE Fully managed execution Pay only for what you use Change backend with a single line of code Amazon Braket Classical compute Simulator Amazon S3 Amazon CloudWatch QPU
  • 28. Programming QPUs – Random Number Generator circuit = Circuit() a = circuit.h(0) # Put data into Qubit Superposition b = a.cnot(0, 1) # Entangle Qubits task = device.run(b, s3_prefix, shots=100000) task.result().measurement_counts Counter({'11': 50084, '00': 49916})
  • 29. Programming QPUs - Teleportation • Step 1: WRITE Conventional Data from CPU into QPU • Step 2: Put Data into Superposition using HAD Operation • Step 3: Entangle Qubits using CNOT Operation • Step 4: Perform PHASE Ops on Qubits in Superposition • Step 5: READ Qubits from QPU back into CPU
  • 30. © 2020, Amazon Web Services, Inc. or its affiliates. All rights reserved. DEMO: Amazon Braket
  • 31. Looking Ahead • Quantum Research Improves Conventional Algos • ie. Recommendation Systems • Error-Correcting Qubits • Enables Deeper Circuits Across More Qubits • Quantum Now = Deep Learning 10-15 Years Ago • Limited by Hardware Availability … Until Now!
  • 32. AWS Center for Quantum Computing (Opens 2021) Near-term applications Error correction & programming models Quantum hardware and technologies
  • 33. Amazon Quantum Solutions Lab (Available Now) Quantum Computing Machine learning HPC Build and optimize deep learning approaches on AWS Build and optimize classical and quantum-inspired approaches on AWS Target use case Build quantum computing prototypes Develop new algorithms and approaches Benchmarking Results Collaboration Education Solutions
  • 34. Technology and consulting partners provide software and service solutions on Amazon Braket
  • 35. Thank you! © 2020, Amazon Web Services, Inc. or its affiliates. All rights reserved. Chris Fregly github.com/data-science-on-aws @cfregly linkedin.com/in/cfregly https://www.datascienceonaws.com https://aws.amazon.com/quantum-solutions-lab/