From Bits to Qubits: Can Medicine Benefit From Quantum Computing?
Download as PPTX, PDF4 likes619 views
Quantum computing offers potential advantages over classical computing by utilizing principles of quantum mechanics like superposition and entanglement. A quantum bit or "qubit" can represent more than the binary states of 0 and 1, allowing quantum computers to potentially solve certain problems like searching large databases and optimizing complex systems much faster than classical computers. Several algorithms like Grover's algorithm and Shor's algorithm demonstrate quantum computing's potential. Experimental quantum computers with a handful of qubits have been built by companies like IBM, D-Wave, and others. While still in early stages, quantum computing shows promise for applications in optimization problems in areas like healthcare, machine learning, materials science, and more.
From Bits to Qubits: Can Medicine Benefit From Quantum Computing?
- 1. FROM BITS TO QUBITS: CAN MEDICINE BENEFIT FROM QUANTUM COMPUTING? Mike Hogarth, MD, FACP, FACMI mahogarth@ucdavis.edu http://hogarth.ucdavis.edu
- 2. Early Classical Computers: “Colossus” • The Z3, Atanasoff-Berry Computer (ABC), and “Colossus” – All developed independently of each other – All three used *binary* arithmetic for processing – Colossus (seen to the right) developed by British codebreakers in 1943 to help in decryption • *not* the machine built by Alan Turing to decrypt Enigma.
- 3. Early classical computing: ENIAC 1946 • ENIAC - Electronic Numerical Integrator and Computer (1946- 1955) – One of the earliest general- purpose computer – Performed *decimal* arithmetic – 100Khz speed – Multiplying two 10-digit numbers took 2.8 seconds – Used in the Manhattan Project
- 4. What is a “bit”? ● Term coined by Claude Shannon in his paper A Mathematical Theory of Communication (1948) ● Claude Shannon also founded digital circuit design theory (1937) ● He attributed “bit” to John Tukey of Bell Labs who used it as a short for ‘binary information digit’
- 5. The solid state device - the transistor • TRANSfer resISTOR • Dec 16, 1947 - invented at Bell Labs by William Schockley, John Bardeen, and Walter Brattain • Provides a switch (on/off) the replaced the vacuum tube that could be miniaturized and wasted less heat
- 6. First transistor computer: CADET • Harwell CADET – Transistor Electronic Digital Automatic Computer - TEDAC (CADET backwards) – First fully transistorized computer – Dev. in 1951 by Harwell Dekatron Computer – Laid the foundations for the industry and methods for direct numerical solutions
- 7. The boolean nature of transistors The switch = transistor
- 9. Logic “circuits” and computation A “half adder” with carry “out”. Adds A and B and accounts for carry over http://www.electronics-tutorials.ws/combination/comb_7.html
- 10. Integrated circuits that can “add”: A 4-bit binary adder http://www.electronics-tutorials.ws/combination/comb_7.html
- 11. How classical computers ‘compute’ for you! Addition Multiplication Square root Intel i7 CPU ‘registers’
- 12. Computing with binary circuits and “gates” http://www.electronics-tutorials.ws/combination/comb_1.html “integrated circuit” using transistors - CMOS 4071 with four OR gates “integrated circuit” using transistors - 7408 with four AND gates
- 13. A modern microcomputer - designed with integrated circuits
- 14. Moore’s Law - when will circuits become quantum mechanical systems? 10 billion transistors!
- 15. There are also complex computations that cannot be done on classical computers Time taken to run (N) as a function of the length of the input Compute time Input length (digits for n)
- 16. Demonstrating the limits of classical computing with python... Try 2 (2 to the 10 million)10000000 In python: >>> 2**10000000 (use 7 zeros) Does not return after 60 seconds…. Try 2 (2 to the 1 million)1000000 In python: >>> 2**1000000 (1million -- use 6 zeros) Returns a number in about 3-4 seconds
- 17. Improving on the “bit” and Moore’s law limitations • What if you could have a “bit” that could have more than binary (“0” or “1”) states? • What if you could have multi- state bits with states that can depend on other bits?
- 18. How we arrived at “quantum computing” • Story begins before any concerns about classical computing or practical generalized applications for quantum based computers • Rolf Landauer - physicist, IBM Fellow at IBM’s Thomas Watson Research Center in NY – First to highlight that computation is physics at the hardware level – One is harnessing physics to perform information processing – Landauer Principle -- energy must be expended to erase information Rolf Landauer (1927-1999)
- 19. Richard Feynman (1918-1988) • Nobel prize in physics in 1965 - for work on quantum electrodynamics and the physics of elementary particles • Invited keynote at MIT “First Conference on the Physics of Computation” • Feynman proposed simulating quantum mechanical system using a computer based on the same principles (a quantum computer)
- 20. David Deutsch • Originally described an experimental computational system where: • Submitted idea in 1978 in a paper sent to Physics Review but was rejected. Did not re- submit to other journals. • Resubmitted paper after being invited to do so - to International Journal of Theoretical Physics in 1985. It is the seminal paper on quantum computing today “a conventional computer operating by quantum means that had some additional quantum hardware that allowed it to do something extra” (Deutch) (age 62)
- 21. Key Concepts: Deutsch 1985
- 22. What physical things can be used as a quantum “multi-state” bit for computing? • Photon • Atomic nucleus • Electron • Magnetic field
- 23. Electrons as multi-state “bits” • An electron’s spin -- can only have two values (up/down) - like a bit • In a quantum mechanical system, electron spin can be in many (probable) states at once before it is measures (superposition) • The fundamental quantum-mechanical nature of “spin” makes it an ideal candidate for use as a quantum bit (qubit) • Individual spins can be initialized, coherently controlled, and read out using a variety of techniques (optical, electronic)
- 24. Welcome to the “qubit” https://www.cbinsights.com/blog/quantum-computing-explainer/ ● Concept originally introduced by Stephen Wiesner (1983) in his proposal for “quantum money” ● Term “qubit” attributed to Benjamin Schumacher - invented in jest because it sounded like “cubit” ● Schumacher described a way of compressing states from a quantum source of information so they require less physical resources to store ○ “Schumacher compression”
- 25. 3-qubits and 8 states at the same time https://www.doc.ic.ac.uk/~nd/surprise_97/journal/vol4/spb3/#1.1 Quantum computer basics
- 26. Three key features of quantum computation • Superposition • Entanglement • Annealing
- 27. Quantum-mechanical systems and improving on the “bit” • Superposition –A quantum-mechanical system can have a particle be in multiple states at once! –“Quantum bits” can be in a superposition of states –Adds significant computational advantage to a qubit over a bit • 2 classical bits → a state representing either 0,1,2,3 (1 number at any time • 2 qubits → can represent all 4 numbers at the same time
- 28. Quantum Superposition • The quantum mechanical property that has an atom, electron, or its spin, or its magnetic field to be in two ‘positions’ (states) at the same time
- 29. Entanglement • Physical phenomenon that occurs when pairs of particles are generated in ways that the quantum state of each particle cannot be described independently • One must describe the quantum state for the whole system (A Hamiltonian) • Described by Einstein, Podolsky and Rosen (EPR) as the “EPR Paradox” – Einstein considered it “spooky” – The “spin” of the particles are “entangled” changing one, changes the other instantly – Has been proven to happen with particles even as far apart as 15m
- 30. Quantum Entanglement • The ability of quantum systems to exhibit correlations between states within a superposition • If we have two qubits, each in superposition of 0 and 1, the qubits are said to be entangled • Seen as a powerful computational feature of quantum computation • Interference -- if we examine/measure one qubit’s state, entanglement causes us to erase the rest
- 31. Quantum Annealing • “a method for finding solutions to combinatorial optimisation problems and ‘ground states’ of systems” • What it does at the quantum level -- finds the lowest energy state in a system • By letting a system cool and go through sequential states, it will “anneal”, one can find the lowest energy state • Uses equations that describe the total energy of a system - a “Hamiltonian” Finnila, Gomez, Sebenik, Stenson, Doll. Quantum annealing: A new method for minimizing multidimensional functions. Chem Physics Letters. 219(1994) 343-348
- 32. Annealing - reaching the lowest energy point with a specially designed quantum computer
- 33. The physics of “annealing”
- 34. Results using Quantum Annealing
- 35. Practical applications for QA • Combinatorial optimization - “traveling salesman problem” • Integer factorization (breaks RSA) • Search in unsorted databases (Grover’s) • Pattern recognition • Protein folding
- 36. Classical vs. Quantum Gates
- 37. The fundamental Quantum NOT gate
- 38. A new truly ‘quantum’ gate
- 39. The “Hadamard” gate • Hadamard gate - receives a qubit in state 0 as input and can return a qubit as output that is in a superposition of 0 and 1 (simultaneously) • Consistent with Shrodinger’s principle ---> measuring a system in superposition collapses it to 0 or 1 but probabilistically – If the Hadamard gate gets a 0 as input, there is a 50:50 chance of seeing a 0 or 1
- 41. Grover’s Algorithm • Lou Grover 1996 • Uses qubits in superposition to compute ‘searches’ much faster than classical computers • “Searches” = generalized search –Finding an item in an *unstructured* list
- 42. Grover’s -- find item “w” in a list of N items https://www.youtube.com/watch?v=hK6BBluTGhU O - operation N - number of items in the list
- 43. https://www.youtube.com/watch?v=hK6BBluTGhU Grover’s algorithm is a quantum algorithm that will perform search in less time -- lowers it by the square root of the total items in the list
- 44. How does it work? • It puts all the qubits in multiple possible positions (superpositions) • 3 steps – Step 1: Hadamard gates - puts qubits in superpositions (all possible positions for x) • Makes the qubits have a uniform amount of energy – Step 2: Oracle function - flips amplitude of only the item being searched – Step 3: Hadamard gates after Oracle function - applies state change to the qubits and amplifies the value of the item being searched – Repeat steps 2 and 3 until amplitude reaches ket “w”, meaning probability is high the result is correct
- 45. Grover: Step 1 https://www.youtube.com/watch?v=hK6BBluTGhU Use Hadamard gate to put all qubits into superpositions with the same energy state
- 46. Grover: Step 2 https://www.youtube.com/watch?v=hK6BBluTGhU Flip spin amplitude of the item of interest in the list
- 47. Grover: Step 3 https://www.youtube.com/watch?v=hK6BBluTGhU Use another Hadamard gate to re- flip the item of interest and amplify spin - repeat steps 2-3 until amplitude reaches a high enough reliability level
- 48. Peter Shor’s Algorithm and Prime Numbers https://science.mit.edu/research/faculty/shor-peter-williston Look out RSA encryption!!
- 50. Quantum Computing at IBM Free access to IBM 16-qubit machine IBM Quantum Computing Service
- 51. IBM Q - crowd-sourcing quantum computing
- 52. IBM Q Composer
- 53. IBM 5-qubit Quantum Computer Real IBM Quantum Chip (5 qubit)
- 54. You can run Grover’s on the IBM too!
- 55. Implementing algorithms using “gates” in a drag and drop ‘composer’
- 56. D-Wave - the first commercial quantum computer
- 57. The D-Wave quantum transistor - the SQUID ● Superconducting QUantum Interference Device (SQUID) ● Made of niobium, becomes superconducting at low temperatures ● A very sensitive magnetometer that can measure very subtle magnetic fields, based on superconducting loops containing Josephson junctions ● The transistor behavior: ● The SQUID stores two magnetic fields, which either point up (+1) or down (-1) ● Each SQUID is a qubit that can be controlled and put into a superposition of the two states
- 58. D-Wave Coupling ● Multi-qubit D-Wave processor has qubits connected to each other through couplers ● Couplers cause qubits to influence each other ● Mathematically, these elements couple together qubits, set as variables, providing parallelized solutions to multi-dimensional computation ○ Ie, optimization problems where changing one element requires re-computing of the others ● Readout device attached to each qubit - inactive during computation (do not affect qubit behavior), but read output once computation has finished 8 qubit loops with 16 couplers ‘connecting’ each qubit with 4 others
- 59. D-Wave - hardware at a glance
- 62. Utility of Quantum Computing
- 63. Quantum computing in medicine
- 64. Optimization problems in healthcare • Anything that requires a high number of variables and their combinations -- massive variable problems = “optimization problems” • Best ED throughput • Best treatment strategies through pattern matching
- 65. Deep Learning Model and Quantum Annealing
- 66. Correctly identifying handwritten numbers
- 67. Deep learning for digital pathology
- 68. FDA and Deep Learning
- 69. Proteins and modeling structure • Understanding how proteins fold • Modeling malfunctioning proteins and their physical structures http://www.atelier.net/en/trends/articles/quantum-computing-set-revolutionise-health-sector_437915
- 71. “Quantum Informatics” • A new field of information science that optimizes applied information processing using quantum computing devices • Just an idea/concept -- what do you think?
- 72. Quantum Computers in the News
- 73. UK National Quantum Program
- 74. Rapidly Evolving Quantum Computing Designs
- 76. Questions?