THE STORY OF QUANTUM COMPUTING 8 7g Gitergali — VenkatiowmanFOREWORD The book youre holding containe the basics of quantum computing inthe form of a comio book. W'e charming and delightful and awesome. The author, Gitanjal Verkatraman, contacted me more-or-ese out ofthe blue with a draft ofthe book. | waen't ‘ure what to expect, bu [wae blown away by the craftemanip that went into the book ~ especially the beautiful iustratione con every page (the mirimait caricatures of my coleagues nover failed to make me em For a quarter of a century, most popularizatione of quantum computing have tried to take the eacy way out: a qubit ig juct a bit that's O and | at the came time, and a quantum computer solves hard probleme in an instant by trying every possible anewer in parallel. The problem with these pat ancwere ic twofold. Fire, they mice the rel beauty of quantum information ~ a beauty that hae to do with amplitudes, equare roots of probabilities (!, interfering and cancding each other out, kil off the wrong ‘anqwers to a problem and leave only the right ancwers ~~ an effect weird enough that | doubt any ecience fiction uniter would ever have had the imagination to invent i. Second, 've seen again and again how the pat anawore fil people with ideae of what {quantun computers willbe abe to doin practice that ae wildy off leaving them vulnerable to onake-oi ealosmen. Even though ite written fora broad autionee, thie Bock mekes a genuine effort to avoid the tomptatone of falechood and grapple withthe richness of the field. [hope you enjoy it ae much ac I dd Scott Aarongon Austin, TX August 2020INTRODUCTION AT THE SMALLEST SCALES IN THE UNIVERSE, AT THE LEVEL OF AN ATOM, THE LAWS OF PHYSICS ARE WEIRD. WE CAN) PRECISELY kivow WHERE AN ELECTRON Is, BUT WOT HOW FAST IT MOVES. AS FOR WHERE, AN ELECTRON COULD BE IW MANY PLACES AT ONCE. DESCRIBING THESE Is THE 208 OF UANTUM PHYSICS. QUANTUM PHYSICS TOMETHER WITH COMPUTER SCIENCE MAKE A NEW TYPE DF COMPUTER CALLED THE QUANTUM COMPUTER. IT USES THE WEIRDNESS OF THE QUANTUM WORLD 7D SOLVE PROBLEMS WE MIGHT Wor BE ABLE 70 SOLVE WITH SUPERCOMPUTERS. IT COULD CRACK CODES FASTER THAN SUPERCOMPUTERS, HELP US BUILD BETTER DRUGS AWD MATERIALS ALL WITH THE TAWTALISIVA PROMISE OF EXTRADRDINARY SPEED: WHY ARE SOME PROBLEMS HARDER To SOLVE THAN OTHERS? WHO THOUGHT OF MAKINA A QUANTUM COMPUTER? HOW DIFFERENT ARE THEY TD THE ONES WE USE NOW? ARE THERE THINAS IT CANNOT Do? TAKE A VISUAL TOUR OF ITS EVOLUTION THE PEOPLE, THE PHYSICS AND A FLAVOUR OF HOW WE MIAKT PROGRAM IT. THIS IS A FULLY ILLUSTRATED QuiDE, THE IS THE STORY OF QUANTUM COMPUTING,ACKNOWLEDGEMENTS THIS GOOk IS POSSIBLE ONLY BECAUSE OF THE SEVERAL INFLUENCES IV MY LIFE FIRSTLY, MY GRATEFUL THAMES TO SCOTT AARONSON, DIRECTOR OF THE AUANTUM INFORMATION CEMvTRE AT THE UNIVERSITY DF TEXAS AUSTIN, FOR RESPONDING To MY EMAIL AND FOR TAKING TIME TO GIVE VALUABLE mvpuTs To MY DRAFT AND FOR HIS VERY ENCOURAGING WORDS. MANY THANKS TD REBECCA PARSONS, CTO-THOUAHTWORKS, FOR HER UIVRVESTIONINA SUPPORT OF MY WoRk AND TO MY COUEAAVES WITH WHO 1 HAVE EXCHANAED IDEAS. my aRaTiTuDE To DAVE ELLIMAN FOR BELIEVING 1 COULD AYE EVANS MY ART TEACHER my BOWS K ASHOR FOR BEIVA THERE MY MOTHER, FOR EVERYTHINGCONTENTS WHAT 1S A QuanTuM compuTeR? WHY MIGHT WE NEED ONE? Ew oF MooRE’s cAW? ‘A TOUR OF COMPUTATIONAL COMPLEXITY ALL ABOUT RUANTUM, @uawTum Vs CLASSICAL PARTICLES AND WAVES BEGINNING OF QUANTUM THEORY fuantum weiRDwvess UPERPOSITION ENTANLEMENT REACTIONS TD QUANTUM MECHANICS APPLICATIONS OF QUANTUM MECHANICS QUANTUM COMPUTERS: A NEW IDEA HARDWARE OVERVIEW & CHALLEWAES THE EXPONENTIAL POWER OF GuAATUM QUANTUM BITS Bits, quaiTs, ETATES MATHEMATICAL MODEL OF AUBITS WAVES, AMPLITUDES L MEASUREMENT Lome gates ‘A FEW BASIC QUANTUM LomIC GATES GATES be ENvTANaLEMENT QUANTUM ALAORITHMS DEUTSCH 20254 GROVER ALAORITHM SHOR'S ALAORITHM AVANTUM COMPUTERS SOLVE EVERYTHING? D0 AND NOT Do ALQORITHM EFFICIENCY COMPLEXITY CLASS SEcueITY AND §UANTUM COMPUTERS BREAKING CRYPTO QUANTUM-SAFE CRYPTO a NeW Quantum vey DisrereutioN. EREDR CORRECTION PRORRAMMING AND TESTING PRESENT AND FUTURE WOMEN IN QUANTUM COMPUTING REFERENCES 6 “3 “a ” 1% 32. ae % 48 "4 100 tor tor 103 log 106 0 m2 16 naWHAT 1S A QUANTUM COMPUTER? $0 whar is A... +s RVANTVM COMPUTER? ARENT ALL COMPUTERS BuILT| | 1S IT CLEVERER THAW ALL OF Dn PHYSICS ? WHY Is @unATum | Us? DR FASTER? The word is ie? iE can think a trillton thoughts ab once any DIF eRENT? Tk rs a computer built OS sty using Qvantom PHYSICS) though. It lives in the coldest IS IT BETTER AT puyiva climes. lt is both music, SHOPPING, TELLING THE WAY, MAKING A GRAPH ? Does iT ExT? fragile and sharp. and capable of solving the hardest af problemsDO We NEED QUANTUM COMPUTERS NoT $0 MUCH NEED AS IT MAY BE INEVITABLE ESPECIALLY IN LIGHT OF THESE quesTIONS HOW CONG WiLL mooee’s caw ocd? @ Yr 2 ALSO, AVANTUM COMPUTERS ARE WOT INTENDED AS A REPLACEMENT 7D THE ONES WE NOW USEEND OF MOORE'S LAW? THE “LAW* we can expect the seed of computers fo double ery 2 years or so GORDON MOORE 1965 THE TRANSISTOR TRANSISTOR 1S THE SIMPLEST PROCESSOR IT is A swirey To Ler Elecreows Feow - of WOT Size MUCH SMALLER THAN HUMAN HAIR THE PROCESSOR PROCESSORS ONE HP Have cot conrains ‘BECOME Agour A eee Ho Bieeiow SMALLER TRANSISTORS THE END? AS THE TRANSISTOR $126 GETS SMALLER. a & + QUANTUM BFFECTS” WILL COME INTO PLAYPROBLEMS : DIFFICULT OR WOT? ANSWER A RIDDLE PLAY TETRIS SOLVE NUMBER PROsLEMS if ee me you want to me pu share me ou lose me WHAT Am I? WIN AT CHESS Sowve A SuDoKU °TO GET TO A SOLUTION WITH COMPUTERS with CompuTeRs DOES A SOLUTION EXIST? HOW MANY STEPS? 's THERE =A SoLUTION? How LOWA DOES IT TARE? AN AlaoRITH™ ? Awd WOW mucH OF THE ResouRces? DOES THE PRORRAM EVENTUALLY STOP? IS THE SOLUTION EFFICIENT?DIFFICULTY OF TOO MANY STEPS Ze she vw mae, oy el A TASK HOW MANY ADDITIONAL STEPS? - Lesa, ane i 5 A cane an ted, Gwen mareorenTs GIVEN mvaReDIENTS x vin we a aman rw) MULTIPLY NUMBERS swen 3 NUMBERS Given 100 umBeRs Sort go0ks BY TITLE 4IVEN A PILE OF BOOKS GIVEN A LIBRARY OF B00ks $yezs55F $23234 rea a ° SoLvE A RUBIK'S CUBE @ GIVEN AN 8X8 CUBE AIVEN A 2x2 CUBE G08 ©0608 00068DIFFICUCT ON SOLVING a PROBLEMSDIFFICULTY WITH TRIAL & ERROR 0000 001 9002 LS Leg PROBLEM. Look HARDER? FIND A NGEDLE IN A BoK ~ FIND A NEEDLE IN A HAYSTACK FIND PRIME FACTORS OF 1 Ez Find PRIME FActoRs oF 5493 Finis 29 FIND A 3-CETTER PATTERN F359 32 Find THE MATCHING PATTERW 8426033 I A VERY LONA SERUENCE “535 9595 ACROSS A FEW 100 SERUENCES 4197 6495 643?DIFFICULTY WITH MEMORY FIND MATCHING PAIRS A fl Baa = a H ‘KEEP IN MEMDRY HOw MUCH MORE TO RETAIN SwoPona ust oe 3 rrems Ww ECoG nes WHEN THE pur2le Has ~1000 Pieces A 4-PIECE SIASAW PUZZLE gaDIFFICULTY OF OPTIONS optimise | TX WI ol (Ss =- , Se ASS 6 WHAT IS THE SHORTEST ROUTE ‘ A PER, — — *How HARD 18 HANGMAN 2 SOURCE ? DICTIONARY TOTAL NUMBER OF worps : ~ 300,000 “zane Man IT PROBABLY DOES NOT MAKE MUCH DIFFERENCE WHETHER THE WORD YOU ARE AVESSINA 1s @ oR (IB) LETTERS Lona WORD LENATH 3 6CRACK A PASSWORD? USERNAME f-corbato SOURCE ; KEYBOARD an AVAILABLE KEYBORRD . gg FERRER EE CHARACTERS PASSWORD LENGTH > B o-4 A PASSWORD OF & LETTERS coULD BE ANY COMBINATION OF —@=® ArZ \ om mreft/-... THIS SET OF POSSIBILITIES IS EXPONENTIALLY LARKE — GROWING EVEN MORE IF THE PASSWORD IS A PASSPHRASEHOW DOES YOUR PROBLEM ARDW? — f TOUR OF CoMPUTATIOWAL COMPLEXITY —COMPUTATIONAL COMPLEXITY ALAN TURIVA GAME UP WITH A MODEL OF coMpuTATION THAT WE STILE USE TODAY. IT CONSISTS DE AN INFINITELY LOMA TAPE, come lwsreverions, AND A macHive (WITH A srATE) THAT READS AND WRITES SYMBOLS DN THE TAPE cA ofofe i fofoti tT Jofof ti Tots T' fo ALAN TRING 1936 ANYTHING A MODERN COMPUTER CAN DO, A TURING MACHINE CAN Do THE VERY NOTION DF AN ALGDRITHM Is BASED DN THE TURIVA MACHINE AN INTERPRETATION DF THE FAMOUS CHURCH-TURINA THESIS IS THAT ANYTHINA COMPUTABLE CAN BE WORKED ovT BY A TURIVA MACHINE, NATURALLY, THIS LED TO FURTHER INVESTIAATIONS ON WHAT THE LIMITS ARE, IE ANN, TO COMPUTATION, COMPUTATIONAL COMPLEXITY IS A FIELD THAT AREW FRom THE 1970-5. LET US Look AT SOME DF THE CONCEPTS aALAORITHM EFFICIENCY wnar 1s IT? (CW) ew AWB FICIENT PROGRAM = Ri FASTER OBVIOUSLY, AN ALGORITHM THAT an Serums Fi ACHIEVES 17 GOAL IN FEWER an muses cess eesoveces STEPS 1S MORE EFFICIENT an a= oS Ca = J Gar om EFFICIENCY IS INDEPENDENT OF EFFICIENCY DEPENDS ON pur SIRE SHARD WARE WHEN INPUT siZ6 AXONS How MucH DOES RUNTIME SLOW? SIM PLEMENTATION THIS 18 A TIME COMPLEXITY MEASURE IsPOLYNOMIAL TIME P ‘A PROBLEM IS wv CLASS P JE THE RUNNING TIME OF THE BEST ALGORITHM GADWS AS A PoweR OF THE INPUT SIZE vf, n,n, re, nt 20s THAT fA sEckTOP/ SMARTPHONE cA DO ARE WP Such AS Fiepins ROUTES, SEARCHING, PLAYINA GAMES Examples MULTIPLY SORT SOLVE RvBIKS CORE 6WOW DETERMINISTIC POLYNOMIAL TIME : wp SOME PROBLEMS APPEAR MUCH HARDER THAN P, TAKIN FAR COWAER TO SOLVE A PROBLEM 16 Ww CLASS NP WHEN ITS SoLUTION CAN BE EASILY cHECEED CLASS WP PROBLEMS ARE VERIFIABLE sev PoLynomiaL Time MANY OF THE WP PROBLEMS CAN BE THOUAMT oF AL SEARCHING THROVAH A LOT oF PossrB/LITIES EXAMPLES FACTORISE A NUMBER. PACK A SUITCASE SOLVE a supoku | 7 2 [racToRisarion 1 4 cone sreP] [me ewcayprion mivowvina 264)NOT POLYNOMIAL TIME UIVSURPRISIVGLY, THERE ARE ALSO PROBLEMS THAT CAN WEITHER BE SOLVED WOR VERIFIED IN POLYNOMIAL TIME FOR SOME PROBLEMS, THERE IS PROOF Har THEY CANNOT BE SoLvED AND FOR SOME, THEGE I6n'T EVEN A PE00F ERAMPLES CHESS ‘There is & far more detailed Aierarchy of difficulty levels —based om formal definitions and ewolving from mew rettarch EXERCISE = LODKUP — comp lexity200- u waterloo: CaPevs WP EFFICIENTLY GFEICIENTLY LOLVED VERIFIED PROBLEMS PROBLEMS NP CONTAINS P NP THE ALaoRITHM FOR SoLVINA @ THE PROBLEM, BY BEIvA CORRECT, VERIFIES THE 0( VTIOW IS PsP? CAN EFFICIENTLY VERIFIED proBiems BE EFFICIENTLY SOLVED 700? 17 15 BELIEVED THAT Pt WP Bur THERE 1s WO PROOF ‘THIS IS A MAd0R VIVSDLVED PROBLEM IW Computer science /maTHemarieei@so IMPORTANT THAT CLAYMATH. ORG OFFERS FOR THE ANSWER Ae bene IT GREAT INCREASES THE StOPE oF WHat we CAN KWow Amp DISCOVER Cand most erype would break) X poem tne ARE WE LEFT WITH ALAORITHMS THAT &O THROVAH EVERY oprtow UNTIL THE GND oF THE woRco.~ ALE WE % ACCEPT THAT THERE ABE PROBLEMS WE cavvor soxvE.-- LIKE GSDEL suowep THAT THERE ARE Tave ARUMETE STATEMENTS THAT AKIOMS tA MATHS Chw'r HELP To PROVE UKE TURING SHOWED THAT THERE 1 WO PROARAM THAT CAN DECIDE IF ANOTHER PLOMRAM WiLL HALT oR WOT.WHAT OPTIONS DO WE HavE ? BY wo sour i A REASONABLE AmouNT oF TIME 0 SPACE Can we even hope to solve frobkms harder than P 2 Are we clever enough 2 quantum. computers ? SOLVED EFFICIENTLY z 20SO, QUANTUM... QuanTus is QUANTUM IS QUANTUM mecHawics LATIN FOR THE SMALLEST DESCRIBES NATURE ‘How GREAT’ ‘Dlece’ OF EWERAY AT THe TIMIEST DF SCALES THE NOTION OF QUANTUM CAME AT A TIME WHEN THE PHENOMEIVA In THE mAcko /CLAGSICAL WORLD HAD BEEN WELL EXPLAINED BY MATHEMATICAL FORMULAE 2THE CLASSICAL WORLD VIEW ZY NEWTON LAD OUT A DETERMINISTIC PICTURE OF our wokld é His MATHEMATICS ACCELERATION VELOCITY DESCRIBED AND PREDICTED — INERTIA MOMENTUM @ THIS PREDICTABLE Cuockwork vIEW @ OF THE UWIVERSE WAS S00 4D 14 To CHANGE action REACTION PLANETARY MOTION 22QUANTUM Vs CLASSICAL $ THE QUANTUM AND THE CLASSICAL ARE WORLDS APART | “4 THEY SEEM 7 opeeaTé OW DIFFERENT Laws BEMAVE DIFFERENTLY How D0 WE BEGIN TO UNDERSTAND og COMPARE THE TWO? 4 00D PLACE TO START I 70 Look AT PaRTicLes &k WAVESA SIMPLE EXPERIMENT wie PARTICLES DO WHAT WAVES DO Fic A bu KN “wird WATER FIND, A BOARD Dore GAPS A RIPPLE ey ‘wwe SIDE TRAVELS PAS; vat, ars AWD MARES PEAKS VALLEYS SAND OLES LATE T PILES WHY IS THIS RELEVANT? BECAUSE SIMILAR EXPERIMENTS I THE QUANTUM WoRLD GAVE VERY UNEXPECTED RESULTS 24THE SVBATOMIC WORLD THE GREEKS GAVE Us THE WORD FOR ATOM BuT IT TOOk THE MIND 0F @ THis was PROBABLY ALBERT EINSTEIN OUR FIRST PEEK TO PROVE THAT INTO THE ATDMS EXIST QUANTUM woeLD THE FASCINATION WiTH LIGHT HOWEVER, BROUGHT US CLOSER TO UNDERSTANDIVA ‘QUANTUMMVESS’ 2sAT FIRST I(T WAS LIGHT CHRISTIAN HUYGENS LUQHT IS MADE WHAT Is LIAHT? OF CORPUSCLES. ANCIENT AREEKS THOUAHT THAT A BEAM FROM THE Eyes ENABLED US TD SEE 1S UGHT WHITE? WHAT DoES IT weIaH? 1S rT mveivITeLY FAST? LiGHT BOUNCES, ScATTERS, BENDS, DIFFRACTS, ..- wy ? QUESTIONS AND THE DESIRE To Use AND CONTROL LIAHT a ED TO mTeRestinG ANSWERS AND DIECOUERIES CUEERIERNELED LES 26THE BEGINNING OF QUANTUM THEORY ELECTRON ABSORB AND EMIT THUS BEGAN A RICH FIELD OF sTuDY— @ WITH CONTRIBUTIONS FROM SCIENTISTS LIKE WOLFGANA PAULI NIELS BOHR THIS FIELD Is NOW CALLED QUANTUM MECHANICS WE CANNOT KNOW EVERYTHING ABOUT NATURE'S BUILDING BLOCKS AT THE SMALLEST SCALE ATURE 15 UNPREDICTABLE WERNER HEISENBERG ERWIN, SCHRODINAER 27OBVIOUSLY, A NEW VOCABULARY PBA Voor 28BUT FIRST, THE EXPERIMENT: WITH LIGHT LiaHT (PHOTONS) SHOWE THROUGH Two NARROW SLITS oe FORM ALTERWATE DARK AND LinHT BANDS An INTERFERENCE PATTERN EXPERIMENT BY THOMAS YOUNG 1800s 24PHOTON PEEKA&EOO IT Seems, ADD A DETECTOR WHEN LEFT ALONE To SEE wHich sur TT THE PHOTONS. use nom suirs | = UKE A RIPPLE Like pouring san> THROUGH WATER, TWO BANDS FORM WHEN YOU Dow'T Look, wHeN You L008 IT SPREADS LIke A WAVE a iT “Couarses’ To A PARTICLE 1% 6 30THE EXPERIMENT-WI/TH MATTER CURIOUSLY, ELECTRONS SENT ONE BY ONE ALSO FORM THE INTERFERENCE PATTERN. THE STUFF THAT. 7 MATTER 1S MADE OF BEHAVES LIke waves ELECTRON Aun HOW DES THIS ReconciLe WITH Newton's AN OUTCOME (NEARLY) IMPOSSIBLE TD UNDERSTAND OR EXPLAIN 3QUANTUM WEIRDNESS LET us FIRST Look AT THEM CONCEPTUALLY FROM A CLASSICAL VIEW C THEN POKE INTO | See THEIR WATURE moRE 32QUANTUM SUPERPOSITION THE QUANTUM PARTICLE (eg. Puoron, Evecreow) SPREADS OUT LIKE A WAVE IT “ExIsTs” as A e WAVE OF PROBABILITY A STATE OF BEINA IW MULTIPLE LOCATIONS AT ONCE EACH WITH A DIFFERENT PROBABILITY vst AS WAVES IW WATER ADD UP To FoRM A RESULTANT WAVE... QUANTUM STATES ADD uP 7> FORM A SUPE RPDS TIOW 33SUPER POSITION: A MACRO PICTURE e Va A COIN BEING FLIPPED Dead on 7 \ is HEADS thesia? o s Is TAILS A MIX OF BOTH WITH A 50-50 CHANCE THE SCIENTIST'S CAT NEXT TO SOME POISDN Is ALIVE {Lk xm XPRESS IN Ean A WuBE You See THE cLAEsIcAL WORLD WHAT FACE IS IN FRONT? THE BLUE? pens a feature THE ORANGE? 4 of an elechon roughly like AN ELECTRON WITH SPIN a magnetic THAT's UP , tee V THATS Down | UNTIL You cHEci ! 34EINSTEIN WASN'T A FAN WHY 20 WE WEVER SEE A HALE ~ ALIVE HALF DEAD CAT FOR REAL? a ——— NIELS BOHR ui A ADE 01 el THEN, AM 1 TOO SPREADING ALBERT EINSTEIN: our- Like A wave? EINSTEIN WROTE A PAPER TD SAY THAT QUANTUM MECHANICS WASN'T COMPLETE AND THAT IT COULD NOT EXPLAIN ITS OWN PREDICTION DF ENTANALEMENT 35ENTANGLEMEWT. THE CORRELATION UNDER THE RIGHT ConDITIONS ae TWO ELECTRONS oR PHOTONS can) BE ENTANALED THIS MEANS: THE PROPERTIES OF THE TWO PARTICLES ARE “RELATED” EVEN IF THE PARTICLES ARE SEPARATED hess which — MEASURING ONE Rock, Maper, Sessors ? INSTANTLY DECIDES THE STATE OF THE OTHER FASTER THAN THE SPEED OF LIGHT! ENTANALED coms wou. {_aoTH mencvee Hens x a the ofher particle Guess which — Rock, Paper, Scissors? 36ENTANALED PARTICLES INFLUENCE EACH OTHER'S STATE EVEN IF THEY ARE GALAXIES APART ENTANALE SEPARATE ; M% & oe ‘eo @ 37ENTANGLEMENT CONUN DRUMS Entanglement? It ts spooky action EvsTern's FINAL PAPER at a distance! ARAUES THAT ENTANALEMENT HAS A PREDETERMINED DUTCOME can avanti, mecnanieat DéscRO TOW OF PrrsicaL Reacrry ag corners? = figstem = hosing? SPEED OF LIGHT ya36 IF ENTANGLED PARTICLES ACROSS GALARIES COULD STALK’ AND INFLUENCE FACH OTHER'S STATE THE MESSAGES WoULD TRAVEL SHOCKING, BECAUSE IT WOULD TAKE LIKHT YEARS 70 TRAVEL BETWEEN GALAAIES 3REACTIONS TO QUANTUM MECHANICS IF QUANTUM MECHANICS HASN'T PROFOUNDLY SHOCKED YOU, YOU HAVEN'T UNDERSTOOD IT — NIELS BOHR I THINK | CAN SAFELY SAY THAT ND ONE UNDERSTANDS QUANTUM MECHANICS ~ RICHARD FeYWMAN 1 pow'r UKE IT AND I'M goRRY 1 HAD ANYTHING EVER TO DO WITH IT “ERWIN SCHRDDINAER AOD DOES NOT PLAY DICE ALBERT ElvsTeIn QUANTUM MECHANICS IS THE OPERATIVG SYSTEM THAT OTHER PHYSICAL THEDRIES RUN oN AS APPLICATION SOFTWARE ScorT anzonson 34WHAT G00D IS QUANTUM MECHANICS ¢ QUANTUM MECHANICS 1S RESPONSIBLE FOR DEVELDPIVG Lasers NUCLEAR REACTIONS TRANSISTORS MEDICAL DEVICES 4oQUANTUM IN COMPUTERS 2 THE IDEA OF A QUANTUM COMPUTER CAME FROM ¢ a» RICHARD FEYNMAN AND How IT MIGHT WORK WAS SUAGESTED BY DAVID DEUTSCH EVERY FINITELY REALISARLE PHYSICAL SYSTEM CAN BE PERFECTLY SIMULATED BY A UAIVERSAL moODEL COMPUTING MACHINE OPERATING BY FINITE MEANS ie Simulating Physics with computers by Ricunen FEYNMAN 1985 Quantum theory, church Tang Principle & Universal Suantumr Computer by Davin Deutsch 41WHAT IS wT ai iii /T MADE OF? also called | such as evecraows|PHoToNs sohose SPIND or can be in | and a handgul ave GNTANALED avanTum ALADRITHMS / ave procedures using AUANTUM LOGIC GATES on quantum particles to if COMPUTATION SPEED UP exploring several options simultaneously ending in a MEASUREMENT with a high PROBABILITY < alan temperatures to quantum particle vraripapied by MICROWAVE PULses ino quantum computer LASER BEAMS ma quantum computer 4eDECOHERENCE . HARD To KEEP QUANTUM computes ) sepa : . ‘om@ J oS aa wns 10 86 of —— MOLATED F0m i peta [ ea hee ITS surRounoinas vibrations 0 Keep t J QUANTUM BEHAVIOUR sich wget olds entanglement superpailon I decoherence er" &. | { $ Tws 1s aw expensive process 65 ANY EXPOSURE TO ‘NOISE’ LEADS TO LOSS OF AUANTUMMESs - THIS IS CALLED DECOHERENCE 43QUANTUM COMPUTERS ARE EXPONENTIALLY POWERFUL = Rom THE SHEER AMOUNT DE INFORMATION IT CAN PoTEMTIALLY HOLD/PRocESS —100 QuBITs THERE ARE ~10°! STARS IW THE WihHT sey 30 100 QUBITS CAN REPRESENT ~ /D STATES300 QU6ITS EY THE ENTIRE OBSERVABLE UMIVERSE Commas ~ 10 PARTICLES 90 300 QUBITS CAN REPRESENT ~ 1D STATESQUANTUM BITS — keepers OF WITH MUCH GRATITUDE THE 10 INFORMATION — umesH vazieani /ye BERKELEY& NOTE ON THE MATHEMATICS AN INDEPTH STUDY OF RUANTUM COMPUTATION RERUIRES SOME KNOWLEDAE OF? @ PROBABILITY THEORY © Linear ALasen vecrors/mareices /élensturr |TENCOR propucrs BIG 0 NOTATION PHYSICS THAT 1S WOT THE AOAL OF TH BOOK. WE WILL IwsTEAD SRAW CONVENIENTLY FROM HIMH SCHOOL MATHEMATICS. @ mucripcy ALaeBeArc EXPRESSIONS ) RECOMISE PYTHARORAS’ THEOREM KNOW REAL IMAGIVARY WUMBERS AWD ABSOLUTE VALUES UNDERSTAND RPOWENTS | POWERS BEIVA FAMILIAR WITH HOW A COMPUTER caw BE PROALAMMED Is ESSENTIAL, ANY OTHER SYMBOLS OR CONCEPTS WILL BE EXPLAWED AS WE AO ALMA 4sBIT & &vueir ~ BINARY DIAIT my ONLY ONE OF Two STATES s+ QUANTUM BIT IN A {uPERPOSITION “FOF O + to some extent vier D0 some eatent can represent & possible stahs each wilh a certain likelihood ».@-@0e @® ° ° 7 ° ® ° 4 ® o can represent 8 possible stales each with a certain likelihood agQUBITS AND STATES ADDING ONE MORE QUBIT Doveles THE MUMBER OF POSSIBLE STATES THIS INCREASE 1S EXPONENTIALMEET THE QUBIT GEOMETRIC VIEW OF THE Qv8IT Basis A RUBIT EXISTS RELATIVE TO THE Basis states (03) any UD QuANTUM STATES Wy Joy | @kounn szare | Down state 11) | excites stare | up srate [yy avert state NOTATION MATHEMATICALLY, A QUBIT IS THE ARROW THE BAR +t BRACKET | > IS THE MARKED W (GReEK LETTER PS) MOVING KET NOTATION CREATED BY THE AROUND INSIDE A BALL OF RADIUS 4 UNIT PHYSICIST PAUL DIRAC BIQUBIT SUPER POSITION , GEOMETRIC VIEW OF THE §UBIT SUPER POSITION Wn " pescarges TO WHAT EXTENT A Ye guair is ww JOY OR Ww iy” Superposition Joy wheve IEERiorae %, Baye comple numbers AMPLITUDES . PRE-PROBABILITIES ” = SCOTT AAROWSON A RUBIT STATE EQUAL SUPERPOSITION VIS A SuPERPoSITION OF |0> AND [I> THE QUBIT 1S ERUALLY LIKELY TO BE IN STATE [o> OR STATE |ID f 1s wor Born joy ano |1> Ipd+ z 10 + gl? 52QUBIT, MAKE UP YouR MIND GEOMETRIC VIEW OF THE Qv8IT In 1¥> 77 }\ to surerosinion oF ® aver 1s |¥> = o2/0>+ BID AMPLITUDE = AMPLITUDE OF |oy f= AMPLITUDE OF JI MEASUREMENT WHEN A QUBIT Is MEASURED, ITS SUPERPOSITION ‘COLLAPSES! IT chooses A state [07 ve |) PROBABILITIES OF OUTCOMES prosasiuity oF [oy = |a/* meosasiity oF [I> = [B/* 53y FROM SUPERPOSITION TO DEFINITE POSITION THE @vBIT DES Rovwp THE CIRCLE of THE PATH IT TRACES 16 REALLY A WAVE CIN ITS AMPLITUDE DETERMINES THE PROBAsiLITY LT OF THE ovTcOME (2trO or one) ff -X MEASURINA, THE QUBIT CHANGES THE AMPLITUDE TO A PROBABILITY THE Hidwee THE progeBiLiTY THE MORE LIKLY THE OUT(OME b LET US TAKE A CLOSER LOOK AT THE JOURNEY FROM QUBIT TO BIT 54WHAT A0ES AROUND [Ss A WAVE @© xerace nue nvmases O me a wn REAVUR How Fak AWAY cuock THEY ARE FROM THE HORIZDNTAL AWAY FROM 26R0 THE HOUR HAND TRAVELS AFTER 12:00 SEE A WAVE: PATTERN FORMING ssABOUT AMPLITUDES AMPLITUDE IS A PROPERTY DF waves PERIOD AMPLITUDE Wwaveu AMPLITUDES CANCEL EACH OTHER AMPLITUDES ADD UP NS + NS WHEN WAVES PASS EACH OTHER, THEY INTERFERE AMPLITUDES INFLUENCE OUTCOMES THIS INTERFERENCE CAN BE USED = TO CANCEL WRONA ANSWERS ~ TO AMPLIFY CORRECT ANSWERS WITH A HIGH PROBABILITY 56PROBABILITY vs AMPLITUDE COIN STATES REAL WwoRLD ouTcoMes HAVE @ @ PRoBagiLiTiEs. Yo Yo PROBABILITIES ARE REAL NUMBERS SUCH AS 0, oF 2/3 , 0-777, 1 NEVER NEGATIVE SUM OF PROBABILITIES = 1 SUM OF PROBABILITIES t+! QUBIT STATES A avaiT [¥>» ecjoy+ Bli> ewos mv [oy O% [1 avanTum woRLD burcomes HAVE AMPLITUDES é AMPLITUDES CAN BE POSITIVE, WEQATIVE oR COMPLEX NUMBERS se EIVABUNG THEM 70 INTERFERE SUM oF PROBABILITIES = |repestune|* = ProsasiLity 7 st pal, ten we amines ce pat, SUM OF PROBABILITIES [el"+ fal” 67MEASUREMENT CHANAES AN AMPLITUDE TO PROBABILITY DESCRIBES A STATE / POSITION gh 4 ALTERS WHAT Is BEING MEACURED IT FORCES THE QUBIT TO MAKE A DECISION IT Loses THE sSuPERPOSITION GIVES A RANDOM ANSWER LT ed /| oeA RANDOM ANSWER IS NO A00D HOWEVER, A QUANTUM COMPUTER cAN BE COAXED IWTO A MORE LIKELY ANSWER DD someresrriow i 3 > . © @ — ze0 0 @ @ 2 QueiTs @~ oe | ourcomes @— 4 anmurvves EVERY QUBIT HAS EVERY ouTLoME @ THREE 2 POTENTIAL OUTCOMES HAS AN AMPLITUDE led 44 \%- AMPLITUDES vee INTERFERE . THE RULES WORK cAN BF AND cancer ovr = P roneren wn conreoceD =) THE unwanTED =o RVANTUM ALAORITHM BY RULES (be Gates) OWES Coe auawtom eseeur) THIS MAKES THEM... CHANDY THAT AMPLITUDES CAN Be wenarive ]ComMEX) SO... 59A LIKELY CORRECT ANSWER A QUANTUM — (AURORITHM) Is MADE UP oF GATES / RULES CH OROPUERPERARPORUBES oF avers Such THAT WHEN AMPLITUDES — [NTERFERE THE AMPLITVDES OF THe weowk ancwers — CANCELNEACHMOTHER, THE AMPLITUDES OF THe Rica answers — IBDUUBYARSWAMPLED So, ow — (MeASvREMEAT we FWD eae \ i ' 7 aia ott 60QATES: A MEANS To CALCULATECLASSICAL LOGIE GATES ELECTRONIC CIRCUITS USE LORIC GATES TO PERFORM CALCULATIONS / = i uy RATHER UKE GATES, OME OB Two STATEMENTS Nor ALL aeT THROUGH ARS examen AT eneH Gare THESE THREE GATES FORM THE BASIS OF ALL CALCULATIONS 3» PASSES TRUE OMLY IF PASSES TRUE IF ONE BOTH STATEMENTS ARE TRUE OF TwO STATEMENTS Is TRUE WHAT passes DEPENDS ON THE RULES OF ThE 4ATE PASSES THE OPPOSITE OF THE STATEMENT 62QUANTUM LOGIC GATES BEHAVES UKE A NOT-4ATE PHASE FLIP CHAWAES THE SIAN oF /I> ACHIEVES ERUAL SUPERPOSITION FLIPS SECOND IF FIRST 16 1 QUsITS OBEY RULES IW Aw ALAOEITHM THESE RULES ARE CALLED HATES GATES ARE APPLIED TO OWE 0k MOLE AUBITS AT A TIME THE ATES MAKE QUBITS MOVE ALOUND THE BLOCH SPHERE THIS CHANAES THE AMPLITUDES OF EACH OUTCOME DF THE guBITS OBVIOUSLY, THERE ARE OTHER GATES, Bur THESE FOUR ALE GOOD FOR A START 638'T FLIP GATE CX) IT IS A QUANTUM NOT- GATE xfer =) x e ral ROTATION ABOUT THE 45° AXIS A OBIT FLIP IS A HALE~CIRCLE ROTATION. AROUND THE 45° ANALE 70 GO FROM STATE Y, TO Yo DO THE BIT FLIP TWICE AND WE ARE BACK TO WHERE WE STARTED FOR A QUBIT IW SUPERPOSITION bhPHASE FLIP GATE (2) IT 1S A PHASE (SIAN) CHANAE GATE PERE “ip =I 2 moxeI eo alo FOR A QUBIT IW SUPERPOSITION ROTATION ABOUT THE |0> Axis ly ‘A PHASE FLIP IS A HALE~CIRCLE ROTATION Axovwd THE |0y STATE TO GO FROM STATE Y, TO Ye DO THE PHASE FUP TWICE AND WE ARE BACK TO WHERE WE STARTED 65HADAMARD GATE CH) PUTS SUBITS IN EQUAL SUPERPOSITION FOR A QUBIT IW SUPERPOSITION Jo» > ilo + 219 = I» 1 (&lo> + AID) = x (4 lor 21?) +B -4 1) SH g/0- lp 21? + lam ain + plop) : Joy + L (ep) Hy Zoran t+ x 21) I>, ]-> ARE AT 90 ROTATION ABOUT THE 22.5° Axis Nl A HADAMARD 1 A HALE CIRCLE ROTATION AROUND THE 22-6 ANALE TO G0 FROM o> To |+> oR FROM I> TO |-> DO THE WADAMARD TWICE AND WE ARE BACK TO WHERE WE STARTED 66CNOT GATE THE (CONTROLLED WOT CwoT GATE ACTS ON Two &UB/TS FLIPS THE SECOND QUIT JE THE Fest RUBIT Is 2 Quantum circuit contro/ a | 82 | cwor [817 [ey o | 0] 00 iy 0 I or (92) /92) 1} otar Err Tia INTERESTIVALY, THE CNOT GATE CAN BE USED 70 CREATE ENTANGLED auBITS 61ENTANALEMENT- A RECAP " 2 2 Oo M$ ERVAL CORRELATION MATCHING EWD-STATES PRoreRTisé DF ENTANGLED Queits Mensuaina owe areécrs|pecives THE CAN ONLY BE DESCRIBED AC A STATE OF THE OTHER EVEN IF toe ~ aor inpivrouaccy THE TWO ALE LIGHT VERBS pPaer / \ ‘@ g ~p mo S OPPOSITE CORRELATION OPPOSITE END— STATESGATES FOR ENTANALEMENT enor : —_ [ 10> /oo> miria, STATE Havamaey ow 10> =, 1 Joy + Ld sg & mre 19) ley = L /o0y + + 109 he a Comemed STATE = ewoT OW 1 Joo ++ [10> , t/1, Zh # flo + el) 6aENTANALEMENT IS STRANAE TWO QUBITS ARE ENTANKLED WHEN THEY ARE IN AM EBVAL SuPER POSITION OF 100y AWD /11 (00> +> measure owe | THIS DECIDES aveir ann THE MEASURE oF INSTANTLY THE OTHER QuBIT THERE ARE 4 STATES OF ENTANGLEMENT CALLED BELL STATES ~ AFTER Jony $- BELL THE STRAWKE IMPLICATIOW OF A BELL STATE Take 1 [DQ>+L/1/>; ProBasiciTy OF Fest QUBIT =) Is 507 z ae JF THE FIRST QUBIT =0, PROBABILITY OF SEcoND QUBIT = D Is to0/- 70QVANTUM ALA0RITHMS — A Couceprual mopEL—QUANTUM ALGORITHM - RECAP $ $e of $$ 4 tt + seas a gag Ta & AVANTUM ALADRITHM USES A GLOBAL’ PROPERTY DF THE VALUES IN THE QUBITS TO CHANGE THEIR AMPLITUDE. REDUCE AMPLITUDE DF WROWA ANSWERS GOAL INCREASE AMPLITUDE DF CORRECT ANSWERS SO, WHEN A MEASUREMENT Is MADE, THERE Is A HIGH PROBABILITY THAT THE ANSWER 15 CORRECT nmHOW D0 QUANTUM ALGORITHMS WORK? In CONTRAST TO CLASSICAL ALADRITHMS, THERE ARE OMLY A HANDFUL OF QUANTUM AC AORITHMS. THEY APPEAR 10 BE CAPABLE DF SOLVING HARDER PROBLEMS FASTER THESE ALGORITHMS ARE COMPOSED AWD EXECUTED VERY DIFFERENTLY FROM WHAT WE ARE USED TO IN CLASSICAL COMPUTERS QUANTUM ALAORITHMS WORK BETTER ON PROBLEMS WITH A CERTAIN STRUCTURE To pRovibe SPEEDUP THE FIRST SVCH Kf THE BERNSTEIA- VAZIRAN) ALAOLITHM To UNDERCTAND THEM 4 Bir BETTER, We wil LOOk & BIT CLOSELY AT THE WoRKIAS DF THE? @ > DevtecH- 20288 acaverHm GROvER's ALADRITHM SHoR's AlgoeitHm Exercise » cookuP quantumalgorithm 200- org 23DJEVTSCH-J02SA ALKORITHMDeuTscH's peoatert AM 1 CONSTANT OR bacancen? THE PROBLEM <= WHO AM 1 2 Tam a function $C) fo fix) +0 Conerant My input (ta 0 or 4 ee f@) 2% EQUALLY Bacawces wor (a) ouTPur @ fH) My output ts @ 0 or 4 fm constant or balanced ? @ THE FAMILIAR CLASSICAL APPROACH Davi> DEUTSCH - 19905 THE QUANTUM WAY @ oe me eres FOR THE VERY FIST fm I Constant ? u AvanTum ALaveTAM am I Balanced? 1985 wirn GOOD You may ask me a MATHEMATICAL ene ABSTRACTION. ee emoree DAVID DEUTSCH? (burch Torry OxFDRDDEVTSCH-J02SA ALGORITHM: CORE x: [oy 4 pe te ma __| function A QUANTUM CikcuIT READS FROM LEFT TO RIAMT mpur avert x = 0 wD THROUAH Tar Ser THEM Iw HevpeR auBiT y = | HADAMARD ATES EQUAL SUPER POSITION AW ORACLE FUNCTION F RETURNS 07 comsrawr I> Bacanced A sivate measurement [7X] ar THE EWD REvEacs 1 F Is comsranT / BALANCED 6IMPLEMENTING THE ORACLE Tuere nee 4 oprions OS, f=) 1, FRY, FD: ore LETS SAY THE Function OUTPUTS /x> AWD /YBfl~)> where B= XoR v I) [yepe) mn <=” ’ oO t os sy ele|~|- (borer: |x? 14> ob lyepey que eee 1 0 0 ! | yepr) acts We (ROT 1 1 0 0 case @ fa)=1 i or) 0 aad sjij 0 Iy@¢e»—{x}— 0 0 v riol ourpuT’= x> wor [yy mi eer ol tol 0.0 10 1 1 lyase ° 0p) « IE 1”DEVTSCH- JOZSA: DETAIL -/ x: [oy 4 FE sanction AFTER 2 HADAMARD GATES a L ([0>+]>) @ L (Joy [> ny > £(lo>+/0) 8-4. (10-12) uy Ft [1009+ /i1> [01 - Jn] —> ae sinte WHEN THEY PASS THROUGH F, THE ovrPUT Is x, YO f(x) 4[r vege» + /100f00 ~ Joregmy ~ [1 104009] 7DEVTSCH- JOZSA: DETAIL -2 ELlooesio)> + 1 oa f6)>- /0 19 £00)>- [1 1afery) toustanT ft0):f (+0 © [009+ [10> - for “i Pall ont common terms © J00(lo- 10) + fp (t>-Iny 2nd gp bit e +(0>+ [1>)(10~ In) ©@ fo0> + [ny - on ~ /w> Pall ont common berms © "[o9(10>-/12) + [0(o~ 109) © CDI» + by) (, “03 I12) and gptit CONSTANT flo): fO: I Jo1>+ /11) - ]00>- 10> Pull out common terms jo» (> -[02)+ Ip (is -19) nd. gprvik = (»+19) (10) ©@ for + [10>~[00>- Jn) Pull ont common terms 0% (2 ~]02) + [1p (Jor ~ |v) ~~ ind git @ -(/0+/2) (10> -)) * 4DEV TSCH- J0ZSA: MEASURE x: [0 4 fe fan Ha fanction 2 Lloowsey> + /1 096) - Jo 18 £00)>- [1 1afeoy] Becomes "Fok CONSTANT FUNCTIONS "Fok BALANCED FUNCTIONS L (10> + 11>) @ LC 10> - 11) 1 (l0>~- 112)@ (t) (10-12) 4 Ba A ve @ aver 4 ww state /+> @ aver 4 6 ow state /-> @ alts aes Jo> @ H/-> Awes JI> MEASURE 70 GET 0 consrawr 1 bALancED BODEVTSCH-J02SA: ANOTHER VIEW EFFECTIVELY, THE FIRST QUBIT IS MULTIPLIED BY Ente MEASUREMENT RETURAIS 0 MEASUREMENT RETURNS 1 16 flo) #01) :[eoMsTANTD, e (0)4 ¢¢) : (BROAMEED) F/4> = Cains 4)0) F/+> = i Cioy 4/0) ve ve #0) fo . #0) ft 9 $ (1) ]0) #61) (> $ (1) ]0) ¢(-1) 17> 2 aac Hl+>= Jo> HIy = ID aIGROVER'S ALGORITHM FOR UNSORTED SEARCHGROVER'S ALGORITHM THE PROBLEM: FIND THE WEEDLE THE FAMILIAR CLASSICAL APPROACH LOOKING FOR A NEEDLE AMONG N UNORDERED ITems REQUIRES Nyt TRIES AT gest N oTRIES AT Most Geover’s ALaoRITHM THE QUANTUM WAY ACHIEVES QUANTUM SPEED UP © wenriey oOo THe ANSWER Finding ‘NEEDLES’ vu iv omy ai tees © Increase O veceense Silent 1s ALC oo A MILLION THINGS an AmpuTUDES Lov K GROVER OMtY 1000 Times! ‘BELL LABS 3QROVER-| > WITIAL STEP cs) +n QUBITS IN EQUAL SUPERPOSITION REPRELENT 2"=—N ITEMS EACH WITH AMPLITUDE '/UNT START THE SEARCH FO AVERIES THE QueiTs AND IT RETURNS 1-1 THE NEEDLE 15 FOUND O- 1 wor FounD 84GROVER — 2: INVERT PHASE INVERT THE PHASE amplitude THE FUNCTION THEN FLIPS THE PHASE oF THE NEEDLE BY muLTIPLYING EACH qnbits aver ay cnt? MEASURE NOW? A MEASUREMENT WON'T HELP AT THIS STAKE, BECAUSE THE SUPERPOSITION WILL COLLAPSE TO A RANDOM STATE WITH A mosagiLiTY Y/N 85GROVER —~3: AMPLIFY ANSWER AMPLIFY THE ANSWER FIND THE AVERAGE OF Au THE AMPLITUDES THEN = ROTATE THE AMPLITUDES ABOUT THE AVERAGE THIS MEAWS THAT THEY ARE JUST AS FAR AWAY FROM THE AVERAKE BUT ow THE OTHER! SIDE 6GROVER-4% : FIND NEEDLE FIND THE NEEDLE REPEAT ENOUGH TIMES —AsouT Vat Times — UNTIL THE AMPLITUDE DF THE WeEEDLE IS HIGH SO WHEN A MEASUREMENT 1s MADE THE MOST LIKELY ouTcOME Is THE ‘wEEDLE! a7SHOR'S ALAORITHM FOR FACTORING — THE MOST FAMOUS QYANTUM ACADEITHM —SHOR'S ALKORITHM THE PROBLEM : FIND THE FACTORS THE CLASSICAL APPROACH TAKES EXPONENTIAL TIME IN A REASONABLE AMOUNT DF TIME Coven the special number fret sieve) THE QUANTUM WAY - FINDS FACTORS IN POLYNOMIAL TIME eee PETER SHOR pene cabs MITTHE CORE DOF SHOR'S ALGORITHM efind x WwW = 2f i @ = = 1 mop 2h 7 . 4 = 1 mop 21 ~ x= 64 7 ex-8 @ 4-/ DivinEes N @ (ar!) (8-) dwwes w (a) (1) ovaes 21 w= IS a @) Facroas OF 21 Aa 3,7 OR = 1 mon © toaves no lb = 1 00 & FEMAINDER | po eien a= Ib ox-s su lerply @ -! dives 7 i ouk er) C4-) dwies w Perfectly (5)(3) ones ir tof @ Facrons OF IS ARE 5,3 check 2 Utd trivial 90HOW TO FIND X CHART TH A avESS Really? YES 21ck A Ramone womeER wee RAISE IT TO EVERY POWER Lo zg22 s2l WRITE DOWN REMAINDERS 248 be. abs imop 21 LUCKY JF THE POWER is even so tf has « square root! Thak is, x98 1 see repeating sets mee “mon 21 2 2 2 2 4 4 a a’ 16 te a 32 Looks Ike =e — as poop W v.28 2 a 286 + Looks like 7 these numbers a_| $2 8 repeat 2? 102616 IN A REPEATING eT? 6 numbers How many SAME AS THE POWER DF THE GUESS I THE 2% 10d Row? Yes.o coincidence aFACTDRISATION CONNECTION TO A AvAWTUM VERSION DF THE FOURIER TRANSFORM IS APPLIED TO THE iwPuT AND OUTPUT QUBITS IW $HOR'S ALGORITHM SO wHen THE QubiTs ARE mensueeD/ SAMPLED FOR FREQUENCIES, WE CAN CALCULATE THE PERIOD WiTH A K00D PROBABILITY. FOURIER TRANSFORM A FOURIER TRAWSFORM Is A MATHEMATICAC PROCESS APPLIED 70 a wave [Fucrion /siqnac THIS IDENTIFIES THE wDIVIDVAL compowenTs & EACH FREQUENCY THE PERIOD OF A WAVE 16 THE INVERSE DF ITS FRERVENCY Taare B0HN = COOLE ‘ack 7 19608 fowrrer Donn TuKeY ae fam von coppsesmirn M06 tan 20008 femrrer eanagorm "2PERIOD FINDING : STEP -{ Powers Mod 2! 1 2 2 FOTIA a ee 2 2 . Z8) a gn eo} > 2 2 #4 ate nears 123 4F6 78 FOUveRBHeE wT wl, FTIR a 2 I Bog THE INPUT AND OUTPUT REGISTERS ARE ENTANGLED wa 4 218 @ measure THE ovrpuT Reausrer a 2-16 G . 2 y IT COLLAPSES TO A RANDOM VALVE - SAY & 18 ats 93PERIOD FINDING : STEP-2 Powers Mod2} 1 e ab gn ; | 7 Pier (@rtse 7 Or eres granu eS 2 & OL THE INPUT REGISTER HAS ALSO NOW COLLAPSED a a To ALL POWERS THAT LEAVE A REMAINDER 4 2 ; @ tHe PATTERN IS NOW: +] X, ht p, te ape 28 ot 6 WHERE p= PeRiod OF THE FuNcTIOW PY a a, (IDEALLY, WE WANT IT 70 BE OF THE FORM Pr 2py 3p.-- %PERIOD FINDING & FACTORING SHIFT THE FlgsT WoW-2E2D VALUE To 2620 nares: THE PHASE > Does WOT ALTER: = ourror on ene = Peei0d TTT a TD MAKE MULTIPLE MEASUREMENTS : 12,18, ... © TARE THE GREATEST COMMON DIVISOR OF THESE 70 AET THE PERIOD = 6 . neal check curssicacey® werner (gee!) duroes we 2t 2) 1s 7 awa fe) 2%. BOTH SHARE FACTORS WITH 21. @ IF wor, CHANGE vaLUEs AND TRY ARAN! an instance compulers working. hgetber I % fe)$0, AVANTUM COMPUTERS CAN SOLVE EVERYTHING 2THOSE PROBLEMS - SOLVABLE ? NP-COMPLETE PROBLEMS (THE MAXIMALLY HARD PROBLEMS IN THE CLASS WP) ARE BELIEVED 70 BE HARD 70 Solve. HOWEVER, AEN A SOLUTION, IT IS EASY To Heck THAT IT Is CoRRECT. SOME EXAMPLES BELOW: (a+! on bso oR c*1) Any Cbs! on 620 of dss) ano cLiave Peosiem (+108 do 02 e= 1) ano.. } IN MY SOCIAL NETWORK IS THERE # SoLUTION To CaM | FIND THE CARAGST LIST DF PEOPLE SATISFY THIS BOOLEAN FORMULA? WHO Ace kwow EACH oTHER? Ge -2 3 ‘Suaser sum peoavem 1s THERE A (Won EMPTY) SvBCET WITH Sums 02 Courd Ate THE PERmurariows/Comaiarios /oPTIOWS BE SETUP INTO ONE MIGHTY SUPERPOSITION FOR A RUANTUM COMPUTER TD CHECK? "7WHAT A QUANTUM COMPUTER CANNOT DO EASY 7D THivk OF THE POWER OF A QUANTUM COMPUTER AS A ALL POSSIBILITIES AT DANCE IT FEELS GAILLION PARALLEL COMPUTERS EXPLORIVA TO FIND THE CORRECT ANSWER EXPOIVENTIALLY FASTER. IN TRUTH! CHECKING ALL PossiaLe A RVANTUM COMPUTER “PARALLEL” STATES ALSO NEEDS A FEW 15 NOT POSSIBLE IW A TRIES BEFORE IT ARRIVES QUANTUM COMPUTER AT AN ANSWER CrdQUANTUM ALGORITHM EFFICIENCY CHOOSING THE RIGHT PROBLEM ‘eet o QUANTUM PARALLELISM (MANIPULATING Peosasiiries EFFICIENCY COMES FROM SUPER POSITION ENTANALEMENT Vihe RRS as a Dr ae ee) 9BAP -A NEW COMPLEXITY CLASS A WEW COMPLEXITY CLASS FoR A RVANTUM woRLD BOUNDED ERROR QVANTUM POLYNOMIAL TIME Be ETHAN BERNSTEIN) UMESH VAziRAMvT PROBLEMS YES/NO DECISION SoLveD BY : AVANTUM COMPUTERS TIME = PoLYWoMIAL. eRRoRs ¢ WITHIN A CERTAIN LIMIT runes CT? BRP EDNTRINS P BUT WHAT WE DON'T YET kNOW. © 18 WP conrained IW BaP 2 © CAN WE SOLVE EVEN MDRE PROBLEMS? 100SECURITY iw THE TIME OF QuANTUMQUANTUM CAN 8REAK CRYPTO 17 SEEMS ONLY A MATTER OF How to fachr Time THAT 204 (EEA) integers wile BREAK ALL tat han ENCRYPTION SYSTEMS THAT HOLD OUR wing 2 million ONLINE WFORMATION SECURE noiky yabits Gracy Gidney, Martin Ekere 20/7 5 THE TOP PROGRAMMABLE Quantum CompuTer @ HAS 53° QUBITS QUANTUM COMPUTERS Applying Srover's Algorithr Bs ONLINE SECURITY : THIIVAS TO LOOK our FOR gnantem resource echnalte ~ QUANTUM-SAFE CRYPTOSYSTEMS: tok = — ENOUGH ERROR-CoReecTED QUEITS ingen berg, boetteler, Shetn wand 102QUANTUM-SAFE CRYPTO CURRENT CRYPTO UNDER THREAT NEW RESEARCH @ ts @ FoR A post- quantum woecy DIF FIE~ HELLMAN eucreric curve DIGITAL SIGNATURES AND SECURITY PROTOCOLS) PRODUCTS BASED oN THEM / ——~ ceayeros ystems —— KNOWN SUANTUM- SAFE CRYPTO QUANTUM KEY DISTRIBUTION - akD eieon Merete entree A Why 70 SECURELY semp ES REVS © ONE-TME pap @ @... rm 9, ‘Nor COMPLETE 1s -coerey SOME HASH FuNcTIONS REFER: ETSI-DRG QUANTUM SAFE WHITE PAPER, REFER! WeSC. GOV. UR KD WHITEPAPER 103AQKD & POLARISATION AVANTUM KEY DISTRIBUTION WORKS WITH POLARISED LIAHT —— HERE If A BRIEF woTE ow POLARISATION — PHOTONS VIBRATE DV PLAIVES PERPENDICULAR 10 THE DIRECTIOW oF LIAHT | yz 4 RECTILINEAR FILTER UNPOLARISED LIGHT DIAGONAL FILTER A PoLARIsING FILTER (LIE SUNMALASSES) CHOOSES A PLANE LETTING DNLY THOSE PHOTONS PASS THROVAH- MAKING POLARISED LIGHT 10%QUANTUM KEY DISTRIBUTION - QKD auce sews 5 \ 208 MOP) oumeised uisur < must avess 2 ¥ WHICH ENCODES BITS 1 oO WHICH METHOD * rawoomey ao e IN ONE 5 ‘READ’ we ge | wo | me morn = . + mermons cath AN envesproprer, BURLEY <— Public Channe/—> (B03) *% Vv compare uw wit EMD uP ia ' me ay. | * hm x fw mann | oe toe 4 Ve K FROM THE METHODS + x THE POLARISATION emery v+ tv 7 AET KEY, convERT 70 pues & ZEROES 105GETTING THE SECRET KEY WITH QKD WRONG FILTER DR AN EAVESDROPPER IDENTIFY DISCREPANCIES (A / CAUSE ERRORS IN THE READING ONLY COMPARE THE FILTERS CwoT BiTs) AND CAN BE IDENTIFIED OVER A PusLiC CHANNEL, USE MATCHING READINGS HAS EVE BEEN EAVESDROPPIIVG ? / e = Compare a subset of the strings ri = @ IF THEY MATCH : ¥# Wo Eve 9 10/01110110101 6 / sioloMono10r, use the other subset as key TRANSLATE THEM To GET IDENTICAL @ THEY DowT maTcH : #Eve PRIVATELY STRINGS OF Os AND Is Stark over BASED OM THE BRay SENNETT AND BRASSARD 106ERROR CORRECTION —T0 MALE QUBIT OPELATIONS MORE AcCULATE — 107CLASSICAL ERROR CORRECTION ORIAIVAL DATA ERROR CORRECTION APPLIED How many owes YAY 76. PER ROW /cocumn ? Wwe | > a ont g = eS Sp mene me 20 a DATA Ce cD THE ERROR CORRECTION CODES? ERRORS CAN BE SPOTTED _2ATA enzors AND FIXED > Row 1- A 8 Row 2~- U con 1- % fe 2- yw THIS TYPE OF ERROR CORRECTION IS INTEGRAL TO HARDWARE DESIGN THAT WE NOW TAKE FOR ARANTEDQUANTUM ERRDR e@ QUBITS ARE PROVE TO ERRORS puna comurarion — FROM DECOHERENCE ey BACKUPS = () MEASUREMENT LOSES SUPERPOSITION @ 'o-ceowiva’ THEOREM sowvrion | APPROACH MAKE A LORICAL QUBIT © SPREAD DATA ACROSS ENTANALED QUBITS © ADD ANCILLA (HELPER) AUBITS © DETECT & CORRECT ERRORS BY READIVA THE ANCILLA QVBITS CORRECTION (GEC) Merwe, reren swor came ve wire : THe FIRST THEORY OF e QUANTUM ERROR CORRECTION USING % QUBITS TO ERROR coeect 1 qusiT oF DATA PETER SHOR 195 Bei Uaas/mir “we mee wv 4 CB NOISY INTERMEDIATE SCALE QUANTUM ERA (50-100 ausiT compurers) ‘A STEP TOWARD MORE POWERFUL QUANTUM TECHNOLOGIES OF THE FUTURE” 20MN) PRESEIOL CALTECH 2017 mePROGRAMMING & TESTING — THIS Is ANOTHER AREA OF OWAOING RESEARCH —PROGRAMMING IT If A CLASSICAL COMPUTER THAT ISSUES COMMANDS 70 A QUANTUM COMPUTER To EXECUTE VARIOUS AATES. OBVIOUSCY , THIE IS LOW LEVEL InsTRUCTION. Also, MOST INSTRUCTION SETS OFFERED BY vaRiour DEVELOPMENT ITS IN THE MARKET ARE TIED TD THE DEVICE. WHAT D0ES IT TAKE TO MAKE a HIGH LEVEL QUANTUM PROARAMMING LANAUAAE 2 HERE ARE SOME OF THE CONCIDERATIONS O weteucrions ARE HARDWARE IWDEPENDENT ¥ COMPILE © oPEWsouRcED FOR A COMMUNITY TO DE @ h00D womBer OF ABSTRACT FUNCTIONS tg ROUTINES FOR ERROR CORRECTION MEASUREMENTS, FOURIER TRAWs FORMS VELOP nmTESTING Inberactive THE REGULTS OF A QUANTUM COMPUTATION ARE HARD paeenioe To PREDICT AS THEY EVOLVE WITH THE CAWS OF aa tin S THEY EVOLVE wi et apr 4017 QUANTUM MecHaics. HOW THEN D0 WE MEAGRE 2 Abaronov THE Success OF AN EnpeaimenT/ALaoRITHM ? 1 ben-or & ban U tmahadey we USE INTERACTIVE PROOFS To QUERY THE ALADEITIM wy A STRUCTURED MANNER. that IT WORKS A BIT LIKE THE ConvERSATION BELOW: @snows wHermen & QUANTUM COMPUTER DOES WHAT IT SUPPOSED TO @estasusnes ravsr iw A SITUATION THAT ALSO int wloure Reavieets Privacy > hn iene vy tame colour weTHE PRESENT AND FUTURE A UITTLE GLIMPSE — "2CURRENT STATE AUANTUM SUPREMACY "WHAT A AVANTUM COMPUTER cay DO TO MAKE A CALCULATION FasTER, (by ® few thousand years!) THAN A CLASSICAL ComPuTER” = DOHNY PRESEILL EXAM year: 2019 Im A004LE ROCESSOR —: SYCAMORE, 53 QUBITS TASK CHECk PATTERN OF A LONG SeRvENCE oF DIAITS DURATION + AROUND 5 MINUTES $0? : To DO THIS VERY Task SUPERCOMPUTERS TARE Agour 10,000 years! DupureD BY: 18M Lee Stott Aavoncons arkele on this: in nytimes.com OF AFFAIRS COMPANIES INVESTING IN RESEARCH HODEe pweBw one more wn Be ves AND MANY THE POTENTIAL OF QUANTUM COMPUTERS AND THEIR APPLICATION IS YET TO BE unpeesroon Fuuey (AMID ALL THE DEBATE) "WePOTENTIAL USES OF QUANTUM COMPUTERS ~ & compurers caw Do? QENERATIVG RVAITUM TELEPORTATIOAI Now THAT quANTUM / et apy RUE RANDOM NUMBERS ubWOMEN IN QUANTUM COMPUTING — BECAUSE We NEED TD KMOW, SHARE AWD HONOUR THEIR ACHIEVEMENTS — ANd WHO xavows, Ir MAY ~IWEPIRE ANOTHER ! WwWOMEN (IN QUANTUM COMPUTING DORIT AHARONOV MAJOR CONTRIBUTOR 70 FOUNDATIONS DF AVANTUM COMPUTATION THEORY — ewravatement = tompurarionaL eonnibeiry [qvanruet aukorirums]inreeacrve moore: TeV4 = AVANTUM CRYPTORRAPHY = noisy systems [skeons) FAULT ToLeeance TALIA AERSHON DESIANER OF THE SPEAKER AND LEAD FOR HanpwAne|cHiP FoR AUANITUM — com PUTING = THE FiRsT 2-QuaIT 4ATE QUANTUM eoMPLTER EEE ¢* srevwunns Wenner oN- cLovD TEAM QUANTUM CRYPTORRAPHER, HACKER, AND BUILDER OF AVANTOM iWTERNET RESEARCHER OF QUANTUM ERROR CORRECTION AND FAULT TOLERANCE 26 5@ E ieWOMEN IN QUANTUM COMPUTING NTT cezecin_woseuin wneisca aiverinn ‘A FIND-OUTER OF WHETHER (MAKER OF POST- QVANTUM IWSTRUCTOR AND QUANTUM er ean emee LATTICE BAGED~ CRYPTO COMPUTER BUILDER ANNTHINA BYANTUM AT ALL ELEANOR RIEFFEL £8 QUANTUM ComPuTINNG DEVELOPER OF QUANTUM- INSPIRED MATHEMATICIAN AIVD AUTHOR RESEARCHER) SPER EER Se een OF QUANTUM ComPuTiNG: A oeiTH RECOMMENDATION ALADEITHM fewrie lrreoouerion TESSICA PoINvTINVA nyWOMEN (N QUANTUM COMPUTING (ECRARRATHEFTI) cen) aiamanrs a+e (O-CREATOR OF QUANTUM COMMUNICATION HIGHER DIMENSION gugITS @UANTUM PeoTocoL 200 RESEARCHER ‘TRIPLE -SLIT EXPERIMENTER TawisHAa Bassa SHELBY KIMMEL S iu DESIGNER DF A NEW SUPER compucTiNg qugiT AVANTUM MACHINE RESEARCHER OF ALADBITHM CEARIING RESEARCHER EFFICIENCY, ERROR ANALYSIS 120MY REFERENCES Books Quantum: A guide for the perplexed - Jim Al-Khalil The Quantum World - New Scientist Quantum Computing Since Democritus - Scott Aaronson Online Courses Quantum Mechanics & Quantum Computing - Umesh Vazirani, UC Berkeley, YouTube The Introduction to Quantum Computing - St Petersburg State University, Coursera Quantum Computing Tutorial - Travis Gritter, YouTube Understanding Quantum Computers - Future Learn Blogs / Websites medium.com - Jonathan Hui ibm.com - What is QC? technologyreview.com - MIT, Explainer Online Talks A Beginner's Guide - Talia Gershon What Quantum Computing isn't - Scott Aaronson, TED On Complexity and Quantum - Scott Aaronson, YCombinator ‘On Quantum Computation - John Preskill, YCombinator QC and Information - Scott Aaronson, Centre for Science of Information Quantum Computing Reality vs Hype - John Preskill, Caltech QC and the Entanglement Frontier - John Preskill, UC Berkeley Einstein Lecture QC Revolution - Michelle Simmons, UNSW Feynman's Opportunity - Christopher Munroe, Caltech Crazy History of Quantum Mechanics - Leonard Mlodinow, TEDMY REFERENCES Complexity Theory P vs NP - MIT, Explainer Knowledge, Creativity and P vs NP - Avi Wigderson, math ias.edu P vs NP, Efficient Computation - Avi Wigderson, math.ias.edu The P vs NP problen - Michael Sipser Lectures Part 1 and 2 - Dorit Aharonov, Dept of Physics Technion Computational Complexity - Erik Demaine, MIT Open Course Ware Church Turing Thesis (CTT) CTT: Logical Limit or Breachable Barrier - ACM.org, Jack Copeland, Oran Shagrir Programming Quantum Computing Languages Landscape QWA, Alba Cervera-Lierta on medium.com Testing Testing Quantum Devices - Umesh Vazirani, MSRI Math Lovers Forum Panel Discussions QC: Far Away? Around the Corner? Both? - ACM.org Beyond Computation P vs NP - Simons InstituteFOR OTHER TECH STUFF More Technology Resources For other illustrated guides and technology and software related resources, please visit thoughtworks.com/insights Thank you for renderg |