Title: Algebraic Aspects of Topological Quantum Computing
1Algebraic Aspects of Topological Quantum Computing
Eric RowellIndiana University
2Collaborators
Project Q with M. Freedman, A. Kitaev, K.
Walker and C. Nayak
3What is a Quantum Computer?
- Any system for computation based on
- quantum mechanical phenomena
Create Manipulate
Measure
Quantum Systems
4Classical vs. Quantum
1
a
- Bits 0,1
- Logical Operations
- on 0,1n
- Deterministic output unique
- Qubits VC2
- (superposition)
- Unitary Operations
- on V n
- Probabilistic output varies
- (Uncertainty principle)
1
0
0
X
5Anyons 2D Electron Gas
1011 electrons/cm2
particle exchange
9 mK
fusion
defectsquasi-particles
10 Tesla
6Topological Computation
Computation
Physics
output
measure
apply operators
braid
create particles
initialize
7Algebraic Characterization
Anyonic System
Top. Quantum Computer
Modular Categories
8Toy Model Rep(G)
- Irreps V1C, V2,,Vk
- Sum V W, product V W, duals W
- Semisimple every W miVi
- Rep Sn EndG(V n)
X
X
9Braid Group Bn Quantum Sn
Generated by
i1,,n-1
bi
Multiplication is by concatenation
10Concept Modular Category
deform
group G
Rep(G)
Modular Category
Sn action
Bn action
Axiomatic definition due to Turaev
11Modular Category
- Simple objects X0C,X1,XM-1
- Rep(G) properties
- Rep. Bn End(X n) (braid group action)
- Non-degeneracy S-matrix invertible
X
12Dictionary MCs vs. TQCs
X
13Constructions of MCs
quantum group
semisimplify
g Uqg Rep(Uqg)
F
Lie algebra
q1
- Survey (E.R. Contemp. Math.) (to appear)
Also,
quantum double
G D(G) Rep(D(G))
finite group
13
14 Physical Feasibility
Realizable TQC
Bn action Unitary
Uqg Unitarity results (Wenzl 98), (Xu 98)
(E.R. 05)
15Computational Power
Physically realizable Ui universal if all
Ui all unitaries
TQC universal
F(Bn) dense in PkSU(k)
Results in (Freedman, Larsen, Wang 02) and
(Larsen, E.R., Wang 05)
Physical Hurdle Realizable as Anyonic Systems?
16General Problem
- Rep. Y G U(N), G discrete
- Y(G) ?
- SU(N)
- Finite group
- SO(N), E7
- For G Bn see (Franko, E.R., Wang) and (Larsen,
E.R.)
17Classify MCs
1-1
Recall distinct particle types
Simple objects in MC
Conjecture (Z. Wang 03) The set MCs of rank M
is finite.
Classified for M1, 2 (V. Ostrik), 3 and 4
(E.R., Stong, Wang)
True for finite groups! (Landau 1903)
18Groethendieck Semiring
- Assume XX. For a MC D
-
- Xi Xj Nijk Xk
- Semiring Gr(D)(Ob(D), , )
- Encoded in matrices (Ni)jk Nijk
X
X
19Modular Group
- Non-dengeneracy S symmetric
- Compatibility T diagonal
-
-
- give a unitary projective rep. of SL(2,Z)
-
T,
S
20Our Approach
- Study Gr(D) and reps. of SL(2,Z)
- Ocneanu Rigidity
- MCs Ni
- Verlinde Formula
- Ni determined by S-matrix
Finite-to-one
21Some Number Theory
- Let pi(x) det(Ni - xI)
- and K
Split(pi,Q). - Study Gal(K/Q) always abelian!
- Nijk integers, Sij algebraic, constraints
polynomials.
22Sketch of Proof (Mlt5)
- Show 1 Gal(K/Q) SM
- Use Gal(K/Q) constraints to determine
(S, Ni)
- For each S find T rep. of SL(2,Z)
- 4. Find realizations.
23Graphs of MCs
- Simple Xi multigraph Gi
- Vertices labeled by 0,,M-1
- Question What graphs possible?
Nijk edges
j
k
24Example (Lie type G2, q10-1)
- Rank 4 MC with fusion rules
- N111N113N123N222N233N3331
- N112N122 N2230
G1
Tensor Decomposable!
G2
0
2
1
3
G3
2
0
3
1
25Classification by Graphs
- Theorem (E.R., Stong, Wang)
- Indecomposable, self-dual MCs of ranklt5 are
classified by
26Future Directions
- Classification of all MCs
- Prove Wangs conjecture
- Images of Bn reps
- Connections to link/manifold invariants, Hopf
algebras, operator algebras
27Thanks!
28Bn Images vs. Link Invariants
- Roughly, we expect
- Finite image Link invariant classical
(easy). - Dense image Link invariant P-hard
- Universal TQC
29Improvements over Classical Computation
- Integer factorization in polyl time
- Simulation of Quantum Mechanics
- Quantum database search
30Analogy
- Theorem (E. Landau 1903)
- The set G Rep(G)N is finite.
- Proof Exercise (Hint Use class equation)
31More Graphs
Lie type B2 q9-1
D(S3)
Lie type B3 q12-1
Lie type G2 q21-1