Table of Contents |
1. | Introduction | 1 |
1. | Superconcentration | 1 |
2. | Chaos | 7 |
3. | Multiple Valleys | 11 |
2. | Markov Semigroups | 15 |
1. | Semigroup Basics | 15 |
2. | The Ornstein-Uhlenbeck Semigroup | 17 |
3. | Connection with Malliavin Calculus | 18 |
4. | Poincare Inequalities | 18 |
5. | Some Applications of the Gaussian Poincare Inequality | 19 |
6. | Fourier Expansion | 20 |
3. | Superconcentration and Chaos | 23 |
1. | Definition of Superconcentration | 23 |
2. | Superconcentration and Noise-Sensitivity | 25 |
3. | Definition of Chaos | 25 |
4. | Equivalence of Superconcentration and Chaos | 27 |
5. | Some Applications of the Equivalence Theorem | 28 |
6. | Chaotic Nature of the First Eigenvector | 29 |
4. | Multiple Valleys | 33 |
1. | Chaos Implies Multiple Valleys: The General Idea | 33 |
2. | Multiple Valleys in Gaussian Polymers | 34 |
3. | Multiple Valleys in the SK Model | 36 |
4. | Multiple Peaks in Gaussian Fields | 38 |
5. | Multiple Peaks in the NK Fitness Landscape | 40 |
5. | Talagrand's Method for Proving Superconcentration | 45 |
1. | Hypercontractivity | 45 |
2. | Talagrand's L1-L2 Bound | 47 |
3. | Talagrand's Method Always Works for Monotone Functions | 49 |
4. | The Benjamini-Kalai-Schramm Trick | 51 |
5. | Superconcentration in Gaussian Polymers | 54 |
6. | Sharpness of the Logarithmic Improvement | 56 |
6. | The Spectral Method for Proving Superconcentration | 57 |
1. | Spectral Decomposition of the OU Semigroup | 58 |
2. | An Improved Poincare Inequality | 60 |
3. | Superconcentration in the SK Model | 60 |
7. | Independent Flips | 63 |
1. | The Independent Flips Semigroup | 63 |
2. | Hypercontractivity for Independent Flips | 65 |
3. | Chaos Under Independent Flips | 66 |
8. | Extremal Fields | 73 |
1. | Superconcentration in Extremal Fields | 73 |
2. | A Sufficient Condition for Extremality | 78 |
3. | Application to Spin Glasses | 79 |
4. | Application to the Discrete Gaussian Free Field | 83 |
9. | Further Applications of Hypercontractivity | 87 |
1. | Superconcentration of the Largest Eigenvalue | 87 |
2. | A Different Hypercontractive Tool | 90 |
3. | Superconcentration in Low Correlation Fields | 92 |
4. | Superconcentration in Subfields | 93 |
5. | Discrete Gaussian Free Field on a Torus | 94 |
6. | Gaussian Fields on Euclidean Spaces | 99 |
10. | The Interpolation Method for Proving Chaos | 105 |
1. | A General Theorem | 105 |
2. | Application to the Sherrington-Kirkpatrick Model | 111 |
3. | Sharpness of the Interpolation Method | 112 |
11. | Variance Lower Bounds | 115 |
1. | Some General Tools | 115 |
2. | Application to the Edwards-Anderson Model | 118 |
3. | Chaos in the Edwards-Anderson Model | 119 |
12. | Dimensions of Level Sets | 125 |
1. | Level Sets of Extremal Fields | 125 |
2. | Induced Dimension | 128 |
3. | Dimension of Near-Maximal Sets | 131 |
4. | Applications | 134 |
Appendix A | Gaussian Random Variables | 137 |
1. | Tail Bounds | 137 |
2. | Size of the Maximum | 137 |
3. | Integration by Parts | 139 |
4. | The Gaussian Concentration Inequality | 139 |
5. | Concentration of the Maximum | 140 |
Appendix B | Hypercontractivity | 143 |
| References | 147 |
| Author Index | 153 |
| Subject Index | 155 |