We give an asymptotic evaluation of the complexity of spherical p-spin spin-glass models via random matrix theory. This study enables us to obtain detailed information about the bottom of the energy landscape, including the absolute minimum (the ground state), the other local minima, and describe an interesting layered structure of the low critical values for the Hamiltonians of these models. We also show that our approach allows us to compute the related TAP-complexity and extend the results known in the physics literature. As an independent tool, we prove a LDP for the k-th largest eigenvalue of the GOE, extending the results of Ben Arous, Dembo and Guionnett (2001).
Circuit complexity lower bounds for quantum spin glasses
Broad Institute of MIT and Harvard(麻省理工学院和哈佛学院博德研究所);Operations Research Center and Center for Statistics and Data Sciences, IDSS, Massachusetts Institute of Technology(麻省理工学院运筹学中心与统计与数据科学中心IDSS)
Compression, inversion, and approximate PCA of dense kernel matrices at near-linear computational complexity
从位置$\{ x_{i} \}_{1 \leq i \leq N} \subset \mathbb{R}^{d}$处的协方差函数$G$的点评估获得的稠密核矩阵$Θ\in \mathbb{R}^{N \times N}$出现在统计、机器学习和数值分析中。对于协方差函数是格林函数的椭圆边值问题和离散分布的采样点,我们展示了如何确定一个子集$S \subset \{ 1 , \dots , N \}^2$,与$\# S = O ( N \log (N) \log^{d} ( N /ε) )$,使零填充不完全Cholesky分解的稀疏矩阵$Θ_{ij} 1_{( i, j ) \in S}$是$ε$ -近似的$Θ$。这种因式分解可以证明在空间和$O ( N \log^{2}( N ) \log^{2d}( N /ε) )$时间的复杂性$O ( N \log( N ) \log^{d}( N /ε) )$,提高了现有技术的一般椭圆算子,我们进一步提出的数值证据表明,$d$可以被视为是固有的尺寸的数据集,而不是周围的空间。该算法只需要知道的$x_{i}$的空间配置,并不需要$G$的解析表示。此外,这种因式分解直接提供了一种近似稀疏PCA,其在算子范数中具有最佳收敛速度。因此,通过仅使用子采样和不完全Cholesky分解,我们得到,在接近线性的复杂性,压缩,反演和近似PCA的一大类协方差矩阵。通…