Tensor cumulants for statistical inference on invariant distributions

• URL: http://arxiv.org/abs/2404.18735v1
• Date: Mon, 29 Apr 2024 14:33:24 GMT
• Title: Tensor cumulants for statistical inference on invariant distributions
• Authors: Dmitriy Kunisky, Cristopher Moore, Alexander S. Wein,
• Abstract summary: We show that PCA becomes computationally hard at a critical value of the signal's magnitude. We define a new set of objects, which provide an explicit, near-orthogonal basis for invariants of a given degree. It also lets us analyze a new problem of distinguishing between different ensembles.
• Score: 49.80012009682584
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Many problems in high-dimensional statistics appear to have a statistical-computational gap: a range of values of the signal-to-noise ratio where inference is information-theoretically possible, but (conjecturally) computationally intractable. A canonical such problem is Tensor PCA, where we observe a tensor $Y$ consisting of a rank-one signal plus Gaussian noise. Multiple lines of work suggest that Tensor PCA becomes computationally hard at a critical value of the signal's magnitude. In particular, below this transition, no low-degree polynomial algorithm can detect the signal with high probability; conversely, various spectral algorithms are known to succeed above this transition. We unify and extend this work by considering tensor networks, orthogonally invariant polynomials where multiple copies of $Y$ are "contracted" to produce scalars, vectors, matrices, or other tensors. We define a new set of objects, tensor cumulants, which provide an explicit, near-orthogonal basis for invariant polynomials of a given degree. This basis lets us unify and strengthen previous results on low-degree hardness, giving a combinatorial explanation of the hardness transition and of a continuum of subexponential-time algorithms that work below it, and proving tight lower bounds against low-degree polynomials for recovering rather than just detecting the signal. It also lets us analyze a new problem of distinguishing between different tensor ensembles, such as Wigner and Wishart tensors, establishing a sharp computational threshold and giving evidence of a new statistical-computational gap in the Central Limit Theorem for random tensors. Finally, we believe these cumulants are valuable mathematical objects in their own right: they generalize the free cumulants of free probability theory from matrices to tensors, and share many of their properties, including additivity under additive free convolution.

Related papers

• Ito Diffusion Approximation of Universal Ito Chains for Sampling, Optimization and Boosting [64.0722630873758]
  We consider rather general and broad class of Markov chains, Ito chains, that look like Euler-Maryama discretization of some Differential Equation. We prove the bound in $W_2$-distance between the laws of our Ito chain and differential equation.
  arXiv  Detail & Related papers  (2023-10-09T18:38:56Z)
• Average-Case Complexity of Tensor Decomposition for Low-Degree Polynomials [93.59919600451487]
  "Statistical-computational gaps" occur in many statistical inference tasks. We consider a model for random order-3 decomposition where one component is slightly larger in norm than the rest. We show that tensor entries can accurately estimate the largest component when $ll n3/2$ but fail to do so when $rgg n3/2$.
  arXiv  Detail & Related papers  (2022-11-10T00:40:37Z)
• Lower Bounds for the Convergence of Tensor Power Iteration on Random Overcomplete Models [3.7565501074323224]
  We show that tensorly many steps are necessary for convergence of tensor power iteration to any true component. We prove that a popular objective function for tensor decomposition is strictly increasing along the power iteration path.
  arXiv  Detail & Related papers  (2022-11-07T19:23:51Z)
• Lattice-Based Methods Surpass Sum-of-Squares in Clustering [98.46302040220395]
  Clustering is a fundamental primitive in unsupervised learning. Recent work has established lower bounds against the class of low-degree methods. We show that, perhaps surprisingly, this particular clustering model textitdoes not exhibit a statistical-to-computational gap.
  arXiv  Detail & Related papers  (2021-12-07T18:50:17Z)
• Smooth tensor estimation with unknown permutations [14.391648046717073]
  We develop a family of smooth tensor models up to arbitrary index permutations. We show that a constrained least-squares estimator in the block-wise family achieves the minimax error bound. We also provide an efficient-time Borda count algorithm that achieves provably optimal rate.
  arXiv  Detail & Related papers  (2021-11-08T17:53:48Z)
• Scaling and Scalability: Provable Nonconvex Low-Rank Tensor Estimation from Incomplete Measurements [30.395874385570007]
  A fundamental task is to faithfully recover tensors from highly incomplete measurements. We develop an algorithm to directly recover the tensor factors in the Tucker decomposition. We show that it provably converges at a linear independent rate of the ground truth tensor for two canonical problems.
  arXiv  Detail & Related papers  (2021-04-29T17:44:49Z)
• Inference for Low-rank Tensors -- No Need to Debias [22.163281794187544]
  In this paper, we consider the statistical inference for several low-rank tensor models. For the rank-one PCA model, we establish the theory for inference on each individual singular tensor. Finally, simulations are presented to corroborate our theoretical discoveries.
  arXiv  Detail & Related papers  (2020-12-29T16:48:02Z)
• Conformal Properties of Hyperinvariant Tensor Networks [0.0]
  We analyze the challenges related to optimizing tensors in a hyMERA with respect to some quasiperiodic critical spin chain. We show two new sets of tensor decompositions which exhibit different properties from the original construction. We find that the constraints imposed on the spectra of local descending superoperators in hyMERA are compatible with the operator spectra of several minimial model CFTs.
  arXiv  Detail & Related papers  (2020-12-17T14:06:15Z)
• Spectral Learning on Matrices and Tensors [74.88243719463053]
  We show that tensor decomposition can pick up latent effects that are missed by matrix methods. We also outline computational techniques to design efficient tensor decomposition methods.
  arXiv  Detail & Related papers  (2020-04-16T22:53:00Z)
• On Linear Stochastic Approximation: Fine-grained Polyak-Ruppert and Non-Asymptotic Concentration [115.1954841020189]
  We study the inequality and non-asymptotic properties of approximation procedures with Polyak-Ruppert averaging. We prove a central limit theorem (CLT) for the averaged iterates with fixed step size and number of iterations going to infinity.
  arXiv  Detail & Related papers  (2020-04-09T17:54:18Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.