no code implementations • 18 Mar 2024 • Hongjie Chen, Jingqiu Ding, Tommaso d'Orsi, Yiding Hua, Chih-Hung Liu, David Steurer
We develop the first pure node-differentially-private algorithms for learning stochastic block models and for graphon estimation with polynomial running time for any constant number of blocks.
no code implementations • 17 May 2023 • Jingqiu Ding, Tommaso d'Orsi, Yiding Hua, David Steurer
We study robust community detection in the context of node-corrupted stochastic block model, where an adversary can arbitrarily modify all the edges incident to a fraction of the $n$ vertices.
no code implementations • 14 Nov 2022 • Tommaso d'Orsi, Rajai Nasser, Gleb Novikov, David Steurer
Using a reduction from the planted clique problem, we provide evidence that the quasipolynomial time is likely to be necessary for sparse PCA with symmetric noise.
no code implementations • 16 Jun 2022 • Hongjie Chen, Tommaso d'Orsi
In this paper, we show that there exists a family of design matrices lacking well-spreadness such that consistent recovery of the parameter vector in the above robust linear regression model is information-theoretically impossible.
no code implementations • 20 Apr 2022 • Tommaso d'Orsi, Luca Trevisan
Strong refutation results based on current approaches construct a certificate that a certain matrix associated to the k-CSP instance is quasirandom.
no code implementations • 14 Feb 2022 • Jingqiu Ding, Tommaso d'Orsi, Chih-Hung Liu, Stefan Tiegel, David Steurer
We develop the first fast spectral algorithm to decompose a random third-order tensor over $\mathbb{R}^d$ of rank up to $O(d^{3/2}/\text{polylog}(d))$.
no code implementations • 16 Nov 2021 • Jingqiu Ding, Tommaso d'Orsi, Rajai Nasser, David Steurer
We develop an efficient algorithm for weak recovery in a robust version of the stochastic block model.
no code implementations • NeurIPS 2021 • Tommaso d'Orsi, Chih-Hung Liu, Rajai Nasser, Gleb Novikov, David Steurer, Stefan Tiegel
For sparse regression, we achieve consistency for optimal sample size $n\gtrsim (k\log d)/\alpha^2$ and optimal error rate $O(\sqrt{(k\log d)/(n\cdot \alpha^2)})$ where $n$ is the number of observations, $d$ is the number of dimensions and $k$ is the sparsity of the parameter vector, allowing the fraction of inliers to be inverse-polynomial in the number of samples.
no code implementations • NeurIPS 2021 • Davin Choo, Tommaso d'Orsi
Even in the restricted case of sparse PCA, known algorithms only recover the sparse vectors for $\lambda \geq \tilde{\mathcal{O}}(k \cdot r)$ while our algorithms require $\lambda \geq \tilde{\mathcal{O}}(k)$.
no code implementations • 12 Nov 2020 • Tommaso d'Orsi, Pravesh K. Kothari, Gleb Novikov, David Steurer
Despite a long history of prior works, including explicit studies of perturbation resilience, the best known algorithmic guarantees for Sparse PCA are fragile and break down under small adversarial perturbations.
no code implementations • 30 Sep 2020 • Tommaso d'Orsi, Gleb Novikov, David Steurer
Concretely, we show that the Huber loss estimator is consistent for every sample size $n= \omega(d/\alpha^2)$ and achieves an error rate of $O(d/\alpha^2n)^{1/2}$.