Adversarial examples are not bugs, they are features

A Ilyas, S Santurkar, D Tsipras… - Advances in neural …, 2019‏ - proceedings.neurips.cc
Adversarial examples have attracted significant attention in machine learning, but the
reasons for their existence and pervasiveness remain unclear. We demonstrate that …

Interactive proofs for verifying machine learning

S Goldwasser, GN Rothblum, J Shafer… - 12th Innovations in …, 2021‏ - drops.dagstuhl.de
We consider the following question: using a source of labeled data and interaction with an
untrusted prover, what is the complexity of verifying that a given hypothesis is" approximately …

Sgd learns one-layer networks in wgans

Q Lei, J Lee, A Dimakis… - … Conference on Machine …, 2020‏ - proceedings.mlr.press
Generative adversarial networks (GANs) are a widely used framework for learning
generative models. Wasserstein GANs (WGANs), one of the most successful variants of …

On the limits of language generation: Trade-offs between hallucination and mode collapse

A Kalavasis, A Mehrotra, G Velegkas - arxiv preprint arxiv:2411.09642, 2024‏ - arxiv.org
Specifying all desirable properties of a language model is challenging, but certain
requirements seem essential. Given samples from an unknown language, the trained model …

Understanding adversarial robustness against on-manifold adversarial examples

J **ao, L Yang, Y Fan, J Wang, ZQ Luo - Pattern Recognition, 2025‏ - Elsevier
Deep neural networks (DNNs) are shown to be vulnerable to adversarial examples. A well-
trained model can be easily attacked by adding small perturbations to the original data. One …

Transfer learning beyond bounded density ratios

A Kalavasis, I Zadik, M Zampetakis - arxiv preprint arxiv:2403.11963, 2024‏ - arxiv.org
We study the fundamental problem of transfer learning where a learning algorithm collects
data from some source distribution $ P $ but needs to perform well with respect to a different …

Computationally and statistically efficient truncated regression

C Daskalakis, T Gouleakis, C Tzamos… - … on learning theory, 2019‏ - proceedings.mlr.press
We provide a computationally and statistically efficient estimator for the classical problem of
truncated linear regression, where the dependent variable $ y=\vec {w}^{\rm T}\vec …

Efficient truncated statistics with unknown truncation

V Kontonis, C Tzamos… - 2019 IEEE 60th Annual …, 2019‏ - ieeexplore.ieee.org
We study the problem of estimating the parameters of a Gaussian distribution when samples
are only shown if they fall in some (unknown) set. This core problem in truncated statistics …

Finite-sample symmetric mean estimation with fisher information rate

S Gupta, JCH Lee, E Price - The Thirty Sixth Annual …, 2023‏ - proceedings.mlr.press
The mean of an unknown variance-$\sigma^ 2$ distribution $ f $ can be estimated from $ n $
samples with variance $\frac {\sigma^ 2}{n} $ and nearly corresponding subgaussian rate …

Learning exponential families from truncated samples

J Lee, A Wibisono… - Advances in Neural …, 2023‏ - proceedings.neurips.cc
Missing data problems have many manifestations across many scientific fields. A
fundamental type of missing data problem arises when samples are\textit {truncated}, ie …