Overparametrized neural networks tend to perfectly fit noisy training data yet generalize well on test data. Inspired by this empirical observation, recent work has sought to understand this phenomenon of benign overfitting or harmless interpolation in the much simpler linear model. Previous theoretical work critically assumes that either the data features are statistically independent or the input data is high-dimensional; this precludes general nonparametric settings with structured feature maps. In this paper, we present a general and flexible framework for upper bounding regression and classification risk in a reproducing kernel Hilbert space. A key contribution is that our framework describes precise sufficient conditions on the data Gram matrix under which harmless interpolation occurs. Our results recover prior independent-features results (with a much simpler analysis), but they furthermore show that harmless interpolation can occur in more general settings such as features that are a bounded orthonormal system. Furthermore, our results show an asymptotic separation between classification and regression performance in a manner that was previously only shown for Gaussian features.

Google engineer Raksha Muthukumar is among more than 500 employees of the tech giant who this week announced the formation of a labor union, a rarity in Silicon Valley. Google engineer Raksha Muthukumar is among more than 500 employees of the tech giant who this week announced the formation of a labor union, a rarity in Silicon Valley. After the death of George Floyd, Google engineer Raksha Muthukumar sent an email to colleagues. In it, she pointed to a list of criminal justice reform groups and bail funds for protesters who were seeking contributions. Soon after, Muthukumar was summoned into a meeting with Google's human relations department.

An artificial intelligence (AI)-trained facial recognition system (FRS) has been installed at the Puratchi Thalaivar Dr. MGR Central railway station for detecting known culprits passing through the gates and alerting authorities. "For the first time, we have introduced the CCTV camera device backed by artificial intelligence. In the existing system, we capture the picture and video of any suspect. But we have to manually analyse the footage to detect their movement. The new system will automatically alert us about known culprits," said a senior police officer of the Government Railway Police (GRP).