We present a convolutional network that is equivariant to rigid body motions.

One approach to defend from adversarial perturbations is to mask gradients.

Neural Information Processing Systems http://nips.cc/

V oice cloning is a highly desired feature for personalized speech interfaces.

To this end, we use algorithms developed in the gradient-based multi-objective optimization literature. These algorithms are not directly applicable to large-scale learning problems since they scale poorly with the dimensionality of the gradients and the number of tasks. We therefore propose an upper bound for the multi-objective loss and show that it can be optimized efficiently. We further prove that optimizing this upper bound yields a Pareto optimal solution under realistic assumptions.

After defying multiple mass extinctions on Earth, the hardy plant passes an intergalactic test. Breakthroughs, discoveries, and DIY tips sent every weekday. While it may appear humble, Earth's moss is built darn tough. It thrives in extreme environments -from the bitter cold, low-oxygen air of the Himalayas, down to the parched sands of Death Valley. Some species even make their home among the lava fields of active volcanoes .

Neural Information Processing Systems http://nips.cc/

We propose a general, theoretically justified mechanism for processing missing data by neural networks. Our idea is to replace typical neuron's response in the

There are many clustering algorithms which perform well towards their aim of finding cohesive groups of instances.

Such large-scale recordings pose a major challenge to existing statistical methods for neural data analysis. Here we develop highly scalable approximate inference methods for Poisson generalized linear models (GLMs) that require only a single pass over the data.