Artificial intelligence publications

We show that box invariance can be checked in cubic time for linear and affine systems, and that it remains decidable for classes of nonlinear systems of interest (with polynomial…

We show that simple efficient many-class learning can perform well for action prediction, significantly improving over previously published results and baselines.

A mixed-initiative agent for personal time management interacts not only with its human owner but also with other agents and humans that share or depend on the same time commitments.

In this paper, we argue that in dealing with ADLs, it is beneficial to exploit both their typical duration patterns and inherent hierarchical structures.

Inspired by popular co-training and domain adaptation methods, we propose a co-adaptation algorithm. The goal is improving the performance of a dialog act segmentation model by exploiting the vast amount…

Abstract The challenge in the DARPA Learning Applied to Ground Robots (LAGR) project is to autonomously navigate a small robot using stereo vision as the main sensor. During this project,…

We report our first experiences with Leaving Flatland, an exploratory project that studies the key challenges of closing the loop between autonomous perception and action on challenging terrain.

This case study article describes the iterative design process of an adaptive, mixed-initiative calendaring tool with embedded artificial intelligence. 

We investigate prediction of users’ desktop activities in the Unix domain. We show that simple efficient many-class learning can perform well for action prediction, significantly improving over previously published results…

In the first part of the paper, we show that efficient many-class learning can perform well for action prediction in the Unix domain, significantly improving over previously published results