We formulate a principle for classification with the knowledge of the marginal distribution over the data points (unlabeled data). The principle is cast in terms of Tikhonov styl...
A fundamental question in causal inference is whether it is possible to reliably infer the manipulation effects from observational data. There are a variety of senses of asymptot...
When aligning texts in very different languages such as Korean and English, structural features beyond word or phrase give useful intbrmation. In this paper, we present a method f...
This paper presents a novel hybrid method combining genetic programming and decision tree learning. The method starts by estimating a benchmark level of reasonable accuracy, based ...
While conventional malware detection approaches increasingly fail, modern heuristic strategies often perform dynamically, which is not possible in many applications due to related ...