Semiparametric estimation of the signal subspace∗ Denis Belomestny, Vladimir Panov, Vladimir Spokoiny 1 — Laboratory for Structural Methods of Data Analysis in Predictive Modeling (MIPT) and University Duisburg-Essen, [email protected] 2 — Laboratory for Structural Methods of Data Analysis in Predictive Modeling (MIPT) and University Duisburg-Essen, [email protected] 3 — Labor...

This paper deals with an overview of predictive modeling approaches adopted by marketers to identify segments in the target market which would have the largest response rates pertaining to a certain product or service being offered. The three approaches in the predictive modeling, which have been discussed are Heuristic approach, Statistical approach and Data Mining approach. However, with refe...

Multi-agent predictive modeling is an essential step for understanding physical, social and team-play systems. Recently, Interaction Networks (INs) were proposed for the task of modeling multi-agent physical systems, INs scale with the number of interactions in the system (typically quadratic or higher order in the number of agents). In this paper we introduce VAIN, a novel attentional architec...

OF THE DISSERTATION Modeling Probability Distributions with Predictive State Representations

Lack of labeled training examples is a common problem for many applications. At the same time, there is often an abundance of labeled data from related tasks, although they have different distributions and outputs (e.g., different class labels, and different scales of regression values). In the medical domain, for example, we may have a limited number of vaccine efficacy examples against a new ...