ICML
2010
IEEE
13 years 6 months ago
2010
IEEE
We propose a novel boosting algorithm which improves on current algorithms for weighted voting classification by striking a better balance between classification accuracy and the ...
ICML
2010
IEEE
13 years 6 months ago
2010
IEEE
Prediction markets are used in real life to predict outcomes of interest such as presidential elections. In this work we introduce a mathematical theory for Artificial Prediction ...
ICML
2010
IEEE
13 years 6 months ago
2010
IEEE
We propose to combine two approaches for modeling data admitting sparse representations: on the one hand, dictionary learning has proven effective for various signal processing ta...
ICML
2010
IEEE
2010
IEEE
Unsupervised Risk Stratification in Clinical Datasets: Identifying Patients at Risk of Rare Outcomes
13 years 6 months ago
Most existing algorithms for clinical risk stratification rely on labeled training data. Collecting this data is challenging for clinical conditions where only a small percentage ...
ICML
2010
IEEE
13 years 6 months ago
2010
IEEE
We develop a penalized kernel smoothing method for the problem of selecting nonzero elements of the conditional precision matrix, known as conditional covariance selection. This p...
ICML
2010
IEEE
13 years 6 months ago
2010
IEEE
In this paper we consider the problem of policy evaluation in reinforcement learning, i.e., learning the value function of a fixed policy, using the least-squares temporal-differe...
ICML
2010
IEEE
13 years 6 months ago
2010
IEEE
ICML
2010
IEEE
13 years 6 months ago
2010
IEEE
Hidden Markov Models (HMMs) are important tools for modeling sequence data. However, they are restricted to discrete latent states, and are largely restricted to Gaussian and disc...
ICML
2010
IEEE
13 years 6 months ago
2010
IEEE
We present MI-CRF, a conditional random field (CRF) model for multiple instance learning (MIL). MI-CRF models bags as nodes in a CRF with instances as their states. It combines di...
ICML
2010
IEEE
13 years 6 months ago
2010
IEEE
We are interested in learning programs for multiple related tasks given only a few training examples per task. Since the program for a single task is underdetermined by its data, ...