In the fields of artificial intelligence and machine learning, we are often interested in making probabilistic predictions from time series data. However, many standard algorithms are not designed to give accurate results when we have only a small amount of data obtained via a heavily biased exploration policy. This thesis explores combining the techniques of matrix completion with spectral learning of predictive models in order to tackle this issue. In the matrix completion step, we give each entry an importance weight in order to emphasize each estimate based on the sample size of the dataset from which is was computed. We also add constraints in order to ensure that the result has the desired structure. By using regularization, we encour...
We frame max-margin learning of latent variable structured prediction models as a convex optimizatio...
The completion of low rank matrices from few entries is a task with many practical applications. We ...
We discuss how a large class of regularization methods, collectively known as spectral regularizatio...
We propose a novel class of algorithms for low rank matrix completion. Our approach builds on novel ...
Predictive state representations (PSRs) are models of dynamical systems that represent state as a ve...
We propose a novel class of algorithms for low rank matrix completion. Our ap-proach builds on novel...
Bayesian matrix completion has been studied based on a low-rank matrix factorization formulation wit...
<p>A central problem in artificial intelligence is to choose actions to maximize reward in a partial...
Student Paper Awards NIPS 2012Many tasks in text and speech processing and computational biology req...
Abstract. It is the main goal of this paper to propose a novel method to per-form matrix completion ...
Predictive state representations (PSR) have emerged as a powerful method for modelling partially obs...
International audienceWe observe $(X_i,Y_i)_{i=1}^n$ where the $Y_i$'s are real valued outputs and t...
The completion of low rank matrices from few entries is a task with many practical applications. We ...
Premi extraordinari doctorat curs 2012-2013, àmbit Enginyeria de les TICThe present thesis addresses...
The first two parts of the thesis study pseudo-Bayesian estimation for the problem of matrix complet...
We frame max-margin learning of latent variable structured prediction models as a convex optimizatio...
The completion of low rank matrices from few entries is a task with many practical applications. We ...
We discuss how a large class of regularization methods, collectively known as spectral regularizatio...
We propose a novel class of algorithms for low rank matrix completion. Our approach builds on novel ...
Predictive state representations (PSRs) are models of dynamical systems that represent state as a ve...
We propose a novel class of algorithms for low rank matrix completion. Our ap-proach builds on novel...
Bayesian matrix completion has been studied based on a low-rank matrix factorization formulation wit...
<p>A central problem in artificial intelligence is to choose actions to maximize reward in a partial...
Student Paper Awards NIPS 2012Many tasks in text and speech processing and computational biology req...
Abstract. It is the main goal of this paper to propose a novel method to per-form matrix completion ...
Predictive state representations (PSR) have emerged as a powerful method for modelling partially obs...
International audienceWe observe $(X_i,Y_i)_{i=1}^n$ where the $Y_i$'s are real valued outputs and t...
The completion of low rank matrices from few entries is a task with many practical applications. We ...
Premi extraordinari doctorat curs 2012-2013, àmbit Enginyeria de les TICThe present thesis addresses...
The first two parts of the thesis study pseudo-Bayesian estimation for the problem of matrix complet...
We frame max-margin learning of latent variable structured prediction models as a convex optimizatio...
The completion of low rank matrices from few entries is a task with many practical applications. We ...
We discuss how a large class of regularization methods, collectively known as spectral regularizatio...