Rates of estimation for determinantal point processes
نویسندگان
چکیده
Determinantal point processes (DPPs) have wide-ranging applications in machine learning, where they are used to enforce the notion of diversity in subset selection problems. Many estimators have been proposed, but surprisingly the basic properties of the maximum likelihood estimator (MLE) have received little attention. In this paper, we study the local geometry of the expected log-likelihood function to prove several rates of convergence for the MLE. We also give a complete characterization of the case where the MLE converges at a parametric rate. Even in the latter case, we also exhibit a potential curse of dimensionality where the asymptotic variance of the MLE is exponentially large in the dimension of the problem. AMS 2000 subject classifications: Primary 62F10; secondary 60G55.
منابع مشابه
Large-Margin Determinantal Point Processes
Investigate determinantal point processes (DPPs) for discriminative subset selection Proposemargin based parameter estimation to explicitly track errors in selecting subsets Balance different types of evaluation metrics, e.g., precision and recall Improve modeling flexibility by multiple-kernel based parameterization Attain state-of-the-art performance on the tasks of video and docume...
متن کاملStructured Determinantal Point Processes
We present a novel probabilistic model for distributions over sets of structures— for example, sets of sequences, trees, or graphs. The critical characteristic of our model is a preference for diversity: sets containing dissimilar structures are more likely. Our model is a marriage of structured probabilistic models, like Markov random fields and context free grammars, with determinantal point ...
متن کاملDeterminantal point processes
In this survey we review two topics concerning determinantal (or fermion) point processes. First, we provide the construction of diffusion processes on the space of configurations whose invariant measure is the law of a determinantal point process. Second, we present some algorithms to sample from the law of a determinantal point process on a finite window. Related open problems are listed.
متن کاملChange Point Estimation in High Yield Processes in the Presence of Serial Correlation
Change point estimation is as an effective method for identifying the time of a change in production and service processes. In most of the statistical quality control literature, it is usually assumed that the quality characteristic of interest is independently and identically distributed over time. It is obvious that this assumption could be easily violated in practice. In this paper, we use m...
متن کاملLoop-free Markov chains as determinantal point processes
We show that any loop-free Markov chain on a discrete space can be viewed as a determinantal point process. As an application we prove central limit theorems for the number of particles in a window for renewal processes and Markov renewal processes with Bernoulli noise. Introduction Let X be a discrete space. A (simple) random point process P on X is a probability measure on the set 2 of all su...
متن کامل