C ount data arise in various organizational settings. When the release of such data is sensitive, organizations need information-disclosure policies that protect data confidentiality while still providing data access. In contrast to extant disclosure policies, we describe a new policy for count tables that is based on disclosing only the sufficient statistics of a flexible discrete distribution...

In this article, we study a Bayesian hierarchical model for profiling health-care facilities using approximately sufficient statistics for aggregate facility-level data when the patient-level data sets are very large or unavailable. Starting with a desired patient-level model, we give several approximate models and the corresponding summary statistics necessary to implement the approximations. ...

Working in the framework of PAC-learning theory, we present special statistics for accomplishing in polynomial time proper learning of DNF boolean formulas having a fixed number of monomials. Our statistics turn out to be near sufficient for a large family of distribution laws—that we call butterfly distributions. We develop a theory of most powerful learning for analyzing the performance of le...

This paper presents a necessary and sufficient condition for discrete-time singular Markovian jump linear systems with defective statistics of modes transitions to be regular, causal, and stochastically stable. The static output-feedback stabilization problem for discrete-time singular Markovian jump linear system with defective statistics of modes transitions is studied. Furthermore, the defec...

Chentsov’s theorem characterizes the Fisher information metrics on statistical models as essentially the only Riemannian metrics that are invariant under sufficient statistics. This implies that statistical models are naturally equipped with geometric structures, so Chentsov’s theorem explains why many statistical properties can be described in geometric terms. However, despite being one of the...

For a wide variety of classification algorithms, scalability to large databases can be achieved by observing that most algorithms are driven by a set of sufficient statistics that are significantly smaller than the data. By relying on a SQL backend to compute the sufficient statistics, we leverage the query processing system of SQL databases and avoid the need for moving data to the client. We ...

This work explores the use of DNN/RNN for extracting Baum-Welch sufficient statistics in place of the conventional GMM-UBM in speaker recognition. In this framework, the DNN/RNN is trained for automatic speech recognition (ASR) and each of the output unit corresponds to a component of GMM-UBM. Then the outputs of network are combined with acoustic features to calculate sufficient statistics for...

The choice of the summary statistics in Bayesian inference and in particular in ABC algorithms is paramount to produce a valid outcome. We derive necessary and sufficient conditions on those statistics for the corresponding Bayes factor to be convergent, namely to asymptotically select the true model. Those conditions, which amount to the expectations of the summary statistics to asymptotically...