On Reversing Jensen's Inequality
نویسندگان
چکیده
Alex Pentland MIT Media Lab Cambridge, MA 02139 [email protected] Jensen's inequality is a powerful mathematical tool and one of the workhorses in statistical learning. Its applications therein include the EM algorithm, Bayesian estimation and Bayesian inference. Jensen computes simple lower bounds on otherwise intractable quantities such as products of sums and latent log-likelihoods. This simplification then permits operations like integration and maximization. Quite often (i.e. in discriminative learning) upper bounds are needed as well. We derive and prove an efficient analytic inequality that provides such variational upper bounds. This inequality holds for latent variable mixtures of exponential family distributions and thus spans a wide range of contemporary statistical models. We also discuss applications of the upper bounds including maximum conditional likelihood, large margin discriminative models and conditional Bayesian inference. Convergence, efficiency and prediction results are shown. 1
منابع مشابه
Grüss Type Discrete Inequalities in Inner Product Spaces, Revisited
Some sharp inequalities of Grüss type for sequences of vectors in real or complex inner product spaces are obtained. Applications for Jensen's inequality for convex functions defined on such spaces are also provided.
متن کاملAnother Proof of an Extension of a Curious Identity
. (1) In 2002 Z. W. Sun [7] proved the case when z = 1 using double recursion. Later four alternative proofs have been provided for the special case. A generating function proof was given by A. Panholzer and H. Prodinger [6]; D. Merlini and R. Sprugnoli [5] established it through Riordan arrays; S. B. Ekhad and M. Mohammed [4] proved it based on a WZ method. Later, W. Chu and L. V. D. Claudio [...
متن کاملA new refinement of Jensen's inequality in linear spaces with applications
A new refinement of Jensen’s celebrated inequality for functions defined on convex sets in linear spaces is given. Applications for norms, mean f -deviation and f -divergences are provided as well.
متن کاملA Univariate Extension of Jensen's Inequality
The result in this paper exp lains some of the qualitative nature of Jensen's inequality. It is shown that the more disperse the distribution of a random variable is, the smaller is the expectation of any concave function of it. This result can be used to show the inadequacy of some current methods of reporting environmental data by using geometric means, and it extends the resu lt of I. Billic...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2000