The expectation maximization (EM) algorithm is a widely used maximum likelihood estimationprocedure for statistical models when the values of some of the variables in the model are not observed. Very often, however, our aim is primarily tofind a model that assigns values to the latent variables that have intended meaning for our data and maximizing expected likelihood only sometimes accomplishes this.Unfortunately, it is typically difficult to add even simple a-priori information about latent variables in graphical models without making the models overly complex or intractable. In this paper, we present an efficient, principled way to inject rich constraints on the posteriors of latent variables into the EM algorithm. Our method can be used to learn tractable graphical models that satisfy additional,otherwise intractable constraints. Focusing on clustering and the alignment problem for statistical machine translation, we show that simple, intuitive posteriorconstraints can greatly improve the performance over standard baselines and be competitive with more complex, intractable models.

Motivated in part by the hierarchical organization of cortex, a number of algorithms have recently been proposed that try to learn hierarchical, or ``deep,'' structure from unlabeled data. While several authors have formally or informally compared their algorithms to computations performed in visual area V1 (and the cochlea), little attempt has been made thus far to evaluate these algorithms in terms of their fidelity for mimicking computations at deeper levels in the cortical hierarchy. This paper presents an unsupervised learning model that faithfully mimics certain properties of visual area V2. Specifically, we develop a sparse variant of the deep belief networks of Hinton et al. (2006). We learn two layers of nodes in the network, and demonstrate that the first layer, similar to prior work on sparse coding and ICA, results in localized, oriented, edge filters, similar to the Gabor functions known to model V1 cell receptive fields. Further, the second layer in our model encodes correlations of the first layer responses in the data. Specifically, it picks up both collinear (``contour'') features as well as corners and junctions. More interestingly, in a quantitative comparison, the encoding of these more complex ``corner'' features matches well with the results from the Ito & Komatsu's study of biological V2 responses. This suggests that our sparse variant of deep belief networks holds promise for modeling more higher-order features.

The expectation maximization (EM) algorithm is a widely used maximum likelihood estimation procedure for statistical models when the values of some of the variables in the model are not observed. Very often, however, our aim is primarily to find a model that assigns values to the latent variables that have intended meaning for our data and maximizing expected likelihood only sometimes accomplishes this. Unfortunately, it is typically difficult to add even simple a-priori information about latent variables in graphical models without making the models overly complex or intractable. In this paper, we present an efficient, principled way to inject rich constraints on the posteriors of latent variables into the EM algorithm. Our method can be used to learn tractable graphical models that satisfy additional, otherwise intractable constraints. Focusing on clustering and the alignment problem for statistical machine translation, we show that simple, intuitive posterior constraints can greatly improve the performance over standard baselines and be competitive with more complex, intractable models.

It is becoming increasingly evident that organisms acting in uncertain dynamical environments often employ exact or approximate Bayesian statistical calculations in order to continuously estimate the environmental state, integrate information from multiple sensory modalities, form predictions and choose actions. What is less clear is how these putative computations are implemented by cortical neural networks. An additional level of complexity is introduced because these networks observe the world through spike trains received from primary sensory afferents, rather than directly. A recent line of research has described mechanisms by which such computations can be implemented using a network of neurons whose activity directly represents a probability distribution across the possible "world states". Much of this work, however, uses various approximations, which severely restrict the domain of applicability of these implementations. Here we make use of rigorous mathematical results from the theory of continuous time point process filtering, and show how optimal real-time state estimation and prediction may be implemented in a general setting using linear neural networks. We demonstrate the applicability of the approach with several examples, and relate the required network properties to the statistical nature of the environment, thereby quantifying the compatibility of a given network with its environment.

Dynamic Bayesian networks are structured representations of stochastic processes. Despite their structure, exact inference in DBNs is generally intractable. One approach to approximate inference involves grouping the variables in the process into smaller factors and keeping independent beliefs over these factors. In this paper we present several techniques for decomposing a dynamic Bayesian network automatically to enable factored inference. We examine a number of features of a DBN that capture different types of dependencies that will cause error in factored inference. An empirical comparison shows that the most useful of these is a heuristic that estimates the mutual information introduced between factors by one step of belief propagation.

There has been an explosion of interest in statistical models for analyzing network data, and considerable interest in the class of exponential random graph (ERG) models, especially in connection with difficulties in computing maximum likelihood estimates. The issues associated with these difficulties relate to the broader structure of discrete exponential families. This paper re-examines the issues in two parts. First we consider the closure of $k$-dimensional exponential families of distribution with discrete base measure and polyhedral convex support $\mathrm{P}$. We show that the normal fan of $\mathrm{P}$ is a geometric object that plays a fundamental role in deriving the statistical and geometric properties of the corresponding extended exponential families. We discuss its relevance to maximum likelihood estimation, both from a theoretical and computational standpoint. Second, we apply our results to the analysis of ERG models. In particular, by means of a detailed example, we provide some characterization of the properties of ERG models, and, in particular, of certain behaviors of ERG models known as degeneracy.

Feature Markov Decision Processes (PhiMDPs) are well-suited for learning agents in general environments. Nevertheless, unstructured (Phi)MDPs are limited to relatively simple environments. Structured MDPs like Dynamic Bayesian Networks (DBNs) are used for large-scale real-world problems. In this article I extend PhiMDP to PhiDBN. The primary contribution is to derive a cost criterion that allows to automatically extract the most relevant features from the environment, leading to the "best" DBN representation. I discuss all building blocks required for a complete general learning algorithm.

We give polynomial-time algorithms for the exact computation of lowest-energy (ground) states, worst margin violators, log partition functions, and marginal edge probabilities in certain binary undirected graphical models. Our approach provides an interesting alternative to the well-known graph cut paradigm in that it does not impose any submodularity constraints; instead we require planarity to establish a correspondence with perfect matchings (dimer coverings) in an expanded dual graph. We implement a unified framework while delegating complex but well-understood subproblems (planar embedding, maximum-weight perfect matching) to established algorithms for which efficient implementations are freely available. Unlike graph cut methods, we can perform penalized maximum-likelihood as well as maximum-margin parameter estimation in the associated conditional random fields (CRFs), and employ marginal posterior probabilities as well as maximum a posteriori (MAP) states for prediction. Maximum-margin CRF parameter estimation on image denoising and segmentation problems shows our approach to be efficient and effective. A C++ implementation is available from http://nic.schraudolph.org/isinf/

Motivation: Profile hidden Markov Models (pHMMs) are a popular and very useful tool in the detection of the remote homologue protein families. Unfortunately, their performance is not always satisfactory when proteins are in the 'twilight zone'. We present HMMER-STRUCT, a model construction algorithm and tool that tries to improve pHMM performance by using structural information while training pHMMs. As a first step, HMMER-STRUCT constructs a set of pHMMs. Each pHMM is constructed by weighting each residue in an aligned protein according to a specific structural property of the residue. Properties used were primary, secondary and tertiary structures, accessibility and packing. HMMER-STRUCT then prioritizes the results by voting. Results: We used the SCOP database to perform our experiments. Throughout, we apply leave-one-family-out cross-validation over protein superfamilies. First, we used the MAMMOTH-mult structural aligner to align the training set proteins. Then, we performed two sets of experiments. In a first experiment, we compared structure weighted models against standard pHMMs and against each other. In a second experiment, we compared the voting model against individual pHMMs. We compare method performance through ROC curves and through Precision/Recall curves, and assess significance through the paired two tailed t-test. Our results show significant performance improvements of all structurally weighted models over default HMMER, and a significant improvement in sensitivity of the combined models over both the original model and the structurally weighted models.

This paper proposes a novel solution to spam detection inspired by a model of the adaptive immune system known as the crossregulation model. We report on the testing of a preliminary algorithm on six e-mail corpora. We also compare our results statically and dynamically with those obtained by the Naive Bayes classifier and another binary classification method we developed previously for biomedical text-mining applications. We show that the cross-regulation model is competitive against those and thus promising as a bio-inspired algorithm for spam detection in particular, and binary classification in general.