Rajesh Ranganath

Convolutional deep belief networks for scalable unsupervised learning of hierarchical representations

There has been much interest in unsupervised learning of hierarchical generative models such as deep belief networks. Scaling such models to full-sized, high-dimensional images remains a difficult problem. To address this problem, we present the convolutional deep belief network, a hierarchical generative model which scales to realistic image sizes. This model is translation-invariant and supports efficient bottom-up and top-down probabilistic inference. Key to our approach is probabilistic max-pooling, a novel technique which shrinks the representations of higher layers in a probabilistically sound way. Our experiments show that the algorithm learns useful high-level visual features, such as object parts, from unlabeled images of objects and natural scenes. We demonstrate excellent performance on several visual recognition tasks and show that our model can perform hierarchical (bottom-up and top-down) inference over full-sized images.

Unsupervised learning of hierarchical representations with convolutional deep belief networks

There has been much interest in unsupervised learning of hierarchical generative models such as deep belief networks (DBNs); however, scaling such models to full-sized, high-dimensional images remains a difficult problem. To address this problem, we present the convolutional deep belief network, a hierarchical generative model that scales to realistic image sizes. This model is translation-invariant and supports efficient bottom-up and top-down probabilistic inference. Key to our approach is probabilistic max-pooling, a novel technique that shrinks the representations of higher layers in a probabilistically sound way. Our experiments show that the algorithm learns useful high-level visual features, such as object parts, from unlabeled images of objects and natural scenes. We demonstrate excellent performance on several visual recognition tasks and show that our model can perform hierarchical (bottom-up and top-down) inference over full-sized images.

Extracting Social Meaning: Identifying Interactional Style in Spoken Conversation

Automatically extracting social meaning and intention from spoken dialogue is an important task for dialogue systems and social computing. We describe a system for detecting elements of interactional style: whether a speaker is awkward, friendly, or flirtatious. We create and use a new spoken corpus of 991 4-minute speed-dates. Participants rated their interlocutors for these elements of style. Using rich dialogue, lexical, and prosodic features, we are able to detect flirtatious, awkward, and friendly styles in noisy natural conversational data with up to 75% accuracy, compared to a 50% baseline. We describe simple ways to extract relatively rich dialogue features, and analyze which features performed similarly for men and women and which were gender-specific.

It's Not You, it's Me: Detecting Flirting and its Misperception in Speed-Dates

Automatically detecting human social intentions from spoken conversation is an important task for dialogue understanding. Since the social intentions of the speaker may differ from what is perceived by the hearer, systems that analyze human conversations need to be able to extract both the perceived and the intended social meaning. We investigate this difference between intention and perception by using a spoken corpus of speed-dates in which both the speaker and the listener rated the speaker on flirtatiousness. Our flirtation-detection system uses prosodic, dialogue, and lexical features to detect a speakers intent to flirt with up to 71.5% accuracy, significantly outperforming the baseline, but also outperforming the human inter-locuters. Our system addresses lexical feature sparsity given the small amount of training data by using an autoencoder network to map sparse lexical feature vectors into 30 compressed features. Our analysis shows that humans are very poor perceivers of intended flirtatiousness, instead often projecting their own intended behavior onto their interlocutors.

Detecting friendly, flirtatious, awkward, and assertive speech in speed-dates

Automatically detecting human social intentions and attitudes from spoken conversation is an important task for speech processing and social computing. We describe a system for detecting interpersonal stance: whether a speaker is flirtatious, friendly, awkward, or assertive. We make use of a new spoken corpus of over 1000 4-min speed-dates. Participants rated themselves and their interlocutors for these interpersonal stances, allowing us to build detectors for style both as interpreted by the speaker and as perceived by the hearer. We use lexical, prosodic, and dialog features in an SVM classifier to detect very clear styles (the strongest 10% in each stance) with up to 75% accuracy on previously seen speakers (50% baseline) and up to 59% accuracy on new speakers (48% baseline). A feature analysis suggests that flirtation is marked by joint focus on the woman as a target of the conversation, awkwardness by decreased speaker involvement, and friendliness by a conversational style including other-directed laughter and appreciations. Our work has implications for our understanding of interpersonal stance, their linguistic expression, and their automatic extraction.

Risk prediction for chronic kidney disease progression using heterogeneous electronic health record data and time series analysis.

Abstract Background As adoption of electronic health records continues to increase, there is an opportunity to incorporate clinical documentation as well as laboratory values and demographics into risk prediction modeling. Objective The authors develop a risk prediction model for chronic kidney disease (CKD) progression from stage III to stage IV that includes longitudinal data and features drawn from clinical documentation. Methods The study cohort consisted of 2908 primary-care clinic patients who had at least three visits prior to January 1, 2013 and developed CKD stage III during their documented history. Development and validation cohorts were randomly selected from this cohort and the study datasets included longitudinal inpatient and outpatient data from these populations. Time series analysis (Kalman filter) and survival analysis (Cox proportional hazards) were combined to produce a range of risk models. These models were evaluated using concordance, a discriminatory statistic. Results A risk model incorporating longitudinal data on clinical documentation and laboratory test results (concordance 0.849) predicts progression from state III CKD to stage IV CKD more accurately when compared to a similar model without laboratory test results (concordance 0.733, P<.001), a model that only considers the most recent laboratory test results (concordance 0.819, P < .031) and a model based on estimated glomerular filtration rate (concordance 0.779, P < .001). Conclusions A risk prediction model that takes longitudinal laboratory test results and clinical documentation into consideration can predict CKD progression from stage III to stage IV more accurately than three models that do not take all of these variables into consideration.

Topographic Factor Analysis: A Bayesian Model for Inferring Brain Networks from Neural Data

The neural patterns recorded during a neuroscientific experiment reflect complex interactions between many brain regions, each comprising millions of neurons. However, the measurements themselves are typically abstracted from that underlying structure. For example, functional magnetic resonance imaging (fMRI) datasets comprise a time series of three-dimensional images, where each voxel in an image (roughly) reflects the activity of the brain structure(s)–located at the corresponding point in space–at the time the image was collected. FMRI data often exhibit strong spatial correlations, whereby nearby voxels behave similarly over time as the underlying brain structure modulates its activity. Here we develop topographic factor analysis (TFA), a technique that exploits spatial correlations in fMRI data to recover the underlying structure that the images reflect. Specifically, TFA casts each brain image as a weighted sum of spatial functions. The parameters of those spatial functions, which may be learned by applying TFA to an fMRI dataset, reveal the locations and sizes of the brain structures activated while the data were collected, as well as the interactions between those structures.

Correlated Random Measures

ABSTRACT We develop correlated random measures, random measures where the atom weights can exhibit a flexible pattern of dependence, and use them to develop powerful hierarchical Bayesian nonparametric models. Hierarchical Bayesian nonparametric models are usually built from completely random measures, a Poisson-process-based construction in which the atom weights are independent. Completely random measures imply strong independence assumptions in the corresponding hierarchical model, and these assumptions are often misplaced in real-world settings. Correlated random measures address this limitation. They model correlation within the measure by using a Gaussian process in concert with the Poisson process. With correlated random measures, for example, we can develop a latent feature model for which we can infer both the properties of the latent features and their dependency pattern. We develop several other examples as well. We study a correlated random measure model of pairwise count data. We derive an efficient variational inference algorithm and show improved predictive performance on large datasets of documents, web clicks, and electronic health records. Supplementary materials for this article are available online.

Hierarchical topographic factor analysis

Recent work has revealed that cognitive processes are often reflected in patterns of functional connectivity throughout the brain (for review see [16]). However, examining functional connectivity patterns using traditional methods carries a substantial computational burden (of computing time and memory). Here we present a technique, termed Hierarchical topographic factor analysis, for efficiently discovering brain networks in large multi-subject neuroimaging datasets.

A probabilistic approach to discovering dynamic full-brain functional connectivity patterns

ABSTRACT Recent research shows that the covariance structure of functional magnetic resonance imaging (fMRI) data – commonly described as functional connectivity – can change as a function of the participants cognitive state (for review see Turk‐Browne, 2013). Here we present a Bayesian hierarchical matrix factorization model, termed hierarchical topographic factor analysis (HTFA), for efficiently discovering full‐brain networks in large multi‐subject neuroimaging datasets. HTFA approximates each subjects network by first re‐representing each brain image in terms of the activities of a set of localized nodes, and then computing the covariance of the activity time series of these nodes. The number of nodes, along with their locations, sizes, and activities (over time) are learned from the data. Because the number of nodes is typically substantially smaller than the number of fMRI voxels, HTFA can be orders of magnitude more efficient than traditional voxel‐based functional connectivity approaches. In one case study, we show that HTFA recovers the known connectivity patterns underlying a collection of synthetic datasets. In a second case study, we illustrate how HTFA may be used to discover dynamic full‐brain activity and connectivity patterns in real fMRI data, collected as participants listened to a story. In a third case study, we carried out a similar series of analyses on fMRI data collected as participants viewed an episode of a television show. In these latter case studies, we found that the HTFA‐derived activity and connectivity patterns can be used to reliably decode which moments in the story or show the participants were experiencing. Further, we found that these two classes of patterns contained partially non‐overlapping information, such that decoders trained on combinations of activity‐based and dynamic connectivity‐based features performed better than decoders trained on activity or connectivity patterns alone. We replicated this latter result with two additional (previously developed) methods for efficiently characterizing full‐brain activity and connectivity patterns. HIGHLIGHTSHierarchical Topographic Factor Analysis identifies full‐brain activity and network dynamics in multi‐subject brain data.We applied HTFA to fMRI data and found that activity and connectivity patterns reflect story and movie timing information.Activity and connectivity patterns contain partially non‐overlapping information about when in a story or movie participants are experiencing.