Dirichlet process mixture models provide a flexible Bayesian framework for density estimation; ho... more Dirichlet process mixture models provide a flexible Bayesian framework for density estimation; however they are inadequate with respect to modeling sequential data due to the full exchangeability assumption they employ. In this paper we present the temporal Dirichlet process mixture model (TDPM) as a framework for modeling complex longitudinal data. In a TDPM, the data is divided into epochs; all data points inside the same epoch are fully exchangeable, whereas the temporal order is maintained across epochs. Moreover, The number of mixture components in each epoch is unbounded: the components can retain, die out or emerge over time, and the actual parameterization of each component can also evolve over time in a Markovian fashion. We give three equivalent construction of this process as well as a Gibbs sampling algorithm to carry out posterior inference. We demonstrate our model by using it to build an infinite dynamic mixture of Gaussian factors, and a simple non-parametric dynamic topic model applied to the NIPS12 collection.
Stochastic networks are a plausible representation of the relational information among entities i... more Stochastic networks are a plausible representation of the relational information among entities in dynamic systems such as living cells or social communities. While there is a rich literature in estimating a static or temporally invariant network from observation data, little has been done toward estimating time-varying networks from time series of entity attributes. In this paper we present two new machine learning methods for estimating time-varying networks, which both build on a temporally smoothed l1-regularized logistic regression formalism that can be cast as a standard convex-optimization problem and solved efficiently using generic solvers scalable to large networks. We report promising results on recovering simulated timevarying networks. For real data sets, we reverse engineer the latent sequence of temporally rewiring political networks between Senators from the US Senate voting records and the latent evolving regulatory networks underlying 588 genes across the life cycle of Drosophila melanogaster from the microarray time course. . This reprint differs from the original in pagination and typographic detail. 1 2 KOLAR, SONG, AHMED AND XING Each of these characteristics adds a degree of complexity to the interpretation and analysis of networks. In this paper we present a new methodology and analysis that address a particular aspect of dynamic network analysis: how can one reverse engineer networks that are latent, and topologically evolving over time, from time series of nodal attributes. While there is a rich and growing literature on modeling time-invariant networks, much less has been done toward modeling dynamic networks that are rewiring over time. We refer to these time or condition specific circuitries as time-varying networks, which are ubiquitous in various complex systems. Consider the following two real world problems:
A major source of information (often the most crucial and informative part) in scholarly articles... more A major source of information (often the most crucial and informative part) in scholarly articles from scientific journals, proceedings and books are the figures that directly provide images and other graphical illustrations of key experimental results and other scientific contents. In biological articles, a typical figure often comprises multiple panels, accompanied by either scoped or global captioned text. Moreover, the text in the caption contains important semantic entities such as protein names, gene ontology, tissues labels, etc., relevant to the images in the figure.
Proceedings of The National Academy of Sciences, 2009
Author contributions: E.P.X. designed research; A.A. and E.P.X. performed research; A.A. and E.P.... more Author contributions: E.P.X. designed research; A.A. and E.P.X. performed research; A.A. and E.P.X. contributed new reagents/analytic tools; A.A. and E.P.X. analyzed data; and E.P.X. and A.A. wrote the paper.
Stochastic networks are a plausible representation of the relational information among entities i... more Stochastic networks are a plausible representation of the relational information among entities in dynamic systems such as living cells or social communities. While there is a rich literature in estimating a static or temporally invariant network from observation data, little has been done toward estimating time-varying networks from time series of entity attributes. In this paper we present two new machine learning methods for estimating time-varying networks, which both build on a temporally smoothed $l_1$-regularized logistic regression formalism that can be cast as a standard convex-optimization problem and solved efficiently using generic solvers scalable to large networks. We report promising results on recovering simulated time-varying networks. For real data sets, we reverse engineer the latent sequence of temporally rewiring political networks between Senators from the US Senate voting records and the latent evolving regulatory networks underlying 588 genes across the life cycle of Drosophila melanogaster from the microarray time course.
MedLDA: maximum margin supervised topic models for regression and classification
Abstract Supervised topic models utilize document's side information for... more Abstract Supervised topic models utilize document's side information for discovering predictive low dimensional representations of docu-ments; and existing models apply likelihood-based estimation. In this paper, we present a max-margin supervised topic model for both ...
Training Hierarchical Feed-Forward Visual Recognition Models Using Transfer Learning from Pseudo-Tasks
Abstract. Building visual recognition models that adapt across different do-mains is a challengin... more Abstract. Building visual recognition models that adapt across different do-mains is a challenging task for computer vision. While feature-learning machines in the form of hierarchial feed-forward models (eg, convolutional neural net-works) showed promise in this direction, they are ...
In this work, we address the problem of joint modeling of text and citations in the topic modelin... more In this work, we address the problem of joint modeling of text and citations in the topic modeling framework. We present two different models called the Pairwise-Link-LDA and the Link-PLSA-LDA models.
On Tight Approximate Inference of the Logistic-Normal Topic Admixture Model
The Logistic-Normal Topic Admixture Model (LoNTAM), also known as correlated topic model (Blei an... more The Logistic-Normal Topic Admixture Model (LoNTAM), also known as correlated topic model (Blei and Lafferty, 2005), is a promis-ing and expressive admixture-based text model. It can capture topic correlations via the use of a logistic-normal distribu-tion to model non-trivial ...
Dirichlet process mixture models provide a flexible Bayesian framework for density estimation; ho... more Dirichlet process mixture models provide a flexible Bayesian framework for density estimation; however they are inadequate with respect to modeling sequential data due to the full exchangeability assumption they employ. In this paper we present the temporal Dirichlet process mixture model (TDPM) as a framework for modeling complex longitudinal data. In a TDPM, the data is divided into epochs; all data points inside the same epoch are fully exchangeable, whereas the temporal order is maintained across epochs. Moreover, The number of mixture components in each epoch is unbounded: the components can retain, die out or emerge over time, and the actual parameterization of each component can also evolve over time in a Markovian fashion. We give three equivalent construction of this process as well as a Gibbs sampling algorithm to carry out posterior inference. We demonstrate our model by using it to build an infinite dynamic mixture of Gaussian factors, and a simple non-parametric dynamic topic model applied to the NIPS12 collection.
The SLIF project combines text-mining and image processing to extract structured information from... more The SLIF project combines text-mining and image processing to extract structured information from biomedical literature.
Uploads
Papers by Amr Ahmed