Mobile health (mHealth) has emerged as a transformative solution to enhance healthcare accessibility and affordability, particularly in resource-constrained regions and low-to-middle-income countries. mHealth leverages mobile platforms to improve healthcare accessibility, addressing radiologist shortages in low-resource settings by enabling remote diagnosis and consultation through mobile devices. Mobile phones allow healthcare workers to transmit radiographic images, such as chest X-rays (CXR), to specialists or AI-driven models for interpretation. However, AI-based diagnosis using CXR images shared via apps like WhatsApp suffers from reduced predictability and explainability due to compression artifacts, and there is a lack of datasets to systematically study these challenges. To address this, we introduce CheXwhatsApp, a dataset of 141,804 paired original and WhatsApp-compressed CXR images. We present a benchmarking study which shows the dataset improves prediction stability and explainability of state-of-the-art models by up to 80%, while also enhancing localization performance. CheXwhatsApp is open-sourced to support advancements in mHealth applications for CXR analysis1.

A central problem related to transformers can be stated as follows: given two $n \times d$ matrices $Q$ and $K$, and a non-negative function $f$, define the matrix $A$ as follows: (1) apply the function $f$ to each entry of the $n \times n$ matrix $Q K^T$, and then (2) normalize each of the row sums of $A$ to be equal to $1$. The matrix $A$ can be computed in $O(n^2 d)$ time assuming $f$ can be applied to a number in constant time, but the quadratic dependence on $n$ is prohibitive in applications where it corresponds to long context lengths. For a large class of functions $f$, we show how to find all the "large attention scores", i.e., entries of $A$ which are at least a positive value $\varepsilon$, in time with linear dependence on $n$ (i.e., $n \cdot \textrm{poly}(d/\varepsilon)$) for a positive parameter $\varepsilon > 0$. Our class of functions include all functions $f$ of the form $f(x) = |x|^p$, as explored recently in transformer models. Using recently developed tools from randomized numerical linear algebra, we prove that for any $K$, there is a "universal set" $U \subset [n]$ of size independent of $n$, such that for any $Q$ and any row $i$, the large attention scores $A_{i,j}$ in row $i$ of $A$ all have $j \in U$. We also find $U$ in $n \cdot \textrm{poly}(d/\varepsilon)$ time. Notably, we (1) make no assumptions on the data, (2) our workspace does not grow with $n$, and (3) our algorithms can be computed in streaming and parallel settings. We empirically show the benefits of our scheme for vision transformers, showing how to train new models that use our universal set while training as well, showing that our model is able to consistently select "important keys'" during training. We also provide theoretical motivation by formulating a planted model in which our efficient algorithms provably identify relevant keys for each query.

Open weight models, which are ubiquitous, rarely provide access to their training data or loss function. This makes modifying such models for tasks such as pruning or unlearning, which are constrained by this unavailability, an active area of research. Existing techniques typically require gradients or ground-truth labels, rendering them infeasible in settings with limited computational resources. In this work, we investigate the fundamental question of identifying components that are critical to the model's predictive performance, without access to either gradients or the loss function, and with only distributional access such as synthetic data. We theoretically demonstrate that the global error is linearly bounded by local reconstruction errors for Lipschitz-continuous networks such as CNNs and well-trained Transformers (which, contrary to existing literature, we find exhibit Lipschitz continuity). This motivates using the locally reconstructive behavior of component subsets to quantify their global importance, via a metric that we term *Subset Fidelity*. In the uncorrelated features setting, selecting individual components based on their Subset Fidelity scores is optimal, which we utilize to propose **ModHiFi**, an algorithm for model modification that requires neither training data nor access to a loss function. **ModHiFi-P**, for structured pruning, achieves an 11\% speedup over the current state of the art on ImageNet models and competitive performance on language models. **ModHiFi-U**, for classwise unlearning, achieves complete unlearning on CIFAR-10 without fine-tuning and demonstrates competitive performance on Swin Transformers.

To fix the `bias in, bias out' problem in fair machine learning, it is important to steer feature distributions of data or internal representations of Large Language Models (LLMs) to \emph{ideal} ones that guarantee group-fair outcomes. Previous work on fair generative models and representation steering could greatly benefit from provable fairness guarantees on the model output. We define a distribution as \emph{ideal} if the minimizer of any cost-sensitive risk on it is guaranteed to have exact group-fair outcomes (e.g., demographic parity, equal opportunity)---in other words, it has no fairness-utility trade-off. We formulate an optimization program for optimal steering by finding the nearest \emph{ideal} distribution in KL-divergence, and provide efficient algorithms for it when the underlying distributions come from well-known parametric families (e.g., normal, log-normal). Empirically, our optimal steering techniques on both synthetic and real-world datasets improve fairness without diminishing utility (and sometimes even improve utility). We demonstrate affine steering of LLM representations to reduce bias in multi-class classification, e.g., occupation prediction from a short biography in Bios dataset (De-Arteaga et al.). Furthermore, we steer internal representations of LLMs towards desired outputs so that it works equally well across different groups.

The success of the convex-concave procedure (CCCP), a widely used technique for non-convex optimization, crucially depends on finding a decomposition of the objective function as a difference of convex functions (dcds). Despite the widespread applicability of CCCP, finding such dcds has attracted little attention in machine learning. For graphical models with polynomial potentials, existing methods for finding dcds require solving a Sum-of-Squares (SOS) program, which is often prohibitively expensive. In this work, we leverage tools from algebraic geometry certifying the positivity of polynomials, to derive LP-based constructions of dcds of polynomials which are particularly suited for graphical model inference. Our experiments demonstrate that using our LP-based technique constructs dcds for polynomial potentials of Markov random fields significantly faster compared to SOS-based approaches used in previous works.

The Separating Hyperplane theorem is a fundamental result in Convex Geometry with myriad applications. The theorem asserts that for a point a not in a closed convex set K, there is a hyperplane with K on one side and a strictly on the other side. Our first result, Random Separating Hyperplane Theorem (RSH), is a strengthening of this for polytopes. RSH asserts that if the distance between a and a polytope K with k vertices and unit diameter in ℜ^d is at least δ, where δ is a fixed constant in (0,1), then a randomly chosen hyperplane separates a and K with probability at least 1/poly(k) and margin at least Ω (δ/√d). RSH has algorithmic applications in learning polytopes. We consider a fundamental problem, denoted the "Hausdorff problem", of learning a unit diameter polytope K within Hausdorff distance δ, given an optimization oracle for K. Using RSH, we show that with polynomially many random queries to the optimization oracle, K can be approximated within error O(δ). To our knowledge, this is the first provable algorithm for the Hausdorff Problem in this setting. Building on this result, we show that if the vertices of K are well-separated, then an optimization oracle can be used to generate a list of points, each within distance O(δ) of K, with the property that the list contains a point close to each vertex of K. Further, we show how to prune this list to generate a (unique) approximation to each vertex of the polytope. We prove that in many latent variable settings, e.g., topic modeling, LDA, optimization oracles do exist provided we project to a suitable SVD subspace. Thus, our work yields the first efficient algorithm for finding approximations to the vertices of the latent polytope under the well-separatedness assumption. This assumption states that each vertex of K is far from the convex hull of the remaining vertices of K, and is much weaker than other assumptions behind algorithms in the literature which find vertices of the latent polytope.

Model editing is a growing area of research that is particularly valuable in contexts where modifying key model components, like neurons or filters, can significantly impact the model’s performance. The key challenge lies in identifying important components useful to the model’s predictions. We apply model editing to address two active areas of research, Structured Pruning, and Selective Class Forgetting. In this work, we adopt a distributional approach to the problem of identifying important components, leveraging the recently proposed discriminative filters hypothesis, which states that well-trained (convolutional) models possess discriminative filters that are essential to prediction. To do so, we define discriminative ability in terms of the Bayes error rate associated with the feature distributions, which is equivalent to computing the Total Variation (TV) distance between the distributions. However, computing the TV distance is intractable, motivating us to derive novel witness function-based lower bounds on the TV distance that require no assumptions on the underlying distributions; using this bound generalizes prior work such as Murti et al. [39] that relied on unrealistic Gaussianity assumptions on the feature distributions. With these bounds, we are able to discover critical subnetworks responsible for classwise predictions, and derive DISCEDIT-SP and DISCEDIT-U , algorithms for structured pruning requiring no access to the training data and loss function, and selective forgetting respectively. We apply DISCEDIT-U to selective class forgetting on models trained on CIFAR10 and CIFAR100, and we show that on average, we can reduce accuracy on a single class by over 80% with a minimal reduction in test accuracy on the remaining classes. Similarly, on Structured pruning problems, we obtain 40.8% sparsity on ResNet50 on Imagenet, with only a 2.6% drop in accuracy with minimal fine-tuning.

A fundamental problem in quantum many-body physics is that of finding ground states of local Hamiltonians. A number of recent works gave provably efficient machine learning (ML) algorithms for learning ground states. Specifically, [Huang et al. Science 2022], introduced an approach for learning properties of the ground state of an $n$-qubit gapped local Hamiltonian $H$ from only $n^{\mathcal{O}(1)}$ data points sampled from Hamiltonians in the same phase of matter. This was subsequently improved by [Lewis et al. Nature Communications 2024], to $\mathcal{O}(\log 𝑛)$ samples when the geometry of the $n$-qubit system is known. In this work, we introduce two approaches that achieve a constant sample complexity, independent of system size $n$, for learning ground state properties. Our first algorithm consists of a simple modification of the ML model used by Lewis et al. and applies to a property of interest known beforehand. Our second algorithm, which applies even if a description of the property is not known, is a deep neural network model. While empirical results showing the performance of neural networks have been demonstrated, to our knowledge, this is the first rigorous sample complexity bound on a neural network model for predicting ground state properties. We also perform numerical experiments that confirm the improved scaling of our approach compared to earlier results.

Deep neural networks (DNNs) have recently emerged as a powerful paradigm for solving Markovian optimal stopping problems. However, a ready extension of DNN-based methods to non-Markovian settings requires significant state and parameter space expansion, manifesting the curse of dimensionality. Further, efficient state-space transformations permitting Markovian approximations, such as those afforded by recurrent neural networks (RNNs), are either structurally infeasible or are confounded by the curse of non-Markovianity. Considering these issues, we introduce, for the first time, an optimal stopping policy gradient algorithm (OSPG) that can leverage RNNs effectively in non-Markovian settings by implicitly optimizing value functions without recursion, mitigating the curse of non-Markovianity. The OSPG algorithm is derived from an inference procedure on a novel Bayesian network representation of discrete-time non-Markovian optimal stopping trajectories and, as a consequence, yields an offline policy gradient algorithm that eliminates expensive Monte Carlo policy rollouts.

Knowledge transfer is shown to be a very successful technique for training neural classifiers: together with the ground truth data, it uses the "privileged information" (PI) obtained by a "teacher" network to train a "student" network. It has been observed that classifiers learn much faster and more reliably via knowledge transfer. However, there has been little or no theoretical analysis of this phenomenon. To bridge this gap, we propose to approach the problem of knowledge transfer by regularizing the fit between the teacher and the student with PI provided by the teacher. Using tools from dynamical systems theory, we show that when the student is an extremely wide two layer network, we can analyze it in the kernel regime and show that it is able to interpolate between PI and the given data. This characterization sheds new light on the relation between the training error and capacity of the student relative to the teacher. Another contribution of the paper is a quantitative statement on the convergence of student network. We prove that the teacher reduces the number of required iterations for a student to learn, and consequently improves the generalization power of the student. We give corresponding experimental analysis that validates the theoretical results and yield additional insights.

Providing a timely intervention before the onset of a critical event, such as a system failure, is of importance in many industrial settings. Before the onset of the critical event, systems typically exhibit behavioral changes which often manifest as stochastic co-variate observations which may be leveraged to trigger intervention. In this paper, for the first time, we formulate the problem of finding an optimally timed intervention (OTI) policy as minimizing the expected residual time to event, subject to a constraint on the probability of missing the event. Existing machine learning approaches to intervention on critical events focus on predicting event occurrence within a pre-defined window (a classification problem) or predicting time-to-event (a regression problem). Interventions are then triggered by setting model thresholds. These are heuristic-driven, lacking guarantees regarding optimality. To model the evolution of system behavior, we introduce the concept of a hazard rate process. We show that the OTI problem is equivalent to an optimal stopping problem on the associated hazard rate process. This key link has not been explored in literature. Under Markovian assumptions on the hazard rate process, we show that an OTI policy at any time can be analytically determined from the conditional hazard rate function at that time. Further, we show that our theory includes, as a special case, the important class of neural hazard rate processes generated by recurrent neural networks (RNNs). To model such processes, we propose a dynamic deep recurrent survival analysis (DDRSA) architecture, introducing an RNN encoder into the static DRSA setting. Finally, we demonstrate RNN-based OTI policies with experiments and show that they outperform popular intervention methods

Many algorithms for clustering high dimensional data assume that k, the number of clusters, is given. However, there has been little work on provably inferring k from the data. This paper gives polynomial time algorithms for finding k from the data assuming it satisfies certain natural deterministic conditions. Informally, we assume that there is a Ground Truth (GT) clustering of the data with the following properties: (i) Each cluster has a certain minimum size, (ii) the inter-mean separation of any two distinct clusters in the GT is large enough (although still weaker than what is typically assumed in the literature), and (iii) we define a novel “no large sub-cluster” (NLSC) property that characterizes the notion of a cluster by stipulating that there be no subsets of low “directional variance”. NLSC is indeed satisfied by large class of distributions including log-concave densities. The first major contribution is an algorithm for finding k where m, the minimum GT cluster size, is assumed to be known. This algorithm uses a novel rounding procedure which finds subsets of size m with low Directional Variance by rounding a SDP relaxation using Cheeger's inequality and it is shown that k is precisely the number of such sets whose means are well-separated. The harder problem of finding k when m not given is addressed by running the previous algorithm for each value of m to find candidate values of k and the corresponding k-clustering. The second major contribution of this paper is a test which certifies the correct candidate thereby yielding a polynomial time algorithm which finds k.

Accurate reconstruction of static environments from LiDAR scans of scenes containing dynamic objects, which we refer to as Dynamic to Static Translation (DST), is an important area of research in Autonomous Navigation. This problem has been recently explored for visual SLAM, but to the best of our knowledge no work has been attempted to address DST for LiDAR scans. The problem is of critical importance due to wide-spread adoption of LiDAR in Autonomous Vehicles. We show that state-of the art methods developed for the visual domain when adapted for LiDAR scans perform poorly. We develop DSLR, a deep generative model which learns a mapping between dynamic scan to its static counterpart through an adversarially trained autoencoder. Our model yields the first solution for DST on LiDAR that generates static scans without using explicit segmentation labels. DSLR cannot always be applied to real world data due to lack of paired dynamic-static scans. Using Unsupervised Domain Adaptation, we propose DSLR-UDA for transfer to real world data and experimentally show that this performs well in real world settings. Additionally, if segmentation information is available, we extend DSLR to DSLR-Seg to further improve the reconstruction quality. DSLR gives the state of the art performance on simulated and real-world datasets and also shows at least 4× improvement. We show that DSLR, unlike the existing baselines, is a practically viable model with its reconstruction quality within the tolerable limits for tasks pertaining to autonomous navigation like SLAM in dynamic environments.

Group-fairness in classification aims for equality of a predictive utility across different sensitive sub-populations, e.g., race or gender. Equality or near-equality constraints in group-fairness often worsen not only the aggregate utility but also the utility for the least advantaged sub-population. In this paper, we apply the principles of Pareto-efficiency and least-difference to the utility being accuracy, as an illustrative example, and arrive at the Rawls classifier that minimizes the error rate on the worst-off sensitive sub-population. Our mathematical characterization shows that the Rawls classifier uniformly applies a threshold to an ideal score of features, in the spirit of fair equality of opportunity. In practice, such a score or a feature representation is often computed by a black-box model that has been useful but unfair. Our second contribution is practical Rawlsian fair adaptation of any given black-box deep learning model, without changing the score or feature representation it computes. Given any score function or feature representation and only its second-order statistics on the sensitive sub-populations, we seek a threshold classifier on the given score or a linear threshold classifier on the given feature representation that achieves the Rawls error rate restricted to this hypothesis class. Our technical contribution is to formulate the above problems using ambiguous chance constraints, and to provide efficient algorithms for Rawlsian fair adaptation, along with provable upper bounds on the Rawls error rate. Our empirical results show significant improvement over state-of-the-art group-fair algorithms, even without retraining for fairness.

The recently introduced Latent $k-$ Polytope($\LkP$) encompasses several stochastic Mixed Membership models including Topic Models. The problem of finding $k$, the number of extreme points of $\LkP$, is a fundamental challenge and includes several important open problems such as determination of number of components in Ad-mixtures. This paper addresses this challenge by introducing Interpolative Convex Rank(\INR) of a matrix defined as the minimum number of its columns whose convex hull is within Hausdorff distance $\varepsilon$ of the convex hull of all columns. The first important contribution of this paper is to show that under \emph{standard assumptions} $k$ equals the \INR of a \emph{subset smoothed data matrix} defined from Data generated from an $\LkP$. The second important contribution of the paper is a polynomial time algorithm for finding $k$ under standard assumptions. An immediate corollary is the first polynomial time algorithm for finding the \emph{inner dimension} in Non-negative matrix factorisation(NMF) with assumptions which are qualitatively different than existing ones such as \emph{Separability}. %An immediate corollary is the first polynomial time algorithm for finding the \emph{inner dimension} in Non-negative matrix factorisation(NMF) with assumptions considerably weaker than \emph{Separability}.

Although Socially Assistive Robotics have been used in Autism Spectrum Disorder (ASD) interventions, such studies often exclude Special Educators (SEs) and often use expensive humanoid robots. In this paper, we investigate whether non-humanoid toy robots can act as teaching aids in ASD Education, in particular, can they reduce the workload of SEs. We target two most common yet divergent problems from Individualized Education Plans (IEPs) of ASD children - communication and gross motor skills. We present results from three studies a) toy robot Cozmo assists SEs in verbal lessons in school premises, b) mini drone Tello helps SEs in exercise lessons in school premises, and c) Cozmo, SEs, and ASD children connect remotely, as mandated due to the Covid-19 pandemic, for verbal lessons. All three studies showed improvement in learning outcomes and reduction in prompts from the SEs, denoting reduced workload. The effect of a robot's virtual presence in online ASD interventions has not been studied before. However, our results show that children spent more time on lessons in online intervention with Cozmo, suggesting that using robots should also be considered when designing online interventions. Furthermore, the roles of Cozmo were analyzed, and we found children showed increased spontaneous interaction when Cozmo acts as a Co-Instructor. Thus, preliminary results indicate toy robots, as opposed to expensive humanoids, may have significant potential in aiding SEs in Autism education.

Deriving Optimal bounds on Sample Complexity of Latent Variable models is an active area of research. Recently such bounds were obtained for Mixture of Gaussians \cite{HSNCAY18}, no such results are known for Ad-mixtures, a generalization of Mixture distributions. In this paper we show that $O^*(dk/m)$ samples are sufficient to learn each of $k-$ topic vectors of LDA, a popular Ad-mixture model, with vocabulary size $d$ and $m\in \Omega(1)$ words per document, to any constant error in $L_1$ norm. The result is a corollary of the major contribution of this paper: the first sample complexity upper bound for the problem (introduced in \cite{BK20}) of learning the vertices of a Latent $k-$ Polytope in $\RR^d$, given perturbed points from it. The bound, $O^*(dk/\beta)$, is optimal and linear in number of parameters. It applies to many stochastic models including a broad class Ad-mixtures. To demonstrate the generality of the approach we specialize the setting to Mixed Membership Stochastic Block Models(MMSB) and show for the first time that if an MMSB has $k$ blocks, the sample complexity is $O^*(k^2)$ under usual assumptions.

Subquadratic norms have been studied recently in the context of structured sparsity, which has been shown to be more beneficial than conventional regularizers in applications such as image denoising, compressed sensing, banded covariance estimation, etc. While existing works have been successful in learning structured sparse models such as trees, graphs, their associated optimization procedures have been inefficient because of hard-to-evaluate proximal operators of the norms. In this paper, we study the computational aspects of learning with subquadratic norms in a general setup. Our main contributions are two proximal-operator based algorithms ADMM-η and CP-η, which generically apply to these learning problems with convex loss functions, and achieve a proven rate of convergence of O(1/T) after T iterations. These algorithms are derived in a primal-dual framework, which have not been examined for subquadratic norms. We illustrate the efficiency of the algorithms developed in the context of tree-structured sparsity, where they comprehensively outperform relevant baselines.

In this paper we show that the learning problem for a large class of Latent variable models, such as Mixed Membership Stochastic Block Models, Topic Models, and Adversarial Clustering can be posed geometrically as follows: find a latent k— vertex simplex, K in Rd, given n data points, each obtained by perturbing a latent point in K. This problem does not seem to have been addressed. Our main contribution is an efficient algorithm for the geometric problem under deterministic assumptions which naturally hold for the models considered here. We observe that for a suitable r ≤ n, K is close to a data-determined polytope K’ (the subset smoothed, polytope) which is the convex hull of the points, each obtained by averaging an r subset of data points. Our algorithm is simply stated: it optimizes k carefully chosen linear functions over K’ to find the k vertices of the latent simplex. The proof of correctness is more involved, drawing on existing and new tools from Numerical Analysis. Our overall runtime of O* (k nnz) is as good as the best times of existing algorithms (modulo O* (1) factor) for the special cases and is better for sparse data which is the norm in Topic Modelling and Mixed Membership models. Some consequences of our algorithm are: Mixed Membership Models and Topic Models: We give the first quasi-input-sparsity time algorithm for parameter estimation for k ϵ O* (1) Adversarial Clustering: In k–means, an adversary is allowed to move many data points from each cluster towards the convex hull of other cluster centers. Our algorithm still estimates cluster centers well.

We present an unsupervised method to generate Word2Sense word embeddings that are interpretable -each dimension of the embedding space corresponds to a fine-grained sense, and the non-negative value of the embedding along the j-th dimension represents the relevance of the j-th sense to the word.The underlying LDA-based generative model can be extended to refine the representation of a polysemous word in a short context, allowing us to use the embeddings in contextual tasks.On computational NLP tasks, Word2Sense embeddings compare well with other word embeddings generated by unsupervised methods.Across tasks such as word similarity, entailment, sense induction, and contextual interpretation, Word2Sense is competitive with the state-of-the-art method for that task.Word2Sense embeddings are at least as sparse and fast to compute as prior art.

Word embeddings have been widely adopted across several NLP applications.Most existing word embedding methods utilize sequential context of a word to learn its embedding.While there have been some attempts at utilizing syntactic context of a word, such methods result in an explosion of the vocabulary size.In this paper, we overcome this problem by proposing SynGCN, a flexible Graph Convolution based method for learning word embeddings.SynGCN utilizes the dependency context of a word without increasing the vocabulary size.Word embeddings learned by SynGCN outperform existing methods on various intrinsic and extrinsic tasks and provide an advantage when used with ELMo.We also propose SemGCN, an effective framework for incorporating diverse semantic knowledge for further enhancing learned word representations.We make the source code of both models available to encourage reproducible research.

We consider the problem of minimizing a convex function over the intersection of finitely many simple sets which are easy to project onto. This is an important problem arising in various domains such as machine learning. The main difficulty lies in finding the projection of a point in the intersection of many sets. Existing approaches yield an infeasible point with an iteration-complexity of $O(1/ε^2)$ for nonsmooth problems with no guarantees on the in-feasibility. By reformulating the problem through exact penalty functions, we derive first-order algorithms which not only guarantees that the distance to the intersection is small but also improve the complexity to $O(1/ε)$ and $O(1/\sqrt{ε})$ for smooth functions. For composite and smooth problems, this is achieved through a saddle-point reformulation where the proximal operators required by the primal-dual algorithms can be computed in closed form. We illustrate the benefits of our approach on a graph transduction problem and on graph matching.

Distantly-supervised Relation Extraction (RE) methods train an extractor by automatically aligning relation instances in a Knowledge Base (KB) with unstructured text.In addition to relation instances, KBs often contain other relevant side information, such as aliases of relations (e.g., founded and co-founded are aliases for the relation founderOfCompany).RE models usually ignore such readily available side information.In this paper, we propose RESIDE, a distantly-supervised neural relation extraction method which utilizes additional side information from KBs for improved relation extraction.It uses entity type and relation alias information for imposing soft constraints while predicting relations.RE-SIDE employs Graph Convolution Networks (GCN) to encode syntactic information from text and improves performance even when limited side information is available.Through extensive experiments on benchmark datasets, we demonstrate RESIDE's effectiveness.We have made RESIDE's source code available to encourage reproducible research.

Understanding the representational power of Restricted Boltzmann Machines (RBMs) with multiple layers is an ill-understood problem and is an area of active research. Motivated from the approach of Inherent Structure formalism (Stillinger & Weber, 1982), extensively used in analysing Spin Glasses, we propose a novel measure called Inherent Structure Capacity (ISC), which characterizes the representation capacity of a fixed architecture RBM by the expected number of modes of distributions emanating from the RBM with parameters drawn from a prior distribution. Though ISC is intractable, we show that for a single layer RBM architecture ISC approaches a finite constant as number of hidden units are increased and to further improve the ISC, one needs to add a second layer. Furthermore, we introduce Lean RBMs, which are multi-layer RBMs where each layer can have at-most O(n) units with the number of visible units being n. We show that for every single layer RBM with Omega(n^{2+r}), r >= 0, hidden units there exists a two-layered lean RBM with Theta(n^2) parameters with the same ISC, establishing that 2 layer RBMs can achieve the same representational power as single-layer RBMs but using far fewer number of parameters. To the best of our knowledge, this is the first result which quantitatively establishes the need for layering.

The failure of LASSO to identify groups of correlated predictors in linear regression has sparked significant research interest. Recently, various norms were proposed, which can be best described as instances of ordered weighted $\ell_1$ norms (OWL), as an alternative to $\ell_1$ regularization used in LASSO. OWL can identify groups of correlated variables but it forces the model to be constant within a group. This artifact induces unnecessary bias in the model estimation. In this paper we take a submodular perspective and show that OWL can be posed as the Lovász extension of a suitably defined submodular function. The submodular perspective not only explains the group-wise constant behavior of OWL, but also suggests alternatives. The main contribution of this paper is smoothed OWL (SOWL), a new family of norms, which not only identifies the groups but also allows the model to be flexible inside a group. We establish several algorithmic and theoretical properties of SOWL including group identification and model consistency. We also provide algorithmic tools to compute the SOWL norm and its proximal operator, whose computational complexity $O(d\log d)$ is significantly better than that of general purpose solvers in $O(d^2\log d)$. In our experiments, SOWL compares favorably with respect to OWL in the regimes of interest.

Recommending lifestyle articles is of immediate interest to the e-commerce industry and is beginning to attract research attention. Often followed strategies, such as recommending popular items are inadequate for this vertical because of two reasons. Firstly, users have their own personal preference over items, referred to as personal styles, which lead to the long-tail phenomenon. Secondly, each user displays multiple personas, each persona has a preference over items which could be dictated by a particular occasion, e.g. dressing for a party would be different from dressing to go to office. Recommendation in this vertical is crucially dependent on discovering styles for each of the multiple personas. There is no literature which addresses this problem.

Standard clustering methods such as K-means, Gaussian mixture models, and hierarchical clustering are beset by local minima, which are sometimes drastically suboptimal. Moreover the number of clusters K must be known in advance. The recently introduced the sum-of-norms (SON) or Clusterpath convex relaxation of k-means and hierarchical clustering shrinks cluster centroids toward one another and ensure a unique global minimizer. We give a scalable stochastic incremental algorithm based on proximal iterations to solve the SON problem with convergence guarantees. We also show that the algorithm recovers clusters under quite general conditions which have a similar form to the unifying proximity condition introduced in the approximation algorithms community (that covers paradigm cases such as Gaussian mixtures and planted partition models). We give experimental results to confirm that our algorithm scales much better than previous methods while producing clusters of comparable quality.

A video is understood by users in terms of entities present in it. Entity Discovery is the task of building appearance model for each entity (e.g., a person), and finding all its occurrences in the video. We represent a video as a sequence of tracklets, each spanning 10-20 frames, and associated with one entity. We pose Entity Discovery as tracklet clustering, and approach it by leveraging Temporal Coherence (TC): the property that temporally neighboring tracklets are likely to be associated with the same entity. Our major contributions are the first Bayesian nonparametric models for TC at tracklet-level. We extend Chinese Restaurant Process (CRP) to TC-CRP, and further to Temporally Coherent Chinese Restaurant Franchise (TC-CRF) to jointly model entities and temporal segments using mixture components and sparse distributions. For discovering persons in TV serial videos without meta-data like scripts, these methods show considerable improvement over state-of-the-art approaches to tracklet clustering in terms of clustering accuracy, cluster purity and entity coverage. The proposed methods can perform online tracklet clustering on streaming videos unlike existing approaches, and can automatically reject false tracklets. Finally we discuss entity-driven video summarization- where temporal segments of the video are selected based on the discovered entities, to create a semantically meaningful summary.

Identifying translations from comparable corpora is a well-known problem with several applications. Existing methods rely on linguistic tools or high-quality corpora. Absence of such resources, especially in Indian languages, makes this problem hard; for example, state-of-the-art techniques achieve a mean reciprocal rank of 0.66 for English-Italian, and a mere 0.187 for Telugu-Kannada. In this work, we address the problem of comparable corpora-based translation correspondence induction (CC-TCI) when the only resources available are small noisy comparable corpora extracted from Wikipedia. We observe that translations in the source and target languages have many topically related words in common in other “auxiliary” languages. To model this, we define the notion of a translingual theme , a set of topically related words from auxiliary language corpora, and present a probabilistic framework for CC-TCI. Extensive experiments on 35 comparable corpora showed dramatic improvements in performance. We extend these ideas to propose a method for measuring cross-lingual semantic relatedness (CLSR) between words. To stimulate further research in this area, we make publicly available two new high-quality human-annotated datasets for CLSR. Experiments on the CLSR datasets show more than 200% improvement in correlation on the CLSR task. We apply the method to the real-world problem of cross-lingual Wikipedia title suggestion and build the WikiTSu system. A user study on WikiTSu shows a 20% improvement in the quality of titles suggested.

The Noisy Non-negative Matrix factorization (NMF) is: given a data matrix A (d x n), find non-negative matrices B;C (d x k, k x n respy.) so that A = BC +N, where N is a noise matrix. Existing polynomial time algorithms with proven error guarantees require EACH column N_⋅j to have l1 norm much smaller than ||(BC)_⋅j ||_1, which could be very restrictive. In important applications of NMF such as Topic Modeling as well as theoretical noise models (e.g. Gaussian with high sigma), almost EVERY column of N_.j violates this condition. We introduce the heavy noise model which only requires the average noise over large subsets of columns to be small. We initiate a study of Noisy NMF under the heavy noise model. We show that our noise model subsumes noise models of theoretical and practical interest (for e.g. Gaussian noise of maximum possible sigma). We then devise an algorithm TSVDNMF which under certain assumptions on B,C, solves the problem under heavy noise. Our error guarantees match those of previous algorithms. Our running time of O(k.(d+n)^2) is substantially better than the O(d.n^3) for the previous best. Our assumption on B is weaker than the “Separability” assumption made by all previous results. We provide empirical justification for our assumptions on C. We also provide the first proof of identifiability (uniqueness of B) for noisy NMF which is not based on separability and does not use hard to check geometric conditions. Our algorithm outperforms earlier polynomial time algorithms both in time and error, particularly in the presence of high noise.

Caching is an important determinant of storage system performance. Bulk cache preloading is the process of preloading large batches of relevant data into cache, minutes or hours in advance of actual requests by the application. We address bulk preloading by analyzing high-level spatio-temporal motifs from raw and noisy I/O traces by aggregating the trace into a temporal sequence of correlated count vectors. Such temporal multivariate data from trace aggregation arise from a diverse set of workloads leading to diverse data distributions with complex spatio-temporal dependencies. Motivated by this, we propose the Copula-HDP-HMM, a new Bayesian non-parametric modeling technique based on Gaussian Copula, suitable for temporal multivariate data with arbitrary marginals, avoiding limiting assumptions on the marginal distributions. We are not aware of prior work on copula based extensions of Bayesian non-parametric modeling algorithms for discrete data. Inference with copulas is hard when data is not continuous. We propose inference based on extended rank likelihood that circumvents specifying marginals, making our inference suitable for count data and even data with a combination of discrete and continuous marginals, enabling the use of Bayesian nonparametric modeling, for several data types, without assumptions on marginals. Finally, we propose HULK, a strategy for I/O efficient bulk cache preloading using our Copula-HDP-HMM model to leverage high-level spatio-temporal motifs in Block I/O traces. In experiments on benchmark traces, we show near perfect hitrate of 0.95 using HULK, a tremendous improvement over baseline using Multivariate Poisson, with only a fourth of I/O overhead.

MOTIVATION: T-cell epitopes serve as molecular keys to initiate adaptive immune responses. Identification of T-cell epitopes is also a key step in rational vaccine design. Most available methods are driven by informatics and are critically dependent on experimentally obtained training data. Analysis of a training set from Immune Epitope Database (IEDB) for several alleles indicates that the sampling of the peptide space is extremely sparse covering a tiny fraction of the possible nonamer space, and also heavily skewed, thus restricting the range of epitope prediction. RESULTS: We present a new epitope prediction method that has four distinct computational modules: (i) structural modelling, estimating statistical pair-potentials and constraint derivation, (ii) implicit modelling and interaction profiling, (iii) feature representation and binding affinity prediction and (iv) use of graphical models to extract peptide sequence signatures to predict epitopes for HLA class I alleles. CONCLUSIONS: HLaffy is a novel and efficient epitope prediction method that predicts epitopes for any Class-1 HLA allele, by estimating the binding strengths of peptide-HLA complexes which is achieved through learning pair-potentials important for peptide binding. It relies on the strength of the mechanistic understanding of peptide-HLA recognition and provides an estimate of the total ligand space for each allele. The performance of HLaffy is seen to be superior to the currently available methods. AVAILABILITY AND IMPLEMENTATION: The method is made accessible through a webserver http://proline.biochem.iisc.ernet.in/HLaffy CONTACT: : [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

This paper introduces ordered stick-breaking process (OSBP), where the atoms in a stick-breaking process (SBP) appear in order. The choice of weights on the atoms of OSBP ensure that; (1) probability of adding new atoms exponentially decrease, and (2) OSBP, though non-exchangeable, admit predictive probability functions (PPFs). In a Bayesian nonparametric (BNP) setting, OSBP serves as a natural prior over sequential mini-batches, facilitating exchange of relevant statistical information by sharing the atoms of OSBP. One of the major contributions of this paper is SUMO, an MCMC algorithm, for solving the inference problem arising from applying OSBP to BNP models. SUMO uses the PPFs of OSBP to obtain a Gibbs-sampling based truncation-free algorithm which applies generally to BNP models. For large scale inference problems existing algorithms such as particle filtering (PF) are not practical and variational procedures such as TSVI (Wang & Blei, 2012) are the only alternative. For Dirichlet process mixture model (DPMM), SUMO outperforms TSVI on perplexity by 33% on 3 datasets with million data points, which are beyond the scope of PF, using only 3GB RAM.

We introduce a unifying generalization of the Lovász theta function, and the associated geometric embedding, for graphs with weights on both nodes and edges. We show how it can be computed exactly by semidefinite programming, and how to approximate it using SVM computations. We show how the theta function can be interpreted as a measure of diversity in graphs and use this idea, and the graph embedding in algorithms for Max-Cut, correlation clustering and document summarization, all of which are well represented as problems on weighted graphs.

Recent literature~\cite{ando} suggests that embedding a graph on an unit sphere leads to better generalization for graph transduction. However, the choice of optimal embedding and an efficient algorithm to compute the same remains open. In this paper, we show that orthonormal representations, a class of unit-sphere graph embeddings are PAC learnable. Existing PAC-based analysis do not apply as the VC dimension of the function class is infinite. We propose an alternative PAC-based bound, which do not depend on the VC dimension of the underlying function class, but is related to the famous Lov\'{a}sz~$\vartheta$ function. The main contribution of the paper is SPORE, a SPectral regularized ORthonormal Embedding for graph transduction, derived from the PAC bound. SPORE is posed as a non-smooth convex function over an \emph{elliptope}. These problems are usually solved as semi-definite programs (SDPs) with time complexity $O(n^6)$. We present, Infeasible Inexact proximal~(IIP): an Inexact proximal method which performs subgradient procedure on an approximate projection, not necessarily feasible. IIP is more scalable than SDP, has an $O(\frac{1}{\sqrt{T}})$ convergence, and is generally applicable whenever a suitable approximate projection is available. We use IIP to compute SPORE where the approximate projection step is computed by FISTA, an accelerated gradient descent procedure. We show that the method has a convergence rate of $O(\frac{1}{\sqrt{T}})$. The proposed algorithm easily scales to 1000's of vertices, while the standard SDP computation does not scale beyond few hundred vertices. Furthermore, the analysis presented here easily extends to the multiple graph setting.

We consider the problem of recommending comment-worthy articles such as news and blog-posts. An article is defined to be comment-worthy for a particular user if that user is interested to leave a comment on it. We note that recommending comment-worthy articles calls for elicitation of commenting-interests of the user from the content of both the articles and the past comments made by users. We thus propose to develop content-driven user profiles to elicit these latent interests of users in commenting and use them to recommend articles for future commenting. The difficulty of modeling comment content and the varied nature of users' commenting interests make the problem technically challenging. The problem of recommending comment-worthy articles is resolved by leveraging article and comment content through topic modeling and the co-commenting pattern of users through collaborative filtering, combined within a novel hierarchical Bayesian modeling approach. Our solution, Collaborative Correspondence Topic Models (CCTM), generates user profiles which are leveraged to provide a personalized ranking of comment-worthy articles for each user. Through these content-driven user profiles, CCTM effectively handle the ubiquitous problem of cold-start without relying on additional meta-data. The inference problem for the model is intractable with no off-the-shelf solution and we develop an efficient Monte Carlo EM algorithm. CCTM is evaluated on three real world data-sets, crawled from two blogs, ArsTechnica (AT) Gadgets (102,087 comments) and AT-Science (71,640 comments), and a news site, DailyMail (33,500 comments). We show average improvement of 14% (warm-start) and 18% (cold-start) in AUC, and 80% (warm-start) and 250% (cold-start) in [email protected], over state of the art.

Tracklet Clustering is central to several Computer vision tasks [17][20]. A video can be represented as a sequence of tracklets, each spanning over 10–20 successive video frames, and each tracklet is associated with one entity (eg. person in case of TV-serial videos). Tracklets are instances of data-types exhibiting rich spatio-temporal structure. Existing approaches model tracklets by deploying detailed parametric models with a large number of parameters, making the inference unwieldy. The task of Person Discovery in long TV-series videos (40–45 minutes) with many persons can be naturally posed as tracklet clustering, and existing approaches give unsatisfactory performance on it. In this paper we attempt to leverage Temporal Coherence(TC) of videos to improve tracklet clustering. TC is the fundamental property of videos that each tracklet is likely to be associated with the same entity as its predecessor or successor. We propose the first Bayesian nonparametric approach for modelling TC, which can automatically infer the number of clusters to be formed. The major contribution of this paper is Temporally Coherent Chinese Restaurant Process (TC-CRP), which extends CRP by using TC. On the task of discovering persons in TV serials via tracklet clustering, without meta-data such as scripts, TC-CRP shows up to 25% improvement in cluster purity compared to state-of-the-art parametric models, and upto 36% improvement in number of persons discovered. We use a simple representation of tracklets: a vector of very generic features (like pixel intensity) which can correspond to any type of entity (not necessarily person), and empirically demonstrate the utility of TC-CRP for discovering entities like cars and planes. Moreover, unlike existing approaches TC-CRP can perform online tracklet clustering on streaming videos with very little performance deterioration, and can also automatically reject outliers (tracklets resulting from false detections).

Existing caching strategies, in the storage domain, though well suited to exploit short range spatio-temporal patterns, are unable to leverage long-range motifs for improving hitrates. Motivated by this, we investigate novel Bayesian non-parametric modeling(BNP) techniques for count vectors, to capture long range correlations for cache preloading, by mining Block I/O traces. Such traces comprise of a sequence of memory accesses that can be aggregated into high-dimensional sparse correlated count vector sequences. While there are several state of the art BNP algorithms for clustering and their temporal extensions for prediction, there has been no work on exploring these for correlated count vectors. Our first contribution addresses this gap by proposing a DP based mixture model of Multivariate Poisson (DP-MMVP) and its temporal extension(HMM-DP-MMVP) that captures the full covariance structure of multivariate count data. However, modeling full covariance structure for count vectors is computationally expensive, particularly for high dimensional data. Hence, we exploit sparsity in our count vectors, and as our main contribution, introduce the Sparse DP mixture of multivariate Poisson(Sparse-DP-MMVP), generalizing our DP-MMVP mixture model, also leading to more efficient inference. We then discuss a temporal extension to our model for cache preloading. We take the first step towards mining historical data, to capture long range patterns in storage traces for cache preloading. Experimentally, we show a dramatic improvement in hitrates on benchmark traces and lay the groundwork for further research in storage domain to reduce latencies using data mining techniques to capture long range motifs.

Applications of machine learning methods increasingly deal with graph structured data through kernels. Most existing graph kernels compare graphs in terms of features defined on small subgraphs such as walks, paths or graphlets, adopting an inherently local perspective. However, several interesting properties such as girth or chromatic number are global properties of the graph, and are not captured in local substructures. This paper presents two graph kernels defined on unlabeled graphs which capture global properties of graphs using the celebrated Lovász number and its associated orthonormal representation. We make progress towards theoretical results aiding kernel choice, proving a result about the separation margin of our kernel for classes of graphs. We give empirical results on classification of synthesized graphs with important global properties as well as established benchmark graph datasets, showing that the accuracy of our kernels is better than or competitive to existing graph kernels.

Understanding user generated comments in response to news and blog posts is an important area of research. After ignoring irrelevant comments, one finds that a large fraction, approximately 50%, of the comments are very specific and can be further related to certain parts of the article instead of the entire story. For example, in a recent product review of Google Nexus 7 in ArsTechnica (a popular blog), the reviewer talks about the prospect of "Retina equipped iPad mini" in a few sentences. It is interesting that although the article is on Nexus 7, but a significant number of comments are focused on this specific point regarding "iPad". We pose the problem of detecting such comments as specific comments location (SCL) problem. SCL is an important open problem with no prior work. SCL can be posed as a correspondence problem between comments and the parts of the relevant article, and one could potentially use Corr-LDA type models. Unfortunately, such models do not give satisfactory performance as they are restricted to using a single topic vector per article-comments pair. In this paper we propose to go beyond the single topic vector assumption and propose a novel correspondence topic model, namely SCTM, which admits multiple topic vectors (MTV) per article-comments pair. The resulting inference problem is quite complicated because of MTV and has no off-the-shelf solution. One of the major contributions of this paper is to show that using stick-breaking process as a prior over MTV, one can derive a collapsed Gibbs sampling procedure, which empirically works well for SCL.

Elasticity in cloud systems provides the flexibility to acquire and relinquish computing resources on demand. However, in current virtualized systems resource allocation is mostly static. Resources are allocated during VM instantiation and any change in workload leading to significant increase or decrease in resources is handled by VM migration. Hence, cloud users tend to characterize their workloads at a coarse grained level which potentially leads to under-utilized VM resources or under performing application. A more flexible and adaptive resource allocation mechanism would benefit variable workloads, such as those characterized by web servers. In this paper, we present an elastic resources framework for IaaS cloud layer that addresses this need. The framework provisions for application workload forecasting engine, that predicts at run-time the expected demand, which is input to the resource manager to modulate resource allocation based on the predicted demand. Based on the prediction errors, resources can be over-allocated or under-allocated as compared to the actual demand made by the application. Over-allocation leads to unused resources and under allocation could cause under performance. To strike a good trade-off between over-allocation and under-performance we derive an excess cost model. In this model excess resources allocated are captured as over-allocation cost and under-allocation is captured as a penalty cost for violating application service level agreement (SLA). Confidence interval for predicted workload is used to minimize this excess cost with minimal effect on SLA violations. An example case-study for an academic institute web server workload is presented. Using the confidence interval to minimize excess cost, we achieve significant reduction in resource allocation requirement while restricting application SLA violations to below 2-3%.

In a recent pioneering approach LDA was used to discover cross cutting concerns(CCC) automatically from software codebases. LDA though successful in detecting prominent concerns, fails to detect many useful CCCs including ones that may be heavily executed but elude discovery because they do not have a strong prevalence in source-code. We pose this problem as that of discovering topics that rarely occur in individual documents, which we will refer to as subtle topics. Recently an interesting approach, namely focused topic models(FTM) was proposed for detecting rare topics. FTM, though successful in detecting topics which occur prominently in very few documents, is unable to detect subtle topics. Discovering subtle topics thus remains an important open problem. To address this issue we propose subtle topic models(STM). STM uses a generalized stick breaking process(GSBP) as a prior for defining multiple distributions over topics. This hierarchical structure on topics allows STM to discover rare topics beyond the capabilities of FTM. The associated inference is non-standard and is solved by exploiting the relationship between GSBP and generalized Dirichlet distribution. Empirical results show that STM is able to discover subtle CCC in two benchmark code-bases, a feat which is beyond the scope of existing topic models, thus demonstrating the potential of the model in automated concern discovery, a known difficult problem in Software Engineering. Furthermore it is observed that even in general text corpora STM outperforms the state of art in discovering subtle topics.

Lovász introduced the theta function in his seminal paper [23] giving his celebrated solution to the problem of computing the Shannon capacity of the pentagon. Since then, the Lovász theta function has come to play a central role in information theory, graph theory and combinatorial optimization [11, 10], indeed Goemans [10] was led to remark: “it seems all paths lead to ϑ!”. The definition of the theta function also gives an elegant geometrical representation of the graph via an embedding in a spherical cap on the unit sphere which has many applications in graph theory and machine learning, some of them perhaps not yet fully appreciated. It is one of the goals of this paper to highlight how the Lovász embedding is a powerful and unifying tool in diverse graph theory and data mining applications.

We propose a novel space-time descriptor for region-based tracking which is very concise and efficient. The regions represented by covariance matrices within a temporal fragment, are used to estimate this space-time descriptor which we call the Eigenprofiles(EP). EP so obtained is used in estimating the Covariance Matrix of features over spatio-temporal fragments. The Second Order Statistics of spatio-temporal fragments form our target model which can be adapted for variations across the video. The model being concise also allows the use of multiple spatially overlapping fragments to represent the target. We demonstrate good tracking results on very challenging datasets, shot under insufficient illumination conditions.

We study the problem of analyzing influence of various factors affecting individual messages posted in social media. The problem is challenging because of various types of influences propagating through the social media network that act simultaneously on any user. Additionally, the topic composition of the influencing factors and the susceptibility of users to these influences evolve over time. This problem has not been studied before, and off-the-shelf models are unsuitable for this purpose. To capture the complex interplay of these various factors, we propose a new non-parametric model called the Dynamic Multi-Relational Chinese Restaurant Process. This accounts for the user network for data generation and also allows the parameters to evolve over time. Designing inference algorithms for this model suited for large scale social-media data is another challenge. To this end, we propose a scalable and multi-threaded inference algorithm based on online Gibbs Sampling. Extensive evaluations on large-scale Twitter and Face book data show that the extracted topics when applied to authorship and commenting prediction outperform state-of-the-art baselines. More importantly, our model produces valuable insights on topic trends and user personality trends beyond the capability of existing approaches.

With proliferation of chip multicores (CMPs) on desktops and embedded platforms, multi-threaded programs have become ubiquitous. Existence of multiple threads may cause resource contention, such as, in on-chip shared cache and interconnects, depending upon how they access resources. Hence, we propose a tool - Thread Contention Predictor (TCP) to help quantify the number of threads sharing data and their sharing pattern. We demonstrate its use to predict a more profitable shared, last level on-chip cache (LLC) access policy on CMPs. Our cache configuration predictor is 2.2 times faster compared to the cycle-accurate simulations. We also demonstrate its use for identifying hot data structures in a program which may cause performance degradation due to false data sharing. We fix layout of such data structures and show up-to 10% and 18% improvement in execution time and energy-delay product (EDP), respectively.

Comments constitute an important part of Web 2.0. In this paper, we consider comments on news articles. To simplify the task of relating the comment content to the article content the comments are about, we propose the idea of showing comments alongside article segments and explore automatic mapping of comments to article segments. This task is challenging because of the vocabulary mismatch between the articles and the comments. We present supervised and unsupervised techniques for aligning comments to segments the of article the comments are about. More specifically, we provide a novel formulation of supervised alignment problem using the framework of structured classification. Our experimental results show that structured classification model performs better than unsupervised matching and binary classification model.

Motivated by the task of vernacular news analysis using known news topics from national news-papers, we study the task of topic analysis, where given source datasets with observed topics, data items from a target dataset need to be assigned either to observed source topics or to new ones. Using Hierarchical Dirichlet Processes for addressing this task imposes unnecessary and often inappropriate generative assumptions on the observed source topics. In this paper, we explore Dirichlet Processes with partially observed groups (POG-DP). POG-DP avoids modeling the given source topics. Instead, it directly models the conditional distribution of the target data as a mixture of a Dirichlet Process and the posterior distribution of a Hierarchical Dirichlet Process with known groups and topics. This introduces coupling between selection probabilities of all topics within a source, leading to effective identification of source topics. We further improve on this with a Combinatorial Dirichlet Process with partially observed groups (POG-CDP) that captures finer grained coupling between related topics by choosing intersections between sources. We evaluate our models in three different real-world applications. Using extensive experimentation, we compare against several baselines to show that our model performs significantly better in all three applications.

Users can rarely reveal their information need in full detail to a search engine within 1--2 words, so search engines need to "hedge their bets" and present diverse results within the precious 10 response slots. Diversity in ranking is of much recent interest. Most existing solutions estimate the marginal utility of an item given a set of items already in the response, and then use variants of greedy set cover. Others design graphs with the items as nodes and choose diverse items based on visit rates (PageRank). Here we introduce a radically new and natural formulation of diversity as finding centers in resistive graphs. Unlike in PageRank, we do not specify the edge resistances (equivalently, conductances) and ask for node visit rates. Instead, we look for a sparse set of center nodes so that the effective conductance from the center to the rest of the graph has maximum entropy. We give a cogent semantic justification for turning PageRank thus on its head. In marked deviation from prior work, our edge resistances are learnt from training data. Inference and learning are NP-hard, but we give practical solutions. In extensive experiments with subtopic retrieval, social network search, and document summarization, our approach convincingly surpasses recently-published diversity algorithms like subtopic cover, max-marginal relevance (MMR), Grasshopper, DivRank, and SVMdiv.

Facet-based sentiment analysis involves discovering the latent facets, sentiments and their associations. Traditional facet-based sentiment analysis algorithms typically perform the various tasks in sequence, and fail to take advantage of the mutual reinforcement of the tasks. Additionally, inferring sentiment levels typically requires domain knowledge or human intervention. In this paper, we propose a series of probabilistic models that jointly discover latent facets and sentiment topics, and also order the sentiment topics with respect to a multi-point scale, in a language and domain independent manner. This is achieved by simultaneously capturing both short-range syntactic structure and long range semantic dependencies between the sentiment and facet words. The models further incorporate coherence in reviews, where reviewers dwell on one facet or sentiment level before moving on, for more accurate facet and sentiment discovery. For reviews which are supplemented with ratings, our models automatically order the latent sentiment topics, without requiring seed-words or domain-knowledge. To the best of our knowledge, our work is the first attempt to combine the notions of syntactic and semantic dependencies in the domain of review mining. Further, the concept of facet and sentiment coherence has not been explored earlier either. Extensive experimental results on real world review data show that the proposed models outperform various state of the art baselines for facet-based sentiment analysis.

The problem of identifying user intent has received considerable attention in recent years, particularly in the context of improving the search experience via query contextualization. Intent can be characterized by multiple dimensions, which are often not observed from query words alone. Accurate identification of Intent from query words remains a challenging problem primarily because it is extremely difficult to discover these dimensions. The problem is often significantly compounded due to lack of representative training sample. We present a generic, extensible framework for learning the multi-dimensional representation of user intent from the query words. The approach models the latent relationships between facets using tree structured distribution which leads to an efficient and convergent algorithm, FastQ, for identifying the multi-faceted intent of users based on just the query words. We also incorporated WordNet to extend the system capabilities to queries which contain words that do not appear in the training data. Empirical results show that FastQ yields accurate identification of intent when compared to a gold standard.

BACKGROUND: Temporal analysis of gene expression data has been limited to identifying genes whose expression varies with time and/or correlation between genes that have similar temporal profiles. Often, the methods do not consider the underlying network constraints that connect the genes. It is becoming increasingly evident that interactions change substantially with time. Thus far, there is no systematic method to relate the temporal changes in gene expression to the dynamics of interactions between them. Information on interaction dynamics would open up possibilities for discovering new mechanisms of regulation by providing valuable insight into identifying time-sensitive interactions as well as permit studies on the effect of a genetic perturbation. RESULTS: We present NETGEM, a tractable model rooted in Markov dynamics, for analyzing the dynamics of the interactions between proteins based on the dynamics of the expression changes of the genes that encode them. The model treats the interaction strengths as random variables which are modulated by suitable priors. This approach is necessitated by the extremely small sample size of the datasets, relative to the number of interactions. The model is amenable to a linear time algorithm for efficient inference. Using temporal gene expression data, NETGEM was successful in identifying (i) temporal interactions and determining their strength, (ii) functional categories of the actively interacting partners and (iii) dynamics of interactions in perturbed networks. CONCLUSIONS: NETGEM represents an optimal trade-off between model complexity and data requirement. It was able to deduce actively interacting genes and functional categories from temporal gene expression data. It permits inference by incorporating the information available in perturbed networks. Given that the inputs to NETGEM are only the network and the temporal variation of the nodes, this algorithm promises to have widespread applications, beyond biological systems.The source code for NETGEM is available from https://github.com/vjethava/NETGEM.

In this paper we consider the problem of learning an n x n Kernel matrix from m similarity matrices under general convex loss. Past research have extensively studied the m =1 case and have derived several algorithms which require sophisticated techniques like ACCP, SOCP, etc. The existing algorithms do not apply if one uses arbitrary losses and often can not handle m > 1 case. We present several provably convergent iterative algorithms, where each iteration requires either an SVM or a Multiple Kernel Learning (MKL) solver for m > 1 case. One of the major contributions of the paper is to extend the well known Mirror Descent(MD) framework to handle Cartesian product of psd matrices. This novel extension leads to an algorithm, called EMKL, which solves the problem in O(m^2 log n) iterations; in each iteration one solves an MKL involving m kernels and m eigen-decomposition of n x n matrices. By suitably defining a restriction on the objective function, a faster version of EMKL is proposed, called REKL, which avoids the eigen-decomposition. An alternative to both EMKL and REKL is also suggested which requires only an SVM solver. Experimental results on real world protein data set involving several similarity matrices illustrate the efficacy of the proposed algorithms.

Learning to rank is an important area at the interface of machine learning, information retrieval and Web search. The central challenge in optimizing various measures of ranking loss is that the objectives tend to be non-convex and discontinuous. To make such functions amenable to gradient based optimization procedures one needs to design clever bounds. In recent years, boosting, neural networks, support vector machines, and many other techniques have been applied. However, there is little work on directly modeling a conditional probability Pr(y|xq) where y is a permutation of the documents to be ranked and xqrepresents their feature vectors with respect to a query q. A major reason is that the space of y is huge: n! if n documents must be ranked. We first propose an intuitive and appealing expected loss minimization objective, and give an efficient shortcut to evaluate it despite the huge space of ys. Unfortunately, the optimization is non-convex, so we propose a convex approximation. We give a new, efficient Monte Carlo sampling method to compute the objective and gradient of this approximation, which can then be used in a quasi-Newton optimizer like LBFGS. Extensive experiments with the widely-used LETOR dataset show large ranking accuracy improvements beyond recent and competitive algorithms.

Detecting impending failure of hard disks is an important prediction task which might help computer systems to prevent loss of data and performance degradation. Currently most of the hard drive vendors support self-monitoring, analysis and reporting technology (SMART) which are often considered unreliable for such tasks. The problem of finding alternatives to SMART for predicting disk failure is an area of active research. In this paper, we consider events recorded from live disks and show that it is possible to construct decision support systems which can detect such failures. It is desired that any such prediction methodology should have high accuracy and ease of interpretability. Black box models can deliver highly accurate solutions but do not provide an understanding of events which explains the decision given by it. To this end we explore rule based classifiers for predicting hard disk failures from various disk events. We show that it is possible to learn easy to understand rules, from disk events, which have extremely low false alarm rates on real world data.

Motivated from real world problems, like object categorization, we study a particular mixed-norm regularization for Multiple Kernel Learning (MKL). It is assumed that the given set of kernels are grouped into distinct components where each component is crucial for the learning task at hand. The formulation hence employs $l_\infty$ regularization for promoting combinations at the component level and $l_1$ regularization for promoting sparsity among kernels in each component. While previous attempts have formulated this as a non-convex problem, the formulation given here is an instance of non-smooth convex optimization problem which admits an efficient Mirror-Descent (MD) based procedure. The MD procedure optimizes over product of simplexes, which is not a well-studied case in literature. Results on real-world datasets show that the new MKL formulation is well-suited for object categorization tasks and that the MD based algorithm outperforms state-of-the-art MKL solvers like \texttt{simpleMKL} in terms of computational effort.

Learning to rank from relevance judgment is an active research area. Itemwise score regression, pairwise preference satisfaction, and listwise structured learning are the major techniques in use. Listwise structured learning has been applied recently to optimize important non-decomposable ranking criteria like AUC (area under ROC curve) and MAP (mean average precision). We propose new, almost-linear-time algorithms to optimize for two other criteria widely used to evaluate search systems: MRR (mean reciprocal rank) and NDCG (normalized discounted cumulative gain) in the max-margin structured learning framework. We also demonstrate that, for different ranking criteria, one may need to use different feature maps. Search applications should not be optimized in favor of a single criterion, because they need to cater to a variety of queries. E.g., MRR is best for navigational queries, while NDCG is best for informational queries. A key contribution of this paper is to fold multiple ranking loss functions into a multi-criteria max-margin optimization. The result is a single, robust ranking model that is close to the best accuracy of learners trained on individual criteria. In fact, experiments over the popular LETOR and TREC data sets show that, contrary to conventional wisdom, a test criterion is often not best served by training with the same individual criterion.

We propose a novel, language-neutral approach for searching online handwritten text using Fr´echet distance. Online handwritten data, which is available as a time series (x,y,t), is treated as representing a parameterized curve in two-dimensions and the problem of searching online hand- written text is posed as a problem of matching two curves in a two-dimensional Euclidean space. Fr´echet distance is a natural measure for matching curves. The main contribu- tion of this paper is the formulation of a variant of Fr´echet distance that can be used for retrieving words even when only a prefix of the word is given as query. Extensive ex- periments on UNIPEN dataset1 consisting of over 16,000 words written by 7 users show that our method outperforms the state-of-the-art DTW method. Experiments were also conducted on a multilingual dataset, generated on a PDA, with encouraging results. Our approach can be used to implement useful, exciting features like auto-completion of handwriting in PDAs.

In this paper we propose a novel, scalable, clustering based Ordinal Regression formulation, which is an instance of a Second Order Cone Program (SOCP) with one Second Order Cone (SOC) constraint. The main contribution of the paper is a fast algorithm, CB-OR, which solves the proposed formulation more eficiently than general purpose solvers. Another main contribution of the paper is to pose the problem of focused crawling as a large scale Ordinal Regression problem and solve using the proposed CB-OR. Focused crawling is an efficient mechanism for discovering resources of interest on the web. Posing the problem of focused crawling as an Ordinal Regression problem avoids the need for a negative class and topic hierarchy, which are the main drawbacks of the existing focused crawling methods. Experiments on large synthetic and benchmark datasets show the scalability of CB-OR. Experiments also show that the proposed focused crawler outperforms the state-of-the-art.

Structural alignments are the most widely used tools for comparing proteins with low sequence similarity. The main contribution of this paper is to derive various kernels on proteins from structural alignments, which do not use sequence information. Central to the kernels is a novel alignment algorithm which matches substructures of fixed size using spectral graph matching techniques. We derive positive semi-definite kernels which capture the notion of similarity between substructures. Using these as base more sophisticated kernels on protein structures are proposed. To empirically evaluate the kernels we used a 40% sequence non-redundant structures from 15 different SCOP superfamilies. The kernels when used with SVMs show competitive performance with CE, a state of the art structure comparison program.

In this paper we propose a novel family of kernels for multivariate time-series classification problems. Each time-series is approximated by a linear combination of piecewise polynomial functions in a reproducing kernel Hilbert space by a novel kernel interpolation technique. Using the associated kernel function a large margin classification formulation is proposed which can discriminate between two classes. The formulation leads to kernels, between two multivariate time-series, which can be efficiently computed. The kernels have been successfully applied to writer independent handwritten character recognition.

We propose a randomized algorithm for large scale SVM learning which solves the problem by iterating over random subsets of the data. Crucial to the algorithm for scalability is the size of the subsets chosen. In the context of text classification we show that, by using ideas from random projections, a sample size of O(log n) can be used to obtain a solution which is close to the optimal with a high probability. Experiments done on synthetic and real life data sets demonstrate that the algorithm scales up SVM learners, without loss in accuracy.

This paper introduces kernels on attributed pointsets, which are sets of vectors embedded in an euclidean space. The embedding gives the notion of neighborhood, which is used to define positive semidefinite kernels on pointsets. Two novel kernels on neighborhoods are proposed, one evaluating the attribute similarity and the other evaluating shape similarity. Shape similarity function is motivated from spectral graph matching techniques. The kernels are tested on three real life applications: face recognition, photo album tagging, and shot annotation in video sequences, with encouraging results.

This paper addresses the problem of maximum margin classification given the moments of class conditional densities and the false positive and false negative error rates. Using Chebyshev inequalities, the problem can be posed as a second order cone programming problem. The dual of the formulation leads to a geometric optimization problem, that of computing the distance between two ellipsoids, which is solved by an iterative algorithm. The formulation is extended to non-linear classifiers using kernel methods. The resultant classifiers are applied to the case of classification of unbalanced datasets with asymmetric costs for misclassification. Experimental results on benchmark datasets show the efficacy of the proposed method.

BACKGROUND: Design of protein structure comparison algorithm is an important research issue, having far reaching implications. In this article, we describe a protein structure comparison scheme, which is capable of detecting correct alignments even in difficult cases, e.g. non-topological similarities. The proposed method computes protein structure alignments by comparing, small substructures, called neighborhoods. Two different types of neighborhoods, sequence and structure, are defined, and two algorithms arising out of the scheme are detailed. A new method for computing equivalences having non-topological similarities from pairwise similarity score is described. A novel and fast technique for comparing sequence neighborhoods is also developed. RESULTS: The experimental results show that the current programs show better performance on Fischer and Novotny's benchmark datasets, than state of the art programs, e.g. DALI, CE and SSM. Our programs were also found to calculate correct alignments for proteins with huge amount of indels and internal repeats. Finally, the sequence neighborhood based program was used in extensive fold and non-topological similarity detection experiments. The accuracy of the fold detection experiments with the new measure of similarity was found to be similar or better than that of the standard algorithm CE. CONCLUSION: A new scheme, resulting in two algorithms, have been developed, implemented and tested. The programs developed are accessible at http://mllab.csa.iisc.ernet.in/mp2/runprog.html.

This paper presents a novel Second Order Cone Programming (SOCP) formulation for large scale binary classification tasks. Assuming that the class conditional densities are mixture distributions, where each component of the mixture has a spherical covariance, the second order statistics of the components can be estimated efficiently using clustering algorithms like BIRCH. For each cluster, the second order moments are used to derive a second order cone constraint via a Chebyshev-Cantelli inequality. This constraint ensures that any data point in the cluster is classified correctly with a high probability. This leads to a large margin SOCP formulation whose size depends on the number of clusters rather than the number of training data points. Hence, the proposed formulation scales well for large datasets when compared to the state-of-the-art classifiers, Support Vector Machines (SVMs). Experiments on real world and synthetic datasets show that the proposed algorithm outperforms SVM solvers in terms of training time and achieves similar accuracies.

BACKGROUND: In recent times, there has been an exponential rise in the number of protein structures in databases e.g. PDB. So, design of fast algorithms capable of querying such databases is becoming an increasingly important research issue. This paper reports an algorithm, motivated from spectral graph matching techniques, for retrieving protein structures similar to a query structure from a large protein structure database. Each protein structure is specified by the 3D coordinates of residues of the protein. The algorithm is based on a novel characterization of the residues, called projections, leading to a similarity measure between the residues of the two proteins. This measure is exploited to efficiently compute the optimal equivalences. RESULTS: Experimental results show that, the current algorithm outperforms the state of the art on benchmark datasets in terms of speed without losing accuracy. Search results on SCOP 95% nonredundant database, for fold similarity with 5 proteins from different SCOP classes show that the current method performs competitively with the standard algorithm CE. The algorithm is also capable of detecting non-topological similarities between two proteins which is not possible with most of the state of the art tools like Dali.

We propose a novel second order cone programming formulation for designing robust classifiers which can handle uncertainty in observations. Similar formulations are also derived for designing regression functions which are robust to uncertainties in the regression setting. The proposed formulations are independent of the underlying distribution, requiring only the existence of second order moments. These formulations are then specialized to the case of missing values in observations for both classification and regression problems. Experiments show that the proposed formulations outperform imputation.

We propose a convex optimization based strategy to deal with uncertainty in the observations of a classification problem. We assume that instead of a sample (xi, yi) a distribution over (xi, yi) is specified. In particu- lar, we derive a robust formulation when the distribution is given by a normal distribution. It leads to Second Order Cone Programming formu- lation. Our method is applied to the problem of missing data, where it outperforms direct imputation.

When constructing a classifier, the probability of correct classifi(cid:173) cation of future data points should be maximized. In the current paper this desideratum is translated in a very direct way into an optimization problem, which is solved using methods from con(cid:173) vex optimization. We also show how to exploit Mercer kernels in this setting to obtain nonlinear decision boundaries. A worst-case bound on the probability of misclassification of future data is ob(cid:173) tained explicitly.

The chief aim of this paper is to propose mean-field approximations for a broad class of Belief networks, of which sigmoid and noisy-or networks can be seen as special cases. The approximations are based on a powerful mean-field theory suggested by Plefka. We show that Saul, Jaakkola and Jordan' s approach is the first order approximation in Plefka's approach, via a variational derivation. The application of Plefka's theory to belief networks is not computationally tractable. To tackle this problem we propose new approximations based on Taylor series. Small scale experiments show that the proposed schemes are attractive.

This article points out an important source of inefficiency in Platt's sequential minimal optimization (SMO) algorithm that is caused by the use of a single threshold value. Using clues from the KKT conditions for the dual problem, two threshold parameters are employed to derive modifications of SMO. These modified algorithms perform significantly faster than the original SMO on all benchmark data sets tried.

A variational derivation of Plefka's mean-field theory is presented. This theory is then applied to sigmoidal belief networks with the aid of further approximations. Empirical evaluation on small scale networks show that the proposed approximations are quite com(cid:173) petitive.

In this paper we give a new fast iterative algorithm for support vector machine (SVM) classifier design. The basic problem treated is one that does not allow classification violations. The problem is converted to a problem of computing the nearest point between two convex polytopes. The suitability of two classical nearest point algorithms, due to Gilbert, and Mitchell et al., is studied. Ideas from both these algorithms are combined and modified to derive our fast algorithm. For problems which require classification violations to be allowed, the violations are quadratically penalized and an idea due to Cortes and Vapnik and Friess is used to convert it to a problem in which there are no classification violations. Comparative computational evaluation of our algorithm against powerful SVM methods such as Platt's sequential minimal optimization shows that our algorithm is very competitive.

This paper points out an important source of inefficiency in Smola and Schölkopf's sequential minimal optimization (SMO) algorithm for support vector machine (SVM) regression that is caused by the use of a single threshold value. Using clues from the KKT conditions for the dual problem, two threshold parameters are employed to derive modifications of SMO for regression. These modified algorithms perform significantly faster than the original SMO on the datasets tried.