Searcharxiv⌕ Search

arXiv subjects

Aritra Ghosh

Publications and source records attributed to Aritra Ghosh.

At least 91 records · Page 5Linked to original sources
arXiv2021

DiPS: Differentiable Policy for Sketching in Recommender Systems

In sequential recommender system applications, it is important to develop models that can capture users' evolving interest over time to successfully recommend future items that they are likely to interact with. For users with long histories, typical models based on recurrent neural networks tend to forget important items in the distant past. Recent works have shown that storing a small sketch of past items can improve sequential recommendation tasks. However, these works all rely on static sketching policies, i.e., heuristics to select items to keep in the sketch, which are not necessarily optimal and cannot improve over time with more training data. In this paper, we propose a differentiable policy for sketching (DiPS), a framework that learns a data-driven sketching policy in an end-to-end manner together with the recommender system model to explicitly maximize recommendation quality in the future. We also propose an approximate estimator of the gradient for optimizing the sketching algorithm parameters that is computationally efficient. We verify the effectiveness of DiPS on real-world datasets under various practical settings and show that it requires up to $50\%$ fewer sketch items to reach the same predictive quality than existing sketching policies.

cs.IR↗
arXiv2021

Contact geometry and quantum thermodynamics of nanoscale steady states

We develop a geometric formalism suited for describing the quantum thermodynamics of a certain class of nanoscale systems (whose density matrix is expressible in the McLennan--Zubarev form) at any arbitrary non-equilibrium steady state. It is shown that the non-equilibrium steady states are points on control parameter spaces which are in a sense generated by the steady state Massieu--Planck function. By suitably altering the system's boundary conditions, it is possible to take the system from one steady state to another. We provide a contact Hamiltonian description of such transformations and show that moving along the geodesics of the friction tensor results in a minimum increase of the free entropy along the transformation. The control parameter space is shown to be equipped with a natural Riemannian metric that is compatible with the contact structure of the quantum thermodynamic phase space which when expressed in a local coordinate chart, coincides with the Schlögl metric. Finally, we show that this metric is conformally related to other thermodynamic Hessian metrics which might be written on control parameter spaces. This provides various alternate ways of computing the Schlögl metric which is known to be equivalent to the Fisher information matrix.

cond-mat.stat-mech↗
arXiv2021

BOBCAT: Bilevel Optimization-Based Computerized Adaptive Testing

Computerized adaptive testing (CAT) refers to a form of tests that are personalized to every student/test taker. CAT methods adaptively select the next most informative question/item for each student given their responses to previous questions, effectively reducing test length. Existing CAT methods use item response theory (IRT) models to relate student ability to their responses to questions and static question selection algorithms designed to reduce the ability estimation error as quickly as possible; therefore, these algorithms cannot improve by learning from large-scale student response data. In this paper, we propose BOBCAT, a Bilevel Optimization-Based framework for CAT to directly learn a data-driven question selection algorithm from training data. BOBCAT is agnostic to the underlying student response model and is computationally efficient during the adaptive testing process. Through extensive experiments on five real-world student response datasets, we show that BOBCAT outperforms existing CAT methods (sometimes significantly) at reducing test length.

cs.LG↗
arXiv2021

Contrastive Learning Improves Model Robustness Under Label Noise

Deep neural network-based classifiers trained with the categorical cross-entropy (CCE) loss are sensitive to label noise in the training data. One common type of method that can mitigate the impact of label noise can be viewed as supervised robust methods; one can simply replace the CCE loss with a loss that is robust to label noise, or re-weight training samples and down-weight those with higher loss values. Recently, another type of method using semi-supervised learning (SSL) has been proposed, which augments these supervised robust methods to exploit (possibly) noisy samples more effectively. Although supervised robust methods perform well across different data types, they have been shown to be inferior to the SSL methods on image classification tasks under label noise. Therefore, it remains to be seen that whether these supervised robust methods can also perform well if they can utilize the unlabeled samples more effectively. In this paper, we show that by initializing supervised robust methods using representations learned through contrastive learning leads to significantly improved performance under label noise. Surprisingly, even the simplest method (training a classifier with the CCE loss) can outperform the state-of-the-art SSL method by more than 50\% under high label noise when initialized with contrastive learning. Our implementation will be publicly available at {\url{https://github.com/arghosh/noisy_label_pretrain}}.

cs.CV↗
arXiv2021

Option Tracing: Beyond Correctness Analysis in Knowledge Tracing

Knowledge tracing refers to a family of methods that estimate each student's knowledge component/skill mastery level from their past responses to questions. One key limitation of most existing knowledge tracing methods is that they can only estimate an \emph{overall} knowledge level of a student per knowledge component/skill since they analyze only the (usually binary-valued) correctness of student responses. Therefore, it is hard to use them to diagnose specific student errors. In this paper, we extend existing knowledge tracing methods beyond correctness prediction to the task of predicting the exact option students select in multiple choice questions. We quantitatively evaluate the performance of our option tracing methods on two large-scale student response datasets. We also qualitatively evaluate their ability in identifying common student errors in the form of clusters of incorrect options across different questions that correspond to the same error.

cs.LG↗
arXiv2021

Do We Really Need Gold Samples for Sample Weighting Under Label Noise?

Learning with labels noise has gained significant traction recently due to the sensitivity of deep neural networks under label noise under common loss functions. Losses that are theoretically robust to label noise, however, often makes training difficult. Consequently, several recently proposed methods, such as Meta-Weight-Net (MW-Net), use a small number of unbiased, clean samples to learn a weighting function that downweights samples that are likely to have corrupted labels under the meta-learning framework. However, obtaining such a set of clean samples is not always feasible in practice. In this paper, we analytically show that one can easily train MW-Net without access to clean samples simply by using a loss function that is robust to label noise, such as mean absolute error, as the meta objective to train the weighting network. We experimentally show that our method beats all existing methods that do not use clean samples and performs on-par with methods that use gold samples on benchmark datasets across various noise types and noise rates.

cs.LG↗
arXiv2021

Thermodynamic curvature of AdS black holes with dark energy

In this paper, we study the effect of dark energy on the extended thermodynamic structure and interacting microstructures of black holes in AdS, through an analysis of thermodynamic geometry. Considering various limiting cases of the novel equation of state obtained in charged rotating black holes with quintessence, and taking enthalpy $H$ as the key potential in the extended phase space, we scrutinize the behavior of the Ruppeiner curvature scalar $R$ in the entropy-pressure $(S,P)$-plane (or equivalently in the temperature-volume ($T,V$)-plane). Analysis of $R$ empirically reveals that dark energy parameterized by $α$, significantly alters the dominant interactions of neutral, charged and slowly rotating black hole microstructures. In the Schwarzschild-AdS case: black holes smaller than a certain size continue to have attractive interactions whereas larger black holes are completely dominated by repulsive interactions which arise to due dark energy. For charged or rotating AdS black holes with quintessence, $R$ can change sign at multiple points depending upon the relation between $α$ and charge $q$ or angular momentum $J$. In particular, above a threshold value of $α$, $R$ is never negative at all, suggesting heuristically that the repulsive interactions due to quintessence are long ranged as opposed to the previously known short ranged repulsion in charged AdS black holes. A mean field interaction potential is proposed whose extrema effectively capture the points where the curvature $R$ changes sign.

gr-qc↗
arXiv2020

Thermodynamic curvature of the Schwarzschild-AdS black hole and Bose condensation

In the AdS/CFT correspondence, a dynamical cosmological constant $Λ$ in the bulk corresponds to varying the number of colors $N$ in the boundary gauge theory with a chemical potential $μ$ as its thermodynamic conjugate. In this work, within the context of Schwarzschild black holes in $AdS_5 \times S^5$ and its dual finite temperature $\mathcal{N}=4$ superconformal Yang-Mills theory at large $N$, we investigate thermodynamic geometry through the behavior of the Ruppeiner scalar $R$. The sign of $R$ is an empirical indicator of the nature of microscopic interactions and is found to be negative for the large black hole branch implying that its thermodynamic characteristics bear qualitative similarities with that of an attraction dominated system, such as an ideal gas of bosons. We find that as the system's fugacity approaches unity, $R$ takes increasingly negative values signifying long range correlations and strong quantum fluctuations signaling the onset of Bose condensation. On the other hand, $R$ for the small black hole branch is negative at low temperatures and positive at high temperatures with a second order critical point which roughly separates the two regions.

hep-th↗
arXiv2020

Context-Aware Attentive Knowledge Tracing

Knowledge tracing (KT) refers to the problem of predicting future learner performance given their past performance in educational applications. Recent developments in KT using flexible deep neural network-based models excel at this task. However, these models often offer limited interpretability, thus making them insufficient for personalized learning, which requires using interpretable feedback and actionable recommendations to help learners achieve better learning outcomes. In this paper, we propose attentive knowledge tracing (AKT), which couples flexible attention-based neural network models with a series of novel, interpretable model components inspired by cognitive and psychometric models. AKT uses a novel monotonic attention mechanism that relates a learner's future responses to assessment questions to their past responses; attention weights are computed using exponential decay and a context-aware relative distance measure, in addition to the similarity between questions. Moreover, we use the Rasch model to regularize the concept and question embeddings; these embeddings are able to capture individual differences among questions on the same concept without using an excessive number of parameters. We conduct experiments on several real-world benchmark datasets and show that AKT outperforms existing KT methods (by up to $6\%$ in AUC in some cases) on predicting future learner responses. We also conduct several case studies and show that AKT exhibits excellent interpretability and thus has potential for automated feedback and personalization in real-world educational settings.

cs.LG↗
arXiv2020

Galaxy Morphology Network: A Convolutional Neural Network Used to Study Morphology and Quenching in $\sim 100,000$ SDSS and $\sim 20,000$ CANDELS Galaxies

We examine morphology-separated color-mass diagrams to study the quenching of star formation in $\sim 100,000$ ($z\sim0$) Sloan Digital Sky Survey (SDSS) and $\sim 20,000$ ($z\sim1$) Cosmic Assembly Near-Infrared Deep Extragalactic Legacy Survey (CANDELS) galaxies. To classify galaxies morphologically, we developed Galaxy Morphology Network (GaMorNet), a convolutional neural network that classifies galaxies according to their bulge-to-total light ratio. GaMorNet does not need a large training set of real data and can be applied to data sets with a range of signal-to-noise ratios and spatial resolutions. GaMorNet's source code as well as the trained models are made public as part of this work ( http://www.astro.yale.edu/aghosh/gamornet.html ). We first trained GaMorNet on simulations of galaxies with a bulge and a disk component and then transfer learned using $\sim25\%$ of each data set to achieve misclassification rates of $\lesssim5\%$. The misclassified sample of galaxies is dominated by small galaxies with low signal-to-noise ratios. Using the GaMorNet classifications, we find that bulge- and disk-dominated galaxies have distinct color-mass diagrams, in agreement with previous studies. For both SDSS and CANDELS galaxies, disk-dominated galaxies peak in the blue cloud, across a broad range of masses, consistent with the slow exhaustion of star-forming gas with no rapid quenching. A small population of red disks is found at high mass ($\sim14\%$ of disks at $z\sim0$ and $2\%$ of disks at $z \sim 1$). In contrast, bulge-dominated galaxies are mostly red, with much smaller numbers down toward the blue cloud, suggesting rapid quenching and fast evolution across the green valley. This inferred difference in quenching mechanism is in agreement with previous studies that used other morphology classification techniques on much smaller samples at $z\sim0$ and $z\sim1$.

astro-ph.GA↗
arXiv2020

Thermodynamic geometry for charged Gauss-Bonnet black holes in AdS spacetimes

In this paper, we study the thermodynamic geometry of charged Gauss-Bonnet black holes (and Reissner-Nordström black holes, for the sake of comparison) in AdS: in both $(T,V)$- and $(S,P)$-planes. The thermodynamic phase space is known to have an underlying contact and metric structure; Ruppeiner geometry then naturally arises in this framework. Sign of Ruppeiner curvature can be used to probe the nature of interactions between the black hole microstructures. It is found that there are both attraction and repulsion dominated regions which are in general determined by the electric charge, Gauss-Bonnet coupling and horizon radius of the black hole. The results are physically explained by considering that these black hole systems consist of charged as well as neutral microstructures much like a binary mixture of fluids.

gr-qc↗
arXiv2020

Thermodynamic geometry and interacting microstructures of BTZ black holes

In this work, we present a study to probe the nature of interactions between black hole microstructures for the case of the BTZ black holes. Even though BTZ black holes without any angular momentum or electric charge thermodynamically behave as an ideal gas, i.e. with non-interacting microstructures; in the presence of electric charge or angular momentum, BTZ black holes are associated with repulsive interactions among the microstructures. We extend the study to the case of exotic BTZ black holes with mass $M = αm + γ\frac{j}{l}$ and angular momentum $J=αj + γl m$, for arbitrary values of $ (α, γ)$ ranging from purely exotic $(α=0,γ=1)$, slightly exotic $(α> \frac{1}{2},γ< \frac{1}{2})$ and highly exotic $(α< \frac{1}{2}, γ> \frac{1}{2})$. We find that unlike the normal BTZ black holes (the case $α=1,γ=0$), there exist both attraction as well as repulsion dominated regions in all the cases of exotic BTZ black holes.

hep-th↗
arXiv2020

Optimal Bidding Strategy without Exploration in Real-time Bidding

Maximizing utility with a budget constraint is the primary goal for advertisers in real-time bidding (RTB) systems. The policy maximizing the utility is referred to as the optimal bidding strategy. Earlier works on optimal bidding strategy apply model-based batch reinforcement learning methods which can not generalize to unknown budget and time constraint. Further, the advertiser observes a censored market price which makes direct evaluation infeasible on batch test datasets. Previous works ignore the losing auctions to alleviate the difficulty with censored states; thus significantly modifying the test distribution. We address the challenge of lacking a clear evaluation procedure as well as the error propagated through batch reinforcement learning methods in RTB systems. We exploit two conditional independence structures in the sequential bidding process that allow us to propose a novel practical framework using the maximum entropy principle to imitate the behavior of the true distribution observed in real-time traffic. Moreover, the framework allows us to train a model that can generalize to the unseen budget conditions than limit only to those observed in history. We compare our methods on two real-world RTB datasets with several baselines and demonstrate significantly improved performance under various budget settings.

cs.LG↗
arXiv2020

Scalable Bid Landscape Forecasting in Real-time Bidding

In programmatic advertising, ad slots are usually sold using second-price (SP) auctions in real-time. The highest bidding advertiser wins but pays only the second-highest bid (known as the winning price). In SP, for a single item, the dominant strategy of each bidder is to bid the true value from the bidder's perspective. However, in a practical setting, with budget constraints, bidding the true value is a sub-optimal strategy. Hence, to devise an optimal bidding strategy, it is of utmost importance to learn the winning price distribution accurately. Moreover, a demand-side platform (DSP), which bids on behalf of advertisers, observes the winning price if it wins the auction. For losing auctions, DSPs can only treat its bidding price as the lower bound for the unknown winning price. In literature, typically censored regression is used to model such partially observed data. A common assumption in censored regression is that the winning price is drawn from a fixed variance (homoscedastic) uni-modal distribution (most often Gaussian). However, in reality, these assumptions are often violated. We relax these assumptions and propose a heteroscedastic fully parametric censored regression approach, as well as a mixture density censored network. Our approach not only generalizes censored regression but also provides flexibility to model arbitrarily distributed real-world data. Experimental evaluation on the publicly available dataset for winning price estimation demonstrates the effectiveness of our method. Furthermore, we evaluate our algorithm on one of the largest demand-side platforms and significant improvement has been achieved in comparison with the baseline solutions.

cs.LG↗
arXiv2019

Contact geometry and thermodynamics of black holes in AdS spacetimes

In this paper we discuss a formulation of extended phase space thermodynamics of black holes in Anti de Sitter (AdS) spacetimes from the contact geometry point of view. Thermodynamics of black holes can be understood within the framework of contact geometry as flows of vector fields generated by Hamiltonian functions on equilibrium submanifolds in the extended phase space that naturally incorporates the structure of a contact manifold. Deformations induced by the contact vector fields are used to construct various maps among thermodynamic quantities. Thermodynamic variables and equations of state of Schwarzschild black holes are mapped to that of Reissner-Nordström black holes in AdS, with charge as the deformation parameter. In addition, the equations of state of general black holes in AdS are shown to emerge from the high-temperature ideal gas limit equations via suitable deformations induced by contact vector fields. The Hamilton-Jacobi formalism analogous to mechanics is set up, and the corresponding characteristic curves of contact vector fields are explicitly obtained to model thermodynamic processes of black holes. Extension to thermodynamic cycles in this framework is also discussed.

hep-th↗
arXiv2019

On purely nonlinear oscillators generalizing an isotonic potential

In this paper we consider a nonlinear generalization of the isotonic oscillator in the same spirit as one considers the generalization of the harmonic oscillator with a truly nonlinear restoring force. The corresponding potential being asymmetric we invoke the symmetrization principle and construct a symmetric potential in which the period function has the same value as in the original asymmetric potential. The period function is amplitude dependent and expressible in terms of the hypergeometric function and reduces to $2π$ when $α=1$, i.e., corresponding to the special case of an isotonic oscillator.

physics.class-ph↗
arXiv2019

Action-angle variables for the purely nonlinear oscillator

In this letter, we study the purely nonlinear oscillator by the method of action-angle variables of Hamiltonian systems. The frequency of the non-isochronous system is obtained, which agrees well with the previously known result. Exact analytic solutions of the system involving generalized trigonometric functions are presented. We also present arguments to show the adiabatic invariance of the action variable for a time-dependent purely nonlinear oscillator.

physics.class-ph↗