Searcharxiv⌕ Search

arXiv subjects

Swati Gupta

Publications and source records attributed to Swati Gupta.

45 records · Page 3Linked to original sources
arXiv2019

Robust Classification using Robust Feature Augmentation

Existing deep neural networks, say for image classification, have been shown to be vulnerable to adversarial images that can cause a DNN misclassification, without any perceptible change to an image. In this work, we propose shock absorbing robust features such as binarization, e.g., rounding, and group extraction, e.g., color or shape, to augment the classification pipeline, resulting in more robust classifiers. Experimentally, we show that augmenting ML models with these techniques leads to improved overall robustness on adversarial inputs as well as significant improvements in training time. On the MNIST dataset, we achieved 14x speedup in training time to obtain 90% adversarial accuracy com-pared to the state-of-the-art adversarial training method of Madry et al., as well as retained higher adversarial accuracy over a broader range of attacks. We also find robustness improvements on traffic sign classification using robust feature augmentation. Finally, we give theoretical insights for why one can expect robust feature augmentation to reduce adversarial input space

cs.LG↗
arXiv2019

Individual Fairness in Hindsight

Since many critical decisions impacting human lives are increasingly being made by algorithms, it is important to ensure that the treatment of individuals under such algorithms is demonstrably fair under reasonable notions of fairness. One compelling notion proposed in the literature is that of individual fairness (IF), which advocates that similar individuals should be treated similarly (Dwork et al. 2012). Originally proposed for offline decisions, this notion does not, however, account for temporal considerations relevant for online decision-making. In this paper, we extend the notion of IF to account for the time at which a decision is made, in settings where there exists a notion of conduciveness of decisions as perceived by the affected individuals. We introduce two definitions: (i) fairness-across-time (FT) and (ii) fairness-in-hindsight (FH). FT is the simplest temporal extension of IF where treatment of individuals is required to be individually fair relative to the past as well as future, while in FH, we require a one-sided notion of individual fairness that is defined relative to only the past decisions. We show that these two definitions can have drastically different implications in the setting where the principal needs to learn the utility model. Linear regret relative to optimal individually fair decisions is inevitable under FT for non-trivial examples. On the other hand, we design a new algorithm: Cautious Fair Exploration (CaFE), which satisfies FH and achieves sub-linear regret guarantees for a broad range of settings. We characterize lower bounds showing that these guarantees are order-optimal in the worst case. FH can thus be embedded as a primary safeguard against unfair discrimination in algorithmic deployments, without hindering the ability to take good decisions in the long-run.

cs.LG↗
arXiv2018

Limited Memory Kelley's Method Converges for Composite Convex and Submodular Objectives

The original simplicial method (OSM), a variant of the classic Kelley's cutting plane method, has been shown to converge to the minimizer of a composite convex and submodular objective, though no rate of convergence for this method was known. Moreover, OSM is required to solve subproblems in each iteration whose size grows linearly in the number of iterations. We propose a limited memory version of Kelley's method (L-KM) and os OSM that requires limited memory (at most n + 1 constraints for an n-dimensional problem) independent of the iteration. We prove convergence for L-KM when the convex part of the objective (g) is strongly convex and show it converges linearly when g is also smooth. Our analysis relies on duality between minimization of the composite objective and minimization of a convex function over the corresponding submodular base polytope. We introduce a limited memory version, L-FCFW, of the Fully-Corrective Frank-Wolfe (FCFW) method with approximate correction, to solve the dual problem. We show that L-FCFW and L-KM are dual algorithms that produce the same sequence of iterates; hence both converge linearly (when g is smooth and strongly convex) and with limited memory. We propose L-KM to minimize composite convex and submodular objectives; however, our results on L-FCFW hold for general polytopes and may be of independent interest.

math.OC↗
arXiv2018

Robust Look-ahead Three-phase Balancing of Uncertain Distribution Loads

Increasing penetration of highly variable components such as solar generation and electric vehicle charging loads pose significant challenges to keeping three-phase loads balanced in modern distribution systems. Failure to maintain balance across three phases would lead to asset deterioration and increasing delivery losses. Motivated by the real-world needs to automate and optimize the three-phase balancing decision making, this paper introduces a robust look-ahead optimization framework that pursues balanced phases in the presence of demand-side uncertainties. We show that look-ahead moving window optimization can reduce imbalances among phases at the cost of a limited number of phase swapping operations. Case studies quantify the improvements of the proposed methods compared with conventional deterministic phase balancing. Discussions on possible benefits of the proposed methods and extensions are presented.

math.OC↗
arXiv2016

E3 : Keyphrase based News Event Exploration Engine

This paper presents a novel system E3 for extracting keyphrases from news content for the purpose of offering the news audience a broad overview of news events, with especially high content volume. Given an input query, E3 extracts keyphrases and enrich them by tagging, ranking and finding role for frequently associated keyphrases. Also, E3 finds the novelty and activeness of keyphrases using news publication date, to identify the most interesting and informative keyphrases.

cs.IR↗
arXiv2016

Solving Combinatorial Games using Products, Projections and Lexicographically Optimal Bases

In order to find Nash-equilibria for two-player zero-sum games where each player plays combinatorial objects like spanning trees, matchings etc, we consider two online learning algorithms: the online mirror descent (OMD) algorithm and the multiplicative weights update (MWU) algorithm. The OMD algorithm requires the computation of a certain Bregman projection, that has closed form solutions for simple convex sets like the Euclidean ball or the simplex. However, for general polyhedra one often needs to exploit the general machinery of convex optimization. We give a novel primal-style algorithm for computing Bregman projections on the base polytopes of polymatroids. Next, in the case of the MWU algorithm, although it scales logarithmically in the number of pure strategies or experts $N$ in terms of regret, the algorithm takes time polynomial in $N$; this especially becomes a problem when learning combinatorial objects. We give a general recipe to simulate the multiplicative weights update algorithm in time polynomial in their natural dimension. This is useful whenever there exists a polynomial time generalized counting oracle (even if approximate) over these objects. Finally, using the combinatorial structure of symmetric Nash-equilibria (SNE) when both players play bases of matroids, we show that these can be found with a single projection or convex minimization (without using online learning).

cs.LG↗
arXiv2006

A Time Dependent Leptonic Model for Microquasar Jets: Application to LSI 61 303

The Galactic high-mass X-ray binary and jet source (microquasar) LSI +61 303 has recently been detected at TeV gamma-ray energies by the MAGIC telescope. We have applied a time-dependent leptonic jet model to the broadband spectral energy distribution and suggested (though not unambiguously detected) orbital modulation of the very high energy gamma-ray emission of this source. Our model takes into account time dependent electron injection and acceleration, and the adiabatic and radiative cooling of non-thermal electrons. It includes synchrotron, synchrotron self-Compton and external inverse Compton (with seed photons from the companion star), as well as gamma-gamma absorption of gamma-rays by starlight photons. The model can successfully reproduce the available multiwavelength observational data. Our best fit to the SED indicates that a magnetic field of B_0 ~ 5 X 10^3 G at ~ 10^3 R_g is required, and electrons need to be accelerated out to TeV energies (gamma_2 = 10^6) with a nonthermal injection spectrum with a spectral index of q = 1.7, indicating the operation of acceleration mechanisms beyond the standard first-order Fermi mechanism at relativistic or non-relativistic shocks. The orbital modulation of the VHE gamma-ray emission can be explained solely by the geometrical effect of changes in the relative orientation of the stellar companion with respect to the compact object and jet as it impacts the position and depth of the gamma-gamma absorption trough. Such a scenario predicts a trend of spectral hardening during VHE gamma-ray low orbital phases.

astro-ph↗
arXiv2006

Time-Dependent Synchrotron and Compton Spectra from Jets of Microquasars

Jet models for the high-energy emission of Galactic X-ray binary sources have regained significant interest with detailed spectral and timing studies of the X-ray emission from microquasars, the recent detection by the HESS collaboration of very-high-energy gamma-rays from the microquasar LS~5039, and the earlier suggestion of jet models for ultraluminous X-ray sources observed in many nearby galaxies. Here we study the synchrotron and Compton signatures of time-dependent electron injection and acceleration, adiabatic and radiative cooling, and different jet geometries in the jets of Galactic microquasars. Synchrotron, synchrotron-self-Compton, and external-Compton radiation processes with soft photons provided by the companion star and the accretion disk are treated. An analytical solution is presented to the electron kinetic equation for general power-law geometries of the jets for Compton scattering in the Thomson regime. We pay particular attention to predictions concerning the rapid flux and spectral variability signatures expected in a variety of scenarios, making specific predictions concerning possible spectral hysteresis, similar to what has been observed in several TeV blazars. Such predictions should be testable with dedicated monitoring observations of Galactic microquasars and ultraluminous X-ray sources using Chandra and/or XMM-Newton.

astro-ph↗