Results for "Aayush Bansal"

total 522took 0.12s
Recycle-GAN: Unsupervised Video RetargetingAug 15 2018We introduce a data-driven approach for unsupervised video retargeting that translates content from one domain to another while preserving the style native to a domain, i.e., if contents of John Oliver's speech were to be transferred to Stephen Colbert, ... More
Shapes and Context: In-the-Wild Image Synthesis & ManipulationJun 11 2019We introduce a data-driven approach for interactively synthesizing in-the-wild images from semantic label maps. Our approach is dramatically different from recent work in this space, in that we make use of no learning. Instead, our approach uses simple ... More
PixelNN: Example-based Image SynthesisAug 17 2017We present a simple nearest-neighbor (NN) approach that synthesizes high-frequency photorealistic images from an "incomplete" signal such as a low-resolution image, a surface normal map, or edges. Current state-of-the-art deep generative models designed ... More
Marr Revisited: 2D-3D Alignment via Surface Normal PredictionApr 05 2016We introduce an approach that leverages surface normal predictions, along with appearance cues, to retrieve 3D models for objects depicted in 2D still images from a large CAD object library. Critical to the success of our approach is the ability to recover ... More
Circuit Complexity of Bounded Planar Cutwidth Graph MatchingJan 03 2018Recently, perfect matching in bounded planar cutwidth bipartite graphs (\BGGM) was shown to be in ACC$^0$ by Hansen et al.. They also conjectured that the problem is in AC$^0$. In this paper, we disprove their conjecture by showing that the problem is ... More
Be Careful What You Backpropagate: A Case For Linear Output Activations & Gradient BoostingJul 13 2017In this work, we show that saturating output activation functions, such as the softmax, impede learning on a number of standard classification tasks. Moreover, we present results showing that the utility of softmax does not stem from the normalization, ... More
Mid-level Elements for Object DetectionApr 27 2015Building on the success of recent discriminative mid-level elements, we propose a surprisingly simple approach for object detection which performs comparable to the current state-of-the-art approaches on PASCAL VOC comp-3 detection challenge (no external ... More
PixelNet: Towards a General Pixel-level ArchitectureSep 21 2016We explore architectures for general pixel-level prediction problems, from low-level edge detection to mid-level surface normal estimation to high-level semantic segmentation. Convolutional predictors, such as the fully-convolutional network (FCN), have ... More
PixelNet: Representation of the pixels, by the pixels, and for the pixelsFeb 21 2017We explore design principles for general pixel-level prediction problems, from low-level edge detection to mid-level surface normal estimation to high-level semantic segmentation. Convolutional predictors, such as the fully-convolutional network (FCN), ... More
Patch Correspondences for Interpreting Pixel-level CNNsNov 29 2017Sep 04 2018We present compositional nearest neighbors (CompNN), a simple approach to visually interpreting distributed representations learned by a convolutional neural network (CNN) for pixel-level tasks (e.g., image synthesis and segmentation). It does so by reconstructing ... More
Effect of Various Regularizers on Model Complexities of Neural Networks in Presence of Input NoiseJan 31 2019Deep neural networks are over-parameterized, which implies that the number of parameters are much larger than the number of samples used to train the network. Even in such a regime deep architectures do not overfit. This phenomenon is an active area of ... More
Incremental Learning in Deep Convolutional Neural Networks Using Partial Network SharingDec 07 2017May 02 2019Deep convolutional neural network (DCNN) based supervised learning is a widely practiced approach for large-scale image classification. However, retraining these large networks to accommodate new, previously unseen data demands high computational time ... More
A Decision-theoretic Approach to Detection-based Target Search with a UAVJan 04 2018Search and rescue missions and surveillance require finding targets in a large area. These tasks often use unmanned aerial vehicles (UAVs) with cameras to detect and move towards a target. However, common UAV approaches make two simplifying assumptions. ... More
GRID Computing at Belle IINov 20 2015The Belle II experiment at the SuperKEKB collider in Tsukuba, Japan, will start physics data taking in 2018 and will accumulate 50 ab$^{-1}$ of e$^{+}$e$^{-}$ collision data, about 50 times larger than the data set of the earlier Belle experiment. The ... More
ATLAS Sensitivity to Leptoquarks, W_R and Heavy Majorana Neutrinos in Final States with High-pt Dileptons and Jets with Early LHC Data at 14 TeV proton-proton collisionsOct 12 2009Dilepton-jet final states are used to study physical phenomena not predicted by the standard model. The ATLAS discovery potential for leptoquarks and Majorana Neutrinos is presented using a full simulation of the ATLAS detector at the Large Hadron Collider. ... More
On a generalization of iterated and randomized roundingNov 05 2018Apr 10 2019We give a general method for rounding linear programs that combines the commonly used iterated rounding and randomized rounding techniques. In particular, we show that whenever iterated rounding can be applied to a problem with some slack, there is a ... More
Investigation to implicate data on cloudsFeb 07 2012Cloud computing can and does mean different things to different people. The common characteristics most shares are on-demand secure access to metered services from nearly anywhere and dislocation of data from inside to outside the organization. Vision ... More
Constructive Algorithms for Discrepancy MinimizationFeb 11 2010Aug 09 2010Given a set system (V,S), V={1,...,n} and S={S1,...,Sm}, the minimum discrepancy problem is to find a 2-coloring of V, such that each set is colored as evenly as possible. In this paper we give the first polynomial time algorithms for discrepancy minimization ... More
Checkbochs: Use Hardware to Check SoftwareJan 14 2006In this paper, we present a system called Checkbochs, a machine simulator that checks rules about its guest operating system and applications at the hardware level. The properties to be checked can be implemented as `plugins' in the Checkbochs simulator. ... More
New physics searches in $B \rightarrow D^{(*)}τν$ decaysNov 18 2014I review the current status of measurements involving semi-tauonic $B$ meson decay at the $B$-factories. I briefly discuss the experimental methods and highlight the importance of background contributions especially from poorly understood $D^{**}$ in ... More
Incremental Learning in Deep Convolutional Neural Networks Using Partial Network SharingDec 07 2017Sep 17 2018Deep convolutional neural network (DCNN) based supervised learning is a widely practiced approach for large-scale image classification. However, retraining these large networks to accommodate new, previously unseen data demands high computational time ... More
TraNNsformer: Neural network transformation for memristive crossbar based neuromorphic system designAug 26 2017Mar 04 2018Implementation of Neuromorphic Systems using post Complementary Metal-Oxide-Semiconductor (CMOS) technology based Memristive Crossbar Array (MCA) has emerged as a promising solution to enable low-power acceleration of neural networks. However, the recent ... More
Search for Leptoquarks, Excited Leptons and Technicolor at the LHCOct 19 2008The ATLAS and CMS experiments at the Large Hadron Collider (LHC) will soon search for physics phenomena that are not predicted by the Standard Model. Technicolor, Compositeness and GUT-based models are rich in high-pt leptons and could be studied in such ... More
Towards Deep Semantic Analysis Of HashtagsJan 13 2015Hashtags are semantico-syntactic constructs used across various social networking and microblogging platforms to enable users to start a topic specific discussion or classify a post into a desired category. Segmenting and linking the entities present ... More
RESPARC: A Reconfigurable and Energy-Efficient Architecture with Memristive Crossbars for Deep Spiking Neural NetworksFeb 20 2017Neuromorphic computing using post-CMOS technologies is gaining immense popularity due to its promising abilities to address the memory and power bottlenecks in von-Neumann computing systems. In this paper, we propose RESPARC - a reconfigurable and energy ... More
FALCON: Feature Driven Selective Classification for Energy-Efficient Image RecognitionSep 12 2016Mar 08 2017Machine-learning algorithms have shown outstanding image recognition or classification performance for computer vision applications. However, the compute and energy requirement for implementing such classifier models for large-scale problems is quite ... More
Jet fragmentation as a tool to explore double parton scattering using Z-boson + jets processes at the LHCMay 01 2019The Large Hadron Collider witnesses the highest ever production cross-section of double parton scattering processes. The production of a Z-boson along with two jets from double parton scattering provides a unique opportunity to explore the kinematics ... More
End-to-End Relation Extraction using LSTMs on Sequences and Tree StructuresJan 05 2016Jun 08 2016We present a novel end-to-end neural model to extract entities and relations between them. Our recurrent neural network based model captures both word sequence and dependency tree substructure information by stacking bidirectional tree-structured LSTM-RNNs ... More
Interpreting Neural Networks to Improve Politeness ComprehensionOct 09 2016We present an interpretable neural network approach to predicting and understanding politeness in natural language requests. Our models are based on simple convolutional neural networks directly on raw text, avoiding any manual identification of complex ... More
Adversarial Over-Sensitivity and Over-Stability Strategies for Dialogue ModelsSep 06 2018We present two categories of model-agnostic adversarial strategies that reveal the weaknesses of several generative, task-oriented dialogue models: Should-Not-Change strategies that evaluate over-sensitivity to small and semantics-preserving edits, as ... More
On the discrepancy of random low degree set systemsOct 08 2018Motivated by the celebrated Beck-Fiala conjecture, we consider the random setting where there are $n$ elements and $m$ sets and each element lies in $t$ randomly chosen sets. In this setting, Ezra and Lovett showed an $O((t \log t)^{1/2})$ discrepancy ... More
Potential-Function Proofs for First-Order MethodsDec 13 2017Mar 07 2019This note discusses proofs for convergence of first-order methods based on simple potential-function arguments. We cover methods like gradient descent (for both smooth and non-smooth settings), mirror descent, and some accelerated variants.
A Novel Active Contour Model for Texture SegmentationJun 28 2013Texture is intuitively defined as a repeated arrangement of a basic pattern or object in an image. There is no mathematical definition of a texture though. The human visual system is able to identify and segment different textures in a given image. Automating ... More
An Overview of Portable Distributed TechniquesJan 13 2011In this paper, we reviewed of several portable parallel programming paradigms for use in a distributed programming environment. The Techniques reviewed here are portable. These are mainly distributing computing using MPI pure java based, MPI native java ... More
Observation-based Cooperation Enforcement in Ad Hoc NetworksJul 04 2003Jul 06 2003Ad hoc networks rely on the cooperation of the nodes participating in the network to forward packets for each other. A node may decide not to cooperate to save its resources while still using the network to relay its traffic. If too many nodes exhibit ... More
Source-Target Inference Models for Spatial Instruction UnderstandingJul 12 2017Nov 21 2017Models that can execute natural language instructions for situated robotic tasks such as assembly and navigation have several useful applications in homes, offices, and remote scenarios. We study the semantics of spatially-referred configuration and arrangement ... More
Improved Algorithmic Bounds for Discrepancy of Sparse Set SystemsJan 13 2016Feb 02 2016We consider the problem of finding a low discrepancy coloring for sparse set systems where each element lies in at most $t$ sets. We give an algorithm that finds a coloring with discrepancy $O((t \log n \log s)^{1/2})$ where $s$ is the maximum cardinality ... More
Geometric Polynomial Constraints in Higher-Order Graph MatchingMay 24 2014Correspondence is a ubiquitous problem in computer vision and graph matching has been a natural way to formalize correspondence as an optimization problem. Recently, graph matching solvers have included higher-order terms representing affinities beyond ... More
Effect of Symmetry Energy on Intermediate Mass Fragments ProductionJul 27 2011When energy of colliding nuclei is between 100-600 MeV/nucleon then multifragmentation take place. By studying the fragments (at final stage of reaction) we can seize the idea about initial condition and various other parameter which influence the reaction ... More
The Geometry of SchedulingAug 28 2010We consider the following general scheduling problem: The input consists of n jobs, each with an arbitrary release time, size, and a monotone function specifying the cost incurred when the job is completed at a particular time. The objective is to find ... More
Phase-space analysis of fragments formed in heavy-ion collisionsApr 01 2011We study the effect of momentum-dependent interactions and a broader Gaussian on multifragmentation. We also look into the details of the fragment structure for a broader Gaussian and momentum-dependent interactions. We find that nucleons forming the ... More
FALCON: Feature Driven Selective Classification for Energy-Efficient Image RecognitionSep 12 2016Machine-learning algorithms have shown outstanding image recognition or classification performance for computer vision applications. However, the compute and energy requirement for implementing such classifier models for large-scale problems is quite ... More
An All-Memristor Deep Spiking Neural Computing System: A Step Towards Realizing the Low Power,Stochastic BrainDec 05 2017Apr 13 2018Deep 'Analog Artificial Neural Networks' (ANNs) perform complex classification problems with remarkably high accuracy. However, they rely on humongous amount of power to perform the calculations, veiling the accuracy benefits. The biological brain on ... More
Scalable Object Detection for Stylized ObjectsNov 27 2017Nov 29 2017Following recent breakthroughs in convolutional neural networks and monolithic model architectures, state-of-the-art object detection models can reliably and accurately scale into the realm of up to thousands of classes. Things quickly break down, however, ... More
Efficient Hybrid Network Architectures for Extremely Quantized Neural Networks Enabling Intelligence at the EdgeFeb 01 2019The recent advent of `Internet of Things' (IOT) has increased the demand for enabling AI-based edge computing. This has necessitated the search for efficient implementations of neural networks in terms of both computations and storage. Although extreme ... More
Hardy's Theorem for Gabor transformMar 09 2016We establish analogues of Hardy's theorem for Gabor transform on locally compact abelian groups, Euclidean motion group and several general classes of nilpotent Lie groups which include Heisenberg groups, thread-like nilpotent Lie groups, $2$-NPC nilpotent ... More
Minimizing Flow-Time on Unrelated MachinesJan 28 2014Jun 09 2015We consider some flow-time minimization problems in the unrelated machines setting. In this setting, there is a set of $m$ machines and a set of $n$ jobs, and each job $j$ has a machine dependent processing time of $p_{ij}$ on machine $i$. The flow-time ... More
Can Chern-Simons or Rarita-Schwinger be a Volkov-Akulov Goldstone?Jun 15 2018Jul 20 2018We study three-dimensional non-linear models of vector and vector-spinor Goldstone fields associated with the spontaneous breaking of certain higher-spin counterparts of supersymmetry whose Lagrangians are of a Volkov-Akulov type. Goldstone fields in ... More
Integrating Fuzzy and Ant Colony System for Fuzzy Vehicle Routing Problem with Time WindowsNov 14 2014In this paper fuzzy VRPTW with an uncertain travel time is considered. Credibility theory is used to model the problem and specifies a preference index at which it is desired that the travel times to reach the customers fall into their time windows. We ... More
Increasing Herd Immunity with Influenza RevaccinationAug 24 2013Jan 06 2015Seasonal influenza is a significant public health concern in the United States and globally. While influenza vaccines are the single most effective intervention to reduce influenza morbidity and mortality, there is considerable debate surrounding the ... More
Revenue Forecasting for Enterprise ProductsNov 21 2016For any business, planning is a continuous process, and typically business-owners focus on making both long-term planning aligned with a particular strategy as well as short-term planning that accommodates the dynamic market situations. An ability to ... More
Electrical Signature of Excitonic Electroluminescence and Mott Transition at Room TemperatureDec 27 2013Jun 17 2014Mostly optical spectroscopies are used to investigate existence of excitons, Mott transitions and other exquisite excitonic condensed phases of matter. On the other hand, electrical signatures of excitons are hardly explored. Here we examine steady state, ... More
K-Algorithm A Modified Technique for Noise Removal in Handwritten DocumentsJun 06 2013OCR has been an active research area since last few decades. OCR performs the recognition of the text in the scanned document image and converts it into editable form. The OCR process can have several stages like pre-processing, segmentation, recognition ... More
A Note on the Complexity of Model-Checking Bounded Multi-Pushdown SystemsDec 06 2012In this note, we provide complexity characterizations of model checking multi-pushdown systems. Multi-pushdown systems model recursive concurrent programs in which any sequential process has a finite control. We consider three standard notions for boundedness: ... More
Avoiding Reasoning Shortcuts: Adversarial Evaluation, Training, and Model Development for Multi-Hop QAJun 17 2019Multi-hop question answering requires a model to connect multiple pieces of evidence scattered in a long context to answer the question. In this paper, we show that in the multi-hop HotpotQA (Yang et al., 2018) dataset, the examples often contain reasoning ... More
Algorithmic Discrepancy Beyond Partial ColoringNov 06 2016The partial coloring method is one of the most powerful and widely used method in combinatorial discrepancy problems. However, in many cases it leads to sub-optimal bounds as the partial coloring step must be iterated a logarithmic number of times, and ... More
Active Contour Models for Manifold Valued Image SegmentationJun 26 2013Nov 11 2013Image segmentation is the process of partitioning a image into different regions or groups based on some characteristics like color, texture, motion or shape etc. Active contours is a popular variational method for object segmentation in images, in which ... More
On the metric distortion of nearest-neighbour graphs on random point setsApr 23 2008Jul 18 2008We study the graph constructed on a Poisson point process in $d$ dimensions by connecting each point to the $k$ points nearest to it. This graph a.s. has an infinite cluster if $k > k_c(d)$ where $k_c(d)$, known as the critical value, depends only on ... More
The large sieve with power moduli for $\mathbb{Z}[i]$Feb 25 2018May 24 2018We establish a large sieve inequality for power moduli in $\mathbb{Z}[i]$, extending earlier work by L. Zhao and the first-named author on the large sieve for power moduli for the classical case of moduli in $\mathbb{Z}$. Our method starts with a version ... More
Shortcut-Stacked Sentence Encoders for Multi-Domain InferenceAug 07 2017Nov 28 2017We present a simple sequential sentence encoder for multi-domain natural language inference. Our encoder is based on stacked bidirectional LSTM-RNNs with shortcut connections and fine-tuning of word embeddings. The overall supervised model uses the above ... More
Improved Human Emotion Recognition Using Symmetry of Facial Key Points with Dihedral GroupApr 14 2017This article describes how to deploy dihedral group theory to detect Facial Key Points (FKP) symmetry to recognize emotions. The method can be applied in many other areas which those have the same data texture.
Object Ordering with Bidirectional Matchings for Visual ReasoningApr 18 2018Sep 06 2018Visual reasoning with compositional natural language instructions, e.g., based on the newly-released Cornell Natural Language Visual Reasoning (NLVR) dataset, is a challenging task, where the model needs to have the ability to create an accurate mapping ... More
An Improved Multiple Faults Reassignment based Recovery in Cluster ComputingFeb 13 2011In case of multiple node failures performance becomes very low as compare to single node failure. Failures of nodes in cluster computing can be tolerated by multiple fault tolerant computing. Existing recovery schemes are efficient for single fault but ... More
Closed-Book Training to Improve Summarization Encoder MemorySep 12 2018A good neural sequence-to-sequence summarization model should have a strong encoder that can distill and memorize the important information from long input texts so that the decoder can generate salient summaries based on the encoder's memory. In this ... More
Continual and Multi-Task Architecture SearchJun 12 2019Architecture search is the process of automatically learning the neural model or cell structure that best suits the given task. Recently, this approach has shown promising performance improvements (on language modeling and image classification) with reasonable ... More
Algorithmic Discrepancy Beyond Partial ColoringNov 06 2016Jul 11 2017The partial coloring method is one of the most powerful and widely used method in combinatorial discrepancy problems. However, in many cases it leads to sub-optimal bounds as the partial coloring step must be iterated a logarithmic number of times, and ... More
A modified proximity approach in the fusion of heavy-ionsOct 01 2010By using a suitable set of the surface energy coefficient, nuclear radius, and universal function, the original proximity potential 1977 is modified. The overestimate of the data by 4 % reported in the literature is significantly reduced. Our modified ... More
Magnetic field induced band depopulation in intrinsic InSb: A revisitSep 28 2005The effect of Landau level formation on the population of intrinsic electrons in InSb is probed near room temperature in magnetic fields upto 16 Tesla. Although the measured magnetic field dependence of the Hall coefficient is qualitatively similar to ... More
Approximation-Friendly Discrepancy RoundingDec 07 2015Rounding linear programs using techniques from discrepancy is a recent approach that has been very successful in certain settings. However this method also has some limitations when compared to approaches such as randomized and iterative rounding. We ... More
Assesment of multifragmentation under the effect of symmetry energy and cross-sectionOct 02 2011The effect of symmetry energy and cross section had seen on the fragment production in the multifragmentation of $^{20}Ne_{10}+ ^{20}Ne_{10}$ and $^{197}Au_{79} + ^{197}Au_{79}$ at incident energy 50-1000 MeV/nucleon using isospin dependent quantum molecular ... More
Polite Dialogue Generation Without Parallel DataMay 08 2018Stylistic dialogue response generation, with valuable applications in personality-based conversational agents, is a challenging task because the response needs to be fluent, contextually-relevant, as well as paralinguistically accurate. Moreover, parallel ... More
Heisenberg Uncertainty Inequality for Gabor TransformJul 02 2015We discuss Heisenberg uncertainty inequality for groups of the form $K \ltimes \mathbb{R}^n$, $K$ is a separable unimodular locally compact group of type I. This inequality is also proved for Gabor transform for several classes of groups of the form $K ... More
On the Adaptivity Gap of Stochastic OrienteeringNov 14 2013May 09 2014The input to the stochastic orienteering problem consists of a budget $B$ and metric $(V,d)$ where each vertex $v$ has a job with deterministic reward and random processing time (drawn from a known distribution). The processing times are independent across ... More
Accomplish the Application Area in Cloud ComputingMar 10 2012In the cloud computing application area of accomplish, we find the fact that cloud computing covers a lot of areas are its main asset. At a top level, it is an approach to IT where many users, some even from different companies get access to shared IT ... More
Approximation-Friendly Discrepancy RoundingDec 07 2015Dec 02 2016Rounding linear programs using techniques from discrepancy is a recent approach that has been very successful in certain settings. However this method also has some limitations when compared to approaches such as randomized and iterative rounding. We ... More
A Software-only Mechanism for Device Passthrough and SharingAug 26 2015Sep 22 2016Network processing elements in virtual machines, also known as Network Function Virtualization (NFV) often face CPU bottlenecks at the virtualization interface. Even highly optimized paravirtual device interfaces fall short of the throughput requirements ... More
On the study of phase-space analysis of fragments produced in heavy-ion collisionsAug 01 2011Dec 13 2011Effect of momentum-dependent interactions and broader Gaussian is investigated on the emission of various fragments formed in a heavy-ion reaction. We also study the corresponding structure details of those fragments for broader Gaussian and momentum-dependent ... More
Voltage modulated electro-luminescence spectroscopy and negative capacitance - the role of sub-bandgap states in light emitting devicesAug 02 2011Voltage modulated electroluminescence spectra and low frequency ({\leq} 100 kHz) impedance characteristics of electroluminescent diodes are studied. Voltage modulated light emission tracks the onset of observed negative capacitance at a forward bias level ... More
Temperature dependent reversal of voltage modulated light emission and negative capacitance in AlGaInP based multi quantum well light emitting devicesAug 09 2012We report a reversal in negative capacitance and voltage modulated light emission from AlGaInP based multi-quantum well electroluminescent diodes under temperature variation. Unlike monotonically increasing CW light emission with decreasing temperature, ... More
Improving Visual Question Answering by Referring to Generated Paragraph CaptionsJun 14 2019Paragraph-style image captions describe diverse aspects of an image as opposed to the more common single-sentence captions that only provide an abstract description of the image. These paragraph captions can hence contain substantial information of the ... More
New observables for multiple-parton interactions measurements using Z + jets process at the LHCFeb 17 2016Multiple-parton interactions play a vital role in hadron-hadron collisions. This paper presents a study of the multiple-parton interactions with simulated Z + jets events in proton-proton collisions at a centre-of-mass energy of 13 TeV. The events are ... More
On Generalized Fractional Derivatives Involving Generalized k-Mittag Leffler FunctionFeb 07 2019In this paper, certain generalized fractional derivative formulae are introduced involving the k-Mittag-Leffler function. Then their image formulae (using Beta transform, Laplace transform and Whittaker transform) are also established. The results obtained ... More
The Impact of Past Epidemics on Future Disease DynamicsOct 12 2009Many pathogens spread primarily via direct contact between infected and susceptible hosts. Thus, the patterns of contacts or contact network of a population fundamentally shapes the course of epidemics. While there is a robust and growing theory for the ... More
On Large Deviation Property of Recurrence TimesMar 05 2013May 18 2013We extend the study by Ornstein and Weiss on the asymptotic behavior of the normalized version of recurrence times and establish the large deviation property for a certain class of mixing processes. Further, an estimator for entropy based on recurrence ... More
A tutorial on evaluating time-varying discrimination accuracy for survival models used in dynamic decision-makingJun 29 2017Feb 21 2018Many medical decisions involve the use of dynamic information collected on individual patients toward predicting likely transitions in their future health status. If accurate predictions are developed, then a prognostic mode can identify patients at greatest ... More
On Match Lengths and the Asymptotic Behavior of Sliding Window Lempel-Ziv Algorithm for Zero Entropy SequencesMar 05 2013May 18 2013The Sliding Window Lempel-Ziv (SWLZ) algorithm has been studied from various perspectives in information theory literature. In this paper, we provide a general law which defines the asymptotics of match length for stationary and ergodic zero entropy processes. ... More
People Counting in High Density Crowds from Still ImagesJul 30 2015We present a method of estimating the number of people in high density crowds from still images. The method estimates counts by fusing information from multiple sources. Most of the existing work on crowd counting deals with very small crowds (tens of ... More
Unifying Geometric Features and Facial Action Units for Improved Performance of Facial Expression AnalysisJun 02 2016Previous approaches to model and analyze facial expression analysis use three different techniques: facial action units, geometric features and graph based modelling. However, previous approaches have treated these technique separately. There is an interrelationship ... More
Role of non-coplanarity in nuclear reactions using the Wong formula based on the proximity potentialOct 07 2010we assessed Wong's formula for its angular momentum $\ell$-summation and "barrier modification" effects at sub-barrier energies in the dominant fusion-evaporation and capture (equivalently, quasi-fission) reaction cross-sections. For use of the multipole ... More
Geographic Information Systems in Evaluation and Visualization of Construction ScheduleMar 25 2008Commercially available scheduling tools such as Primavera and Microsoft Project fail to provide information pertaining to the spatial aspects of construction project. A methodology using geographical information systems (GIS) is developed to represent ... More
On-Line Balancing of Random InputsMar 16 2019We consider an online vector balancing game where vectors $v_t$, chosen uniformly at random in $\{-1,+1\}^n$, arrive over time and a sign $x_t \in \{-1,+1\}$ must be picked immediately upon the arrival of $v_t$. The goal is to minimize the $L^\infty$ ... More
The Challenge of Crafting Intelligible IntelligenceMar 09 2018Oct 15 2018Since Artificial Intelligence (AI) software uses techniques like deep lookahead search and stochastic optimization of huge neural networks to fit mammoth datasets, it often results in complex behavior that is difficult for people to understand. Yet organizations ... More
On Match Lengths, Zero Entropy and Large Deviations - with Application to Sliding Window Lempel-Ziv AlgorithmNov 05 2014Nov 06 2014The Sliding Window Lempel-Ziv (SWLZ) algorithm that makes use of recurrence times and match lengths has been studied from various perspectives in information theory literature. In this paper, we undertake a finer study of these quantities under two different ... More
Achievable Performance of Blind Policies in Heavy TrafficDec 24 2015Jan 06 2016For a GI/GI/1 queue, we show that the average sojourn time under the (blind) Randomized Multilevel Feedback algorithm is no worse than that under the Shortest Remaining Processing Time algorithm times a logarithmic function of the system load. Moreover, ... More
What to talk about and how? Selective Generation using LSTMs with Coarse-to-Fine AlignmentSep 02 2015Jan 08 2016We propose an end-to-end, domain-independent neural encoder-aligner-decoder model for selective generation, i.e., the joint task of content selection and surface realization. Our model first encodes a full set of over-determined database event records ... More
Detecting Linguistic Characteristics of Alzheimer's Dementia by Interpreting Neural ModelsApr 17 2018Alzheimer's disease (AD) is an irreversible and progressive brain disease that can be stopped or slowed down with medical treatment. Language changes serve as a sign that a patient's cognitive functions have been impacted, potentially leading to early ... More
Simple self-gettering differential-pump for minimizing source oxidation in oxide-MBE environmentJun 03 2011Source oxidation of easily oxidizing elements such as Ca, Sr, Ba, and Ti in an oxidizing ambient leads to their flux instability and is one of the biggest problems in the multi-elemental oxide Molecular Beam Epitaxy technique. Here we report a new scheme ... More
Web-scale Surface and Syntactic n-gram Features for Dependency ParsingFeb 25 2015We develop novel first- and second-order features for dependency parsing based on the Google Syntactic Ngrams corpus, a collection of subtree counts of parsed sentences from scanned books. We also extend previous work on surface $n$-gram features from ... More