Search

We study revenue optimization learning algorithms for repeated posted-priceauctions where a seller interacts with a single strategic buyer that holds afixed private valuation for a good and seeks to maximize his cumulativediscounted surplus. For this setting, first, we propose a novel algorithm...

In this paper we propose a general framework of performing MCMC with only amini-batch of data. We show by estimating the Metropolis-Hasting ratio withonly a mini-batch of data, one is essentially sampling from the true posteriorraised to a known temperature. We show by experiments that our method...

We methodologically address the problem of Q-value overestimation in deepreinforcement learning to handle high-dimensional state spaces efficiently. Byadapting concepts from information theory, we introduce an intrinsic penaltysignal encouraging reduced Q-value estimates. The resultant...

Deep learning has become the state of the art approach in many machinelearning problems such as classification. It has recently been shown that deeplearning is highly vulnerable to adversarial perturbations. Taking the camerasystems of self-driving cars as an example, small adversarial perturbations...

Since the invention of word2vec, the skip-gram model has significantlyadvanced the research of network embedding, such as the recent emergence of theDeepWalk, LINE, PTE, and node2vec approaches. In this work, we show that all ofthe aforementioned models with negative sampling can be unified into the...

We propose a family of variational approximations to Bayesian posteriordistributions, called $\alpha$-VB, with provable statistical guarantees. Thestandard variational approximation is a special case of $\alpha$-VB with$\alpha=1$. When $\alpha \in(0,1]$, a novel class of variational inequalitiesare...

We show that the orthogonal projection operator onto the range of the adjointof a linear operator $T$ can be represented as $UT,$ where $U$ is an invertiblelinear operator. Using this representation we obtain a decomposition of aNormal random vector $Y$ as the sum of a linear transformation of $Y$...

We extend a data-based model-free multifractal method of exoplanet detectionto probe exoplanetary atmospheres. Whereas the transmission spectrum is studiedduring the primary eclipse, we analyze the emission spectrum during thesecondary eclipse, thereby probing the atmospheric limb. In addition to...

Neural samplers such as variational autoencoders (VAEs) or generativeadversarial networks (GANs) approximate distributions by transforming samplesfrom a simple random source---the latent space---to samples from a more complexdistribution represented by a dataset. While the manifold hypothesis...

Mendelian randomization (MR) is a method of exploiting genetic variation tounbiasedly estimate a causal effect in presence of unmeasured confounding. MRis being widely used in epidemiology and other related areas of populationscience. In this paper, we study statistical inference in the...

Convolutional Neural Networks (CNNs) have become the method of choice forlearning problems involving 2D planar images. However, a number of problems ofrecent interest have created a demand for models that can analyze sphericalimages. Examples include omnidirectional vision for drones, robots...

We analyzed 2012 and 2016 YouGov pre-election polls in order to understandhow different population groups voted in the 2012 and 2016 elections. We brokethe data down by demographics and state. We display our findings with a seriesof graphs and maps. The R code associated with this project is...

This paper provides a generalization of a classical result obtained by Wilksabout the asymptotic behavior of the likelihood ratio. The new results dealwith the asymptotic behavior of the joint distribution of a vector oflikelihood ratios which turn out to be stochastically dependent, due to...

Researchers are increasingly incorporating numeric high-order data, i.e.,numeric tensors, within their practice. Just like the matrix/vector (MV)paradigm, the development of multi-purpose, but high-performance, sparse datastructures and algorithms for arithmetic calculations, e.g., those found...

In this paper, we introduce a new 2D modulation scheme referred to as OTFS(Orthogonal Time Frequency & Space) that multiplexes information QAM symbolsover new class of carrier waveforms that correspond to localized pulses in asignal representation called the delay-Doppler representation. OTFS...

In this paper, we address the modeling and identification of controlaugmented dynamics for a small fixed-wing Unmanned Aerial Vehicle (UAV) with awidely available off-the-shelf (OTS) autopilot in the loop, utilizing astandard sensor suite. A high-level Nonlinear Model Predictive Controller(NMPC) is...

Over the past few years, Spiking Neural Networks (SNNs) have become popularas a possible pathway to enable low-power event-driven neuromorphic hardware.However, their application in machine learning have largely been limited tovery shallow neural network architectures for simple problems. In this...

Deep neural networks represent a powerful class of function approximatorsthat can learn to compress and reconstruct images. Existing image compressionalgorithms based on neural networks learn quantized representations with aconstant spatial bit rate across each image. While entropy coding...

Social media are more than just a one-way communication channel. Data can becollected, analyzed and contextualized to support disaster risk management.However, disaster management agencies typically use such added-valueinformation to support only their own decisions. A feedback loop...

We prove a tight lower bound (up to constant factors) on the samplecomplexity of any non-interactive local differentially private protocol foroptimizing a linear function over the simplex. This lower bound also implies atight lower bound (again, up to constant factors) on the sample complexity ofany...