arXiv research

Improved elimination strategies for adaptive bandit identification reduce sample complexity and computational burden.

We introduce Neural Choice by Elimination, a new framework that integrates deep neural networks into probabilistic sequential choice models for learning to rank. Given a set of items to chose from, the elimination strategy starts with the whole item set and iteratively eliminates the least worthy item in the remaining …

PoWER-BERT speeds up BERT inference by eliminating redundant word-vectors.

A new algorithm optimizes local objectives in federated learning with heterogeneous clients.

Compound interest as well as inflation grows exponentially with time, whereas other means to repay debt grow polynomially. For this and other, mostly political, reasons, debt without inflation is unsustainable. We suggest a discontinuous way to eliminate debt by nullifying it. This scenario is preferable to current cen…

We present a new strategy for gap estimation in randomized algorithms for multiarmed bandits and combine it with the EXP3++ algorithm of Seldin and Slivkins (2014). In the stochastic regime the strategy reduces dependence of regret on a time horizon from $(\ln t)^3$ to $(\ln t)^2$ and eliminates an additive factor of o…

We consider a market with fractional Brownian motion with stochastic integrals generated by the Riemann sums. We found that this market is arbitrage free if admissible strategies that are using observations with an arbitrarily small delay. Moreover, we found that this approach eliminates the discontinuity of the stocha…

The paper introduces a method to make neural networks more robust to adversarial attacks.

New methods reduce extrapolation errors in feature importance.

A simple trading model based on pair pattern strategy space with holding periods is proposed. Power-law behaviors are observed for the return variance $σ^2$, the price impact $H$ and the predictability $K$ for both models with linear and square root impact functions. The sum of the traders' wealth displays a positive v…

EB improves asset pricing by mining large strategies without lookahead bias.

CRA improves UL-based CO solvers by dynamically smoothing and enforcing discreteness.

The study eliminates infinite families of knots with nontrivial Alexander polynomials and improves unknotting number data.

PES method reduces bias in gradient estimation for unrolled graphs.

Proposes a flexible tournament design combining knockout and round-robin.

How can we control for latent discrimination in predictive models? How can we provably remove it? Such questions are at the heart of algorithmic fairness and its impacts on society. In this paper, we define a new operational fairness criteria, inspired by the well-understood notion of omitted variable-bias in statistic…

Algorithm reduces regret in multi-player bandits with unknown collision rewards.

Classification may not be reliable for several reasons: noise in the data, insufficient input information, overlapping distributions and sharp definition of classes. Faced with several possibilities neural network may in such cases still be useful if instead of a classification elimination of improbable classes is done…

A new approach estimates propagators for trading risky assets.

It is a significant challenge to design probabilistic programming systems that can accommodate a wide variety of inference strategies within a unified framework. Noting that the versatility of modern automatic differentiation frameworks is based in large part on the unifying concept of tensors, we describe a software a…

Proposes a new semi-parametric framework for batched bandits with covariates.

Study optimizes best-arm identification with minimax and Bayes strategies.

StockGPT predicts stock returns using AI, outperforming traditional strategies.

New algorithm eliminates arms to minimize regret in complex bandit problems.

A fast deep learning method for parallel MRI without calibration.

We study an original problem of pure exploration in a strategic bandit model motivated by Monte Carlo Tree Search. It consists in identifying the best action in a game, when the player may sample random outcomes of sequentially chosen pairs of actions. We propose two strategies for the fixed-confidence setting: Maximin…

Learning how to act when there are many available actions in each state is a challenging task for Reinforcement Learning (RL) agents, especially when many of the actions are redundant or irrelevant. In such cases, it is sometimes easier to learn which actions not to take. In this work, we propose the Action-Elimination…

A2 Learning reduces redundant examples in AL for NLP tasks.

Aims to eliminate domain bias in authentication without domain labels.

Improved self-supervised denoising for Poisson-Gaussian noise.

A novel approach ODAR detects outliers for clustering.

Gradient descent with biased rounding errors converges faster under certain conditions.

We simplify Khovanov homology for torus braids using Gaussian elimination.

A wide class of machine learning algorithms can be reduced to variable elimination on factor graphs. While factor graphs provide a unifying notation for these algorithms, they do not provide a compact way to express repeated structure when compared to plate diagrams for directed graphical models. To exploit efficient t…

Logic approach finds real singularities in differential equations.

LSAM optimizes deep learning training with improved efficiency.

We develop an approach for feature elimination in statistical learning with kernel machines, based on recursive elimination of features.We present theoretical properties of this method and show that it is uniformly consistent in finding the correct feature space under certain generalized assumptions.We present four cas…

Greedy selection works well in a toy model of independent increments.

Safe-DRFS selects features robust to covariate shifts for reliable performance.

In this paper, we theoretically prove that adding one special neuron per output unit eliminates all suboptimal local minima of any deep neural network, for multi-class classification, binary classification, and regression with an arbitrary loss function, under practical assumptions. At every local minimum of any deep n…

Unified framework for response-adaptive targeting in multi-treatment experiments

Object recognition systems usually require fully complete manually labeled training data to train the classifier. In this paper, we study the problem of object recognition where the training samples are missing during the classifier learning stage, a task also known as zero-shot learning. We propose a novel zero-shot l…

In many scientific and engineering applications, we are tasked with the maximisation of an expensive to evaluate black box function $f$. Traditional settings for this problem assume just the availability of this single function. However, in many cases, cheap approximations to $f$ may be obtainable. For example, the exp…

In our empirical study, we examine the price of liquid stocks after experiencing a large intraday price change using data from the NYSE and the NASDAQ. We find significant reversal for both intraday price decreases and increases. The results are stable against varying parameters. While on the NYSE the large widening of…

Two new feature selection algorithms improve on RFE.

In this paper, we consider the problem of online learning of Markov decision processes (MDPs) with very large state spaces. Under the assumptions of realizable function approximation and low Bellman ranks, we develop an online learning algorithm that learns the optimal value function while at the same time achieving ve…

New algorithms eliminate stepsize tuning for bilevel optimization problems.

Study shows market makers can cooperate without communication.