arXiv research

AI agents improve forecast combination in empirical economics.

Compensation methods correct overestimation of adversarial robustness in neural networks.

AI agents improve forecast combination but require transparency.

We devise and analyze algorithms for the empirical policy evaluation problem in reinforcement learning. Our algorithms explore backward from high-cost states to find high-value ones, in contrast to forward approaches that work forward from all states. While several papers have demonstrated the utility of backward explo…

Empirical mode modeling improves state-space analysis of noisy data.

This research evaluates generalization measures in deep learning.

The (contextual) multi-armed bandit problem (MAB) provides a formalization of sequential decision-making which has many applications. However, validly evaluating MAB policies is challenging; we either resort to simulations which inherently include debatable assumptions, or we resort to expensive field trials. Recently …

Large language models correlate in errors, even with different architectures and providers.

Synthetic experiments are crucial for assessing causal machine learning methods.

The paper evaluates biased methods for alpha-divergence minimization.

Randomized control methods improve asset pricing and performance analysis.

Bayesian method infers contextual bandit policies robustly.

Causal inference is central to many areas of artificial intelligence, including complex reasoning, planning, knowledge-base construction, robotics, explanation, and fairness. An active community of researchers develops and enhances algorithms that learn causal models from data, and this work has produced a series of im…

Paper introduces Modular Jets for diagnosing model decompositions in pipelines.

This paper evaluates how different imputation methods affect predictive models.

A non-parametric method for evaluation of the aggregate loss distribution (ALD) by combining and numerically inverting the empirical characteristic functions (CFs) is presented and illustrated. This approach to evaluate ALD is based on purely non-parametric considerations, i.e., based on the empirical CFs of frequency …

We present and prove properties of a new offline policy evaluator for an exploration learning setting which is superior to previous evaluators. In particular, it simultaneously and correctly incorporates techniques from importance weighting, doubly robust evaluation, and nonstationary policy evaluation approaches. In a…

Proposes a stability evaluation criterion for learning models using distributional perturbations.

We present the first differentially private algorithms for reinforcement learning, which apply to the task of evaluating a fixed policy. We establish two approaches for achieving differential privacy, provide a theoretical analysis of the privacy and utility of the two algorithms, and show promising results on simple e…

We offer an experimental benchmark and empirical study for off-policy policy evaluation (OPE) in reinforcement learning, which is a key problem in many safety critical applications. Given the increasing interest in deploying learning-based methods, there has been a flurry of recent proposals for OPE method, leading to …

Saliency methods can aid understanding of deep neural networks. Recent years have witnessed many improvements to saliency methods, as well as new ways for evaluating them. In this paper, we 1) present a novel region-based attribution method, XRAI, that builds upon integrated gradients (Sundararajan et al. 2017), 2) int…

Motivated by the Basel 3 regulations, recent studies have considered joint forecasts of Value-at-Risk and Expected Shortfall. A large family of scoring functions can be used to evaluate forecast performance in this context. However, little intuitive or empirical guidance is currently available, which renders the choice…

Stochastic variational inference for Bayesian deep neural network (DNN) requires specifying priors and approximate posterior distributions over neural network weights. Specifying meaningful weight priors is a challenging problem, particularly for scaling variational inference to deeper architectures involving high dime…

PromptEval estimates LLM performance across many prompts, improving reproducibility.

Develops a new framework for perpetual futures on binary prediction markets.

New algorithms for model selection in off-policy evaluation of reinforcement learning.

LinMED is a new linear bandit algorithm with near-optimal regret bound.

CoinDICE estimates confidence intervals for unknown behavior policies in reinforcement learning.

Off-policy evaluation (OPE) in both contextual bandits and reinforcement learning allows one to evaluate novel decision policies without needing to conduct exploration, which is often costly or otherwise infeasible. The problem's importance has attracted many proposed solutions, including importance sampling (IS), self…

Study evaluates new models using human feedback from another model.

New approaches estimate recommendation metrics using sampling.

Suitability filter detects model performance degradation in real-world deployment.

The study addresses biases in evaluating molecular optimization methods and proposes methods to reduce these biases.

Paper offers a fast convergence theory for offline decision making.

Unified approach for non-stationary and clustered bandits.

Proves accuracy guarantees for self-supervised learning with correlated positive pairs.

Bayesian framework improves LLM evaluation stability and transparency.

OSIRIS reduces variance in off-policy evaluation by omitting irrelevant states.

In this work, we evaluate adversarial robustness in the context of transfer learning from a source trained on CIFAR 100 to a target network trained on CIFAR 10. Specifically, we study the effects of using robust optimisation in the source and target networks. This allows us to identify transfer learning strategies unde…

Study evaluates if LLMs have company-specific biases in financial sentiment analysis.

Sampling a fraction of pairs can match full evaluation in machine learning losses.

EMPIR combines low and full precision DNNs to enhance robustness against adversarial attacks.

Tiny benchmarks reduce LLM evaluation costs by using fewer examples.

A new method for faster multi-objective optimization by evaluating objectives separately.

The paper evaluates and improves uncertainty estimates in neural networks for safety-critical applications.

The study improves VaR forecast accuracy by modeling conditional quantile dynamics.

Contextual policy search (CPS) is a class of multi-task reinforcement learning algorithms that is particularly useful for robotic applications. A recent state-of-the-art method is Contextual Covariance Matrix Adaptation Evolution Strategies (C-CMA-ES). It is based on the standard black-box optimization algorithm CMA-ES…

The paper develops a neural network model for SPX option pricing.