arXiv research

Automates bias control in reinforcement learning algorithms.

Paper tackles overestimation bias in continuous control, improving performance by 25%.

Interpolated-MLPs control inductive bias for better performance in low-compute tasks.

Unified framework suppresses model bias in semi-supervised learning with decoupled sampling control.

Geometric framework explains and controls implicit bias in machine learning.

This paper studies how gradient descent in control systems can perform well on unseen data.

Improved 3D generative models for drug design reduce bias and enhance data efficiency.

Revisits VIC method to correct intrinsic reward bias in stochastic environments.

Truncated backpropagation through time (TBPTT) is a popular method for learning in recurrent neural networks (RNNs) that saves computation and memory at the cost of bias by truncating backpropagation after a fixed number of lags. In practice, choosing the optimal truncation length is difficult: TBPTT will not converge …

New framework minimizes interference and selection bias in network A/B testing.

Unified causal model improves controllable text generation without bias.

Synthetic control method improves policy evaluation in high-dimensional settings.

Deep RL algorithms can overfit to early experiences, leading to poor performance.

We consider the off-policy evaluation problem in Markov decision processes with function approximation. We propose a generalization of the recently introduced \emph{emphatic temporal differences} (ETD) algorithm \citep{SuttonMW15}, which encompasses the original ETD($λ$), as well as several other off-policy evaluation …

Bayesian imputation optimizes bias-variance tradeoff in time-series data.

Confounding bias, missing data, and selection bias are three common obstacles to valid causal inference in the data sciences. Covariate adjustment is the most pervasive technique for recovering casual effects from confounding bias. In this paper, we introduce a covariate adjustment formulation for controlling confoundi…

Study on NNs for forecasting time series with novel control variable combinations.

New method controls bias in unadjusted Hamiltonian Monte Carlo and underdamped Langevin.

We quantify causal bias in continuous treatment settings.

Gradient-based methods for optimisation of objectives in stochastic settings with unknown or intractable dynamics require estimators of derivatives. We derive an objective that, under automatic differentiation, produces low-variance unbiased estimators of derivatives at any order. Our objective is compatible with arbit…

WHOMP optimizes randomized controlled trials by minimizing subgroup bias.

Algorithm reduces variance in causal effect estimation from multiple datasets.

A simple method reduces bias in LLM auto-evaluators by controlling output length.

This report examines the Pinned AUC metric introduced and highlights some of its limitations. Pinned AUC provides a threshold-agnostic measure of unintended bias in a classification model, inspired by the ROC-AUC metric. However, as we highlight in this report, there are ways that the metric can obscure different kinds…

We use tools from geometric statistics to analyze the usual estimation procedure of a template shape. This applies to shapes from landmarks, curves, surfaces, images etc. We demonstrate the asymptotic bias of the template shape estimation using the stratified geometry of the shape space. We give a Taylor expansion of t…

Develops tools to audit ML models for bias and unfairness.

This paper tackles data-efficient nonlinear control in Hamiltonian systems using symplectic geometry.

Financial event studies often misestimate causal effects due to misspecified factor models.

A new reinforcement learning method reduces action complexity for robust control.

Untrained neural networks can unfairly assign predictions to the same class.

New TTP framework fuses control arms while controlling Type-I error.

We present experiments demonstrating that some other form of capacity control, different from network size, plays a central role in learning multilayer feed-forward networks. We argue, partially through analogy to matrix factorization, that this is an inductive bias that can help shed light on deep learning.

The paper tackles Neyman-Pearson classification control issues.

Experience replay is an important technique for addressing sample-inefficiency in deep reinforcement learning (RL), but faces difficulty in learning from binary and sparse rewards due to disproportionately few successful experiences in the replay buffer. Hindsight experience replay (HER) was recently proposed to tackle…

Fine-tunes LLMs to correct bias in predictions.

TSC improves causal effect estimation in panel data.

Paper analyzes how present-bias affects carbon emissions and proposes a method to mitigate it.

New method reduces bias in incomplete data using deliberate missingness.

Proposes tests to control confounding bias in predictive models.

The study examines when to trust confidence thresholding in pseudo-labelling regression.

New tools for assessing and correcting bias in AI algorithms.

The study reveals a persistent bias in the distribution of holonomy on compact hyperbolic 3-manifolds.

Cross-validation under sample selection bias can, in principle, be done by importance-weighting the empirical risk. However, the importance-weighted risk estimator produces sub-optimal hyperparameter estimates in problem settings where large weights arise with high probability. We study its sampling variance as a funct…

New method corrects bias in estimating entropic risk for better decision-making.

FinReflectKG - EvalBench benchmarks financial KG extraction from SEC 10-K filings.

Proposes volumization for neural networks to control bias-variance tradeoff.

The paper analyzes the bias-variance tradeoff for Bregman divergences.

New method controls bias in training data for fair outcomes.