Study detects and explains positional bias in financial LLMs.
problem Positional bias in financial decision-making using LLMs.
method Unified framework and benchmark for detecting and quantifying bias in Qwen2.5 models.
result Positional bias is pervasive, scale-sensitive, and resurfaces under nuanced prompt designs.
New method debiases selection bias in PU classification with exposure data.
problem Binary classification from positive and unlabeled data with selection bias.
method Automatic Debiased PUE (ADPUE) learning method.
result ADPUE outperforms traditional PU learning methods on various datasets.
Study improves unbiased recommender learning by addressing missing-reward bias.
problem Data bias caused by missing-reward observations in recommender systems.
method Proposes a novel estimator using propensity scores to mitigate both position and reward bias.
result The proposed estimator outperforms other methods, even with increased reward observation bias.
Paper measures cognitive bias in positive feedback trading using diffusion process estimates.
problem Measuring cognitive bias in positive feedback trading behavior.
method Conditional estimates of diffusion processes to quantify bias, proving asymptotic properties.
result Bias in positive feedback trading converges to zero over time, leading to adaptive expectations.
Paper proposes a method to estimate true positive proportion without knowing it.
problem Bias in binary classifier performance due to different positive item proportions.
method Maximum likelihood estimator for true proportion of positives.
result Method accurately estimates true positive proportion in data sets.
Reduces false positives in classifying rare online platforms.
problem Challenges in accurately identifying rare online platforms with ML.
method Calibrated probabilities and ensembles to reduce bias.
result Significantly reduces false positives in rare event detection.
Positive-unlabeled (PU) learning trains a binary classifier using only positive and unlabeled data. A common simplifying assumption is that the positive data is representative of the target positive class. This assumption rarely holds in practice due to temporal drift, domain shift, and/or adversarial manipulation. Thi…
Presentation bias is one of the key challenges when learning from implicit feedback in search engines, as it confounds the relevance signal with uninformative signals due to position in the ranking, saliency, and other presentation factors. While it was recently shown how counterfactual learning-to-rank (LTR) approache…
Model predicts political ideology using context vectors to mitigate bias and scarcity.
problem Scarcity and selection bias in political ideology prediction.
method Proposes a statistical model decomposing embeddings into context and position vectors, training an end-to-end model for deployment.
result Model can predict ideological labels even with minimal biased data, outperforming state-of-the-art methods.
The paper corrects bias in synthetic data for imbalanced learning.
problem Challenges in balancing false positive and negative rates in imbalanced data.
method Proposes a bias correction procedure to generate synthetic data for minority groups.
result Enhances prediction accuracy while avoiding overfitting.
TMLE improves causal effect estimation in missing data scenarios with various positivity violations.
problem Estimating causal effects in studies with missing data and positivity violations.
method Targeted Maximum Likelihood Estimation (TMLE) with various missing data methods.
result Complete cases with TMLE incorporating an outcome-missingness model exhibit lower bias and greater robustness against positivity violations.
We detect lookahead bias in LLM forecasts using a novel statistical method.
problem Detecting lookahead bias in LLM-generated economic forecasts.
method Developed a statistical procedure using date-only recall queries and estimated Lookahead Propensity (LAP).
result LLM forecasts are contaminated with lookahead bias, as indicated by a positive interaction between LAP and the forecast in accuracy regressions.
Financial LLMs need explicit bias consideration to avoid invalid results.
problem Finance-specific biases inflate performance and contaminate backtests.
method Identified five recurring biases and proposed a Structural Validity Framework.
result Explicit bias consideration is necessary for valid deployment claims.
It is well known that in stochastic multi-armed bandits (MAB), the sample mean of an arm is typically not an unbiased estimator of its true mean. In this paper, we decouple three different sources of this selection bias: adaptive \emph{sampling} of arms, adaptive \emph{stopping} of the experiment, and adaptively \emph{…
Ranking metrics are a family of metrics largely used to evaluate recommender systems. However they typically suffer from the fact the reward is affected by the order in which recommended items are displayed to the user. A classical way to overcome this position bias is to uniformly shuffle a proportion of the recommend…
We present a large-scale study of gender bias in occupation classification, a task where the use of machine learning may lead to negative outcomes on peoples' lives. We analyze the potential allocation harms that can result from semantic representation bias. To do so, we study the impact on occupation classification of…
Bias correction improves language model training performance.
problem Stochastic update bias in preconditioned optimizers.
method Cross-fitted preconditioning and variance-corrected inversion.
result Reduces held-out pretraining loss by 0.15 nats.
Centered plug-in estimators reduce bias in Wasserstein distance estimation.
problem Conservative bias in plug-in estimators of Wasserstein distances.
method Centering procedure based on linear combinations to reduce bias.
result Centered plug-in estimators provide informative upper and lower bounds on Wasserstein distances.
Machine learning models predict brain age with systematic bias, corrected in this study.
problem Systematic bias in machine learning regression models for brain age prediction.
method General constrained optimization approach to correct bias.
result Our method effectively eliminates the bias from brain age predictions.
Algorithm corrects bias in classification data.
problem Underrepresentation and intersectional bias in classification data.
method Estimate group-wise drop-out rates with small unbiased data, construct reweighting scheme, and present algorithm.
result Efficiently approximate loss of any hypothesis on true distribution.
Study finds gender bias in human evaluators and shows how machine learning can mitigate it.
problem Gender bias in human decision-making on micro-lending platforms.
method Structural econometric model and machine learning algorithms trained on real-world data.
result Machine learning algorithms can mitigate both preference-based and belief-based biases.
BR-SNIS reduces bias in self-normalized IS without increasing variance.
problem Bias in self-normalized IS.
method Iterated sampling-importance resampling (ISIR) to form a bias-reduced estimator.
result Significant reduction in bias without increasing variance.
In many applications, different populations are compared using data that are sampled in a biased manner. Under sampling biases, standard methods that estimate the difference between the population means yield unreliable inferences. Here we develop an inference method that is resilient to sampling biases and is able to …
Machine learning (ML) is increasingly deployed in real world contexts, supplying actionable insights and forming the basis of automated decision-making systems. While issues resulting from biases pre-existing in training data have been at the center of the fairness debate, these systems are also affected by technical a…
Gradient descent struggles with learning a single neuron with bias.
problem Learning a single neuron with a bias term in the realizable setting with ReLU activation.
method Theoretical study using gradient descent, characterizing critical points, and providing convergence guarantees.
result Gradient descent faces significant challenges in learning a single neuron with bias, unlike the bias-less case.
New theory shows how learning algorithms can create a bias towards negative outcomes.
problem Negativity bias in adaptive learning algorithms.
method Generalization of the Hot Stove Effect to settings with negative estimates leading to smaller sample sizes.
result Negativity bias persists even when negative estimates do not lead to avoidance.
This paper examines fairness and arbitrariness in bias mitigation methods.
problem Understanding how different bias mitigation strategies affect individual predictions and whether they introduce arbitrariness.
method FRAME framework to evaluate bias mitigation through five dimensions: Impact Size, Change Direction, Decision Rates, Affected Subpopulations, and Neglected Subpopulations.
result Significant differences in the behaviors of debiasing methods were exhibited, highlighting the limitations of current fairness criteria and the inherent arbitrariness in the debiasing process.
FairCal improves face verification accuracy while making results fairer.
problem Bias in face recognition models disproportionately affects minority groups.
method Post-training approach that builds fairer decision classifiers using pre-trained model features.
result State-of-the-art results with increased accuracy and fairness.
In real world systems, the predictions of deployed Machine Learned models affect the training data available to build subsequent models. This introduces a bias in the training data that needs to be addressed. Existing solutions to this problem attempt to resolve the problem by either casting this in the reinforcement l…
Private statistics estimation faces a bias, accuracy, and privacy trilemma.
problem Balancing privacy, accuracy, and bias in statistical estimation.
method Use differential privacy (DP) for private statistics, but clip samples to control sensitivity and add noise for privacy, introducing bias.
result No algorithm can simultaneously have low bias, low error, and low privacy loss for arbitrary distributions.
In display advertising, predicting the conversion rate (CVR), meaning the probability that a user takes a predefined action on an advertiser's website, is a fundamental task for estimating the value of displaying an advertisement to a user. There are two main challenges in CVR prediction due to delayed feedback. First,…
New method for Transformer models to encode position information without sequential bias.
problem Lack of flexible and learnable position encoding for Transformer models.
method Continuous dynamical model to learn position encoding.
result Consistent improvements over baselines in various NLP tasks.
New insights into bias mitigation show DRO isn't a complete solution.
problem Bias in machine learning systems across different data subsets.
method Theoretical analysis of Distributionally Robust Optimization (DRO) and data curation.
result Neither DRO nor data curation alone can fully address bias issues.
A new method predicts true classes from positive and unlabeled data with additional labeled observations.
problem Predicting true classes from positive and unlabeled data with selection bias.
method Introduces augmented PU prediction, allowing feature-dependent labeling, and compares various empirical Bayes rules.
result The variational autoencoder-based method performs similarly or better than other methods and improves accuracy for unlabeled samples.
Paper tackles survival data analysis with positive and unlabeled observations.
problem Traditional survival analysis yields biased results with positive-unlabeled data.
method Developed parametric, nonparametric, and machine learning models for positive and unlabeled survival data.
result Proposed estimation method provides valid results for positive-unlabeled survival data.
Study shows peers' graduation improves residents' success in TCs.
problem Identifying peer influence in therapeutic communities adjusting for latent homophily.
method Used data on affirmations and exit dates to form peer networks, modeled latent homophily, and proposed bias correction methods.
result Positive effect of peers' graduation on residents' graduation, varying by gender, race, and role model definition.
BICauseTree improves causal effect estimation by identifying clusters and balancing treatment allocation.
problem Improving interpretability and transparency in causal effect models from observational data.
method Hierarchical bias-driven stratification using decision trees with a customized objective function.
result BICauseTree provides interpretable causal effect estimation and is comparable to existing methods.
Over the past two decades, the notion of implicit bias has come to serve as an important component in our understanding of discrimination in activities such as hiring, promotion, and school admissions. Research on implicit bias posits that when people evaluate others -- for example, in a hiring context -- their unconsc…
New method reduces bias in incomplete data using deliberate missingness.
problem Systematic gradient biases in incomplete data for stochastic learning.
method Richardson-SGD debiasing procedure with deliberate missingness.
result Reduces gradient bias from O(∥p∥) to O(∥p∥2). Most positive and unlabeled data is subject to selection biases. The labeled examples can, for example, be selected from the positive set because they are easier to obtain or more obviously positive. This paper investigates how learning can be ena BHbled in this setting. We propose and theoretically analyze an empirica…
Paper analyzes how present-bias affects carbon emissions and proposes a method to mitigate it.
problem Present-bias impacts carbon emission patterns towards a net zero target.
method Stochastic control techniques adapted from insurance risk theory.
result Higher present-bias leads to excess emissions, and carbon taxes can reduce emissions but beyond a certain point have diminishing returns.
New methods ensure fair rankings in web-scale recommender systems.
problem Ensuring fairness in recommender systems, especially in web-scale applications.
method Scalable methods for achieving fairness in rankings, addressing position bias.
result Our methods effectively achieve fairness in various recommender systems.
New method learns from biased data, outperforming existing approaches.
problem Learning from one-sided feedback in machine learning.
method Adaptive approach using variance estimation to mitigate sampling bias.
result Empirically outperforms existing methods for generalized linear models.
Cross-validation is the de facto standard for predictive model evaluation and selection. In proper use, it provides an unbiased estimate of a model's predictive performance. However, data sets often undergo various forms of data-dependent preprocessing, such as mean-centering, rescaling, dimensionality reduction, and o…
New fair regression method improves fairness in chronic kidney disease classification.
problem Mitigating societal bias in health care for multiple groups.
method Penalized fair regression framework for multiple groups, with penalties for true positive rate disparity.
result Achieves fairness-accuracy frontier beyond existing methods in simulations and real-world data.
This report to our stage 2 submission to the NeurIPS 2019 disentanglement challenge presents a simple image preprocessing method for learning disentangled latent factors. We propose to train a variational autoencoder on regionally aggregated feature maps obtained from networks pretrained on the ImageNet database, utili…
Simplified analysis of SGD for linear regression with weight averaging.
problem Understanding SGD optimization in linear regression models.
method Simplified analysis using linear algebra tools, bypassing complex operator manipulations.
result Recovery of bias and variance bounds for SGD in linear regression.
Theoretical study shows AI models can recover from contaminated training data.
problem Data contamination in AI training can degrade model performance.
method Theoretical analysis and experiments on various data types.
result Models converge to true distribution under mild conditions, with rate dependent on real data fraction.