Latest Articles for Model Performance

A new method improves model fairness by generating synthetic data samples.

Recent findings question the belief that more data always improves machine learning outcomes.

This study reveals SGD's advantages in robustness over adaptive training methods.

Examining the impact of AI decision-making on groups and systemic outcomes.

MixBag enhances learning from label proportions by creating new labeled groups.

New method tackles multiple biases in image classification to improve fairness.

A new method for evaluating machine learning models without labeled data.

Examining how Causal Neural Networks enhance resilience to adversarial attacks.

A new approach improves the evaluation of predictive models for patient outcomes.

This article presents a new approach to creating NLP models using natural language prompts.

A competition highlights strategies to defend machine learning models in finance.

New methods improve detection of unusual medical images for better diagnostics.

A new method for assessing language models without human labels.

This study examines advanced methods to predict ED crowding and enhance patient care.

Learn how Padding Aware Neurons impact image processing in machine learning models.

This article investigates how parameter sparsity affects AI model performance and efficiency.

This article discusses a fresh method for managing missing data using contrastive learning.

A study on improving training efficiency for language models using SlimPajama dataset.

A novel approach enhances the quality of noisy biomedical labels.

A new MLOps architecture integrates explanations to enhance trust and efficiency in ML models.

This article examines challenges and solutions related to noisy labels in training data.

New approach enhances deep learning models for reliable medical image segmentation.

A new dataset for improving e-commerce image and text recognition.

Contrastive unlearning efficiently removes data influence while preserving model performance.

A deep learning method enhances kidney problem detection in medical imaging.

This study questions the preference for AUPRC over AUROC in imbalanced datasets.

A study examines the performance and biases of AI language models globally.

Credal learning theory offers new insights for adapting machine learning models to changing data.

This article explores how comment generation improves large language models’ coding abilities.

A new framework aims to improve AI's empathetic responses in conversations.

Learn how negative sampling streamlines model training and improves performance.

Discussing privacy and fairness in machine learning through differential privacy and worst-group risk.

A new approach to image representation with differential privacy through captioning.

This study examines how contextual knowledge enhances the detection of adverse drug events.

Research sheds light on transfer learning applications and model performance relationships.

A tool for improving forecasting accuracy using counterfactuals and data visualization.

Examining how noise in pre-training data impacts model performance.

New methods enhance AI's ability to forget data while maintaining performance.

Research highlights improved early diagnosis methods for lung cancer using patient data.

A study reveals new techniques for backdoor attacks on language models with minimal impact.