Latest Articles for Robustness

A new method improves both accuracy and robustness in deep learning models.

Investigating vulnerabilities of DNNs against adversarial noise in image denoising.

Researchers tackle depth estimation challenges with new robust methods.

New method verifies ownership of DRL models to prevent theft.

Introducing SAP-sLDA for better text classification and user experience.

A novel method enhances DNN performance for real-world applications.

Examining how SSL methods perform under varying image conditions.

This study enhances the robustness of deep learning through dynamic model selection.

New methods improve object detection systems' resistance to adversarial changes.

MSECNet improves normal estimation accuracy and speed in 3D point clouds.

Learn how mmWave sensing is changing communication and health monitoring.

This article explores overlapping community detection methods to tackle network manipulation.

A new layer enhances DNNs' resistance to subtle input changes.

This article discusses the need for better evaluation methods for deep learning classifiers.

The rSVDdpd algorithm offers reliable data insights amidst outliers and noise.

DFM-X improves neural networks by addressing shortcut learning through frequency insights.

This article addresses enhancing AI models' performance against unknown data challenges.

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

New defenses aim to protect machine learning models from harmful data attacks.

New techniques improve the reliability of medical imaging tools against harmful data alterations.

Study reveals vulnerabilities in updated language models against adversarial attacks.

DyNNs provide efficient processing and robust performance against adversarial attacks.

A new framework improves deep learning model reliability against subtle changes.

Evaluating LLMs' ability to detect harmful instructions in user prompts.

This framework combines fairness and robustness for responsible AI systems.

SAM improves feature matching by considering group relationships in images.

Using advanced methods to enhance donor selection for better policy impact analysis.

LEAP improves adversarial testing for natural language processing software.

Graph Neural SDEs enhance model accuracy by incorporating randomness into data analysis.

A new method enhances analysis of statistical models for complex datasets.

This article highlights concerns about the robustness of neural network controllers.

Using real and generated data to boost QA model performance.

A study on how architecture affects model calibration in neural networks.

A new method enhances translation systems' adaptability and robustness in low-resource domains.

GF-RLS offers improved adaptability and stability in parameter estimation methods.

A look into methods for recovering low rank matrices from incomplete data.

Learn techniques to simplify complex passive systems while maintaining key properties.

Adversarial training improves machine learning models' resistance to input manipulation.

Mobile augmented reality combines the digital and physical worlds for engaging experiences.

This article discusses ways to enhance AI model reliability in changing environments.