PulseAugur
EN
LIVE 21:49:13

New AI models for Arabic handwriting recognition vulnerable to attacks

By PulseAugur Editorial · [2 sources] ·

Researchers have developed lightweight embedded ConvNet ensembles to improve Arabic handwritten character recognition, achieving accuracy comparable to larger models. A separate study investigated the security of these models, revealing significant vulnerabilities to black-box adversarial attacks. These attacks, which are nearly imperceptible to humans, achieved success rates of up to 100% on benchmark datasets, highlighting the need for enhanced security in AHR systems. AI

IMPACT Highlights the trade-off between model efficiency and security in AI systems, particularly for specialized tasks like handwriting recognition.

RANK_REASON Two arXiv papers detailing new research on Arabic handwriting recognition models and their security vulnerabilities.

Read on arXiv cs.CV →

AI-generated summary · Google Gemini · from 2 sources. How we write summaries →

New AI models for Arabic handwriting recognition vulnerable to attacks

COVERAGE [2]

  1. arXiv cs.CV TIER_1 English(EN) · Abdelillah Semma ·

    Embedded ConvNet Ensembles: A Lightweight Approach to Recognize Arabic Handwritten Characters

    Arabic Handwritten Character Recognition (AHCR) has recently advanced significantly with deep Convolutional Neural Networks (ConvNets). However, many models in the literature are deep and computationally expensive in terms of parameters and FLOPs, limiting their deployment on res…

  2. arXiv cs.CV TIER_1 English(EN) · Rachid Elouahbi ·

    Threats to Arabic Handwriting Recognition: Investigating Black-Box Adversarial Attacks on embedded ConvNet models

    Arabic handwriting recognition (AHR) has made significant progress with deep learning models. AHR research has largely focused on performance, with security receiving little attention. This study provides what appears to be a new line of inquiry by demonstrating the vulnerability…

RELATED ENTITIES

RELATED TOPICS