Researchers from MBZUAI introduce Forget-MI, a machine unlearning method tailored for multimodal medical data, enhancing privacy by removing specific patient data from AI models. Forget-MI utilizes loss functions and perturbation techniques to unlearn both unimodal and joint data representations. The method demonstrates superior performance in reducing Membership Inference Attacks and improving data removal compared to existing techniques, while preserving overall model performance and enabling data forgetting.
Researchers from the National Center for AI in Saudi Arabia investigated the sensitivity of Large Language Model (LLM) leaderboards to minor benchmark perturbations. They found that small changes, like choice order, can shift rankings by up to 8 positions. The study recommends hybrid scoring and warns against over-reliance on simple benchmark evaluations, providing code for further research.
The paper introduces LLMEffiChecker, a tool to test the computational efficiency robustness of LLMs by identifying vulnerabilities that can significantly degrade performance. LLMEffiChecker uses both white-box (gradient-guided perturbation) and black-box (causal inference-based perturbation) methods to delay the generation of the end-of-sequence token. Experiments on nine public LLMs demonstrate that LLMEffiChecker can substantially increase response latency and energy consumption with minimal input perturbations.
MBZUAI researchers have developed MorphDiff, a diffusion model that predicts cell morphology from gene expression data. MorphDiff uses the transcriptome to generate realistic post-perturbation images, either from scratch or by transforming a control image. The model combines a Morphology Variational Autoencoder (MVAE) with a Latent Diffusion Model, enabling both gene-to-image generation and image-to-image transformation. Why it matters: This could significantly accelerate drug discovery and biological research by allowing scientists to preview cellular changes before conducting experiments.
A new paper from MBZUAI demonstrates that state-of-the-art speech models can be easily jailbroken using audio perturbations to generate harmful content, achieving success rates of 76-93% on models like Qwen2-Audio and LLaMA-Omni. The researchers adapted projected gradient descent (PGD) to the audio domain to optimize waveforms that push the model towards harmful responses. They propose a defense mechanism based on post-hoc activation patching that hardens models at inference time without retraining. Why it matters: This research highlights a critical vulnerability in speech-based LLMs and offers a practical solution, contributing to the development of more secure and trustworthy AI systems in the region and globally.
Researchers at MBZUAI have developed a new automatic method to examine cross-lingual abilities in multilingual language models, testing 10 models across 16 languages. They combined beam search with language-model-based simulation, generating 6,000 bilingual question pairs and found significant performance drops compared to English, even in high-resource languages like Chinese. The method introduces perturbations to test the models' ability to transfer knowledge rather than rely on memorization. Why it matters: This research highlights critical gaps in cross-lingual AI, providing a framework for developing more equitable and effective multilingual models, especially for Arabic and other under-represented languages.
The Russian Immune Diversity Atlas project aims to profile immune cells from people of different ancestries at a multiomics level. The goal is to reconstruct a reference atlas of the healthy immune system and investigate its perturbations in Type II Diabetes (T2D). The project seeks to identify novel mechanisms and genetic/epigenetic markers for early T2D diagnostics, prognosis, and therapy as part of the international Human Cell Atlas. Why it matters: Addressing genetic diversity in biomedical research, particularly in the context of the Human Cell Atlas, is crucial for personalized medicine and ensuring that treatments are effective across diverse populations in the Middle East and globally.