MBZUAI and KAUST researchers collaborated to present new optimization methods at ICML 2024 for composite and distributed machine learning settings. The study addresses challenges in training large models due to data size and computational power. Their work focuses on minimizing the "loss function" by adjusting internal trainable parameters, using techniques like gradient clipping. Why it matters: This research contributes to the ongoing advancement of machine learning optimization, crucial for improving the performance and efficiency of AI models in the region and globally.
The article discusses research on fine-tuning text-to-image diffusion models, including reward function training, online reinforcement learning (RL) fine-tuning, and addressing reward over-optimization. A Text-Image Alignment Assessment (TIA2) benchmark is introduced to study reward over-optimization. TextNorm, a method for confidence calibration in reward models, is presented to reduce over-optimization risks. Why it matters: Improving the alignment and fidelity of text-to-image models is crucial for generating high-quality content, and addressing over-optimization enhances the reliability of these models in creative applications.
MBZUAI researchers presented a new strategy for handling complex optimization problems in machine learning at ICLR 2024. The study, a collaboration with ISAM, combines zeroth-order methods with hard-thresholding to address specific settings in machine learning. This approach aims to improve convergence, ensuring algorithms reach quality solutions efficiently. Why it matters: Improving optimization techniques is crucial for advancing machine learning models used in various applications, potentially accelerating development and enhancing performance.
A CMU researcher, Dr. Hongyi Wang, presented an evaluation of gradient compression methods in distributed training, finding limited speedup in most realistic setups. The research identifies the root causes and proposes desirable properties for gradient compression methods to provide significant speedup. The talk was promoted by MBZUAI. Why it matters: Understanding the limitations of gradient compression techniques can help optimize distributed training strategies for AI models in the region.
This paper introduces BRIQA, a new method for automated assessment of artifact severity in pediatric brain MRI, which is important for diagnostic accuracy. BRIQA uses gradient-based loss reweighting and a rotating batching scheme to handle class imbalance in artifact severity levels. Experiments show BRIQA improves average macro F1 score from 0.659 to 0.706, especially for Noise, Zipper, Positioning and Contrast artifacts.
MBZUAI researchers presented a new second-order method for optimizing neural networks at NeurIPS 2024. The method addresses optimization problems related to variational inequalities common in machine learning. They demonstrated that for monotone inequalities with inexact second-order derivatives, no faster second- or first-order methods can theoretically exist, supporting this with experiments. Why it matters: This research has the potential to reduce the computational cost of training large and complex neural networks, which could accelerate AI development in the region.
Xiaolin Huang from Shanghai Jiao Tong University presented a talk at MBZUAI on training deep neural networks in tiny subspaces. The talk covered the low-dimension hypothesis in neural networks and methods to find subspaces for efficient training. It suggests that training in smaller subspaces can improve training efficiency, generalization, and robustness. Why it matters: Investigating efficient training methods is crucial for resource-constrained environments and can enable broader access to advanced AI.
The paper introduces a novel actor-critic framework called Distillation Policy Optimization that combines on-policy and off-policy data for reinforcement learning. It incorporates variance reduction mechanisms like a unified advantage estimator (UAE) and a residual baseline. The empirical results demonstrate improved sample efficiency for on-policy algorithms, bridging the gap with off-policy methods.