One of artificial intelligence’s most stubborn problems is AI systems being able to accumulate new knowledge without losing what had been previously learned. A team of researchers at the MATRIX AI Consortium at The University of Texas at San Antonio may have solved this issue with Genesis, a spiking neuromorphic accelerator chip that would enable on-device continual learning across its operational lifetime.

A new study analyzed data from multiple independent sources, ranging from satellite remote sensing-based machine learning products to state-of-the-art Earth System Models, to reveal how climate change shapes the Northern Eurasian carbon sink. The study confirms that Northern Eurasia accounts for nearly one-third of the global terrestrial carbon sink. It also shows that, despite warming faster than most regions worldwide, the northern high latitudes contribute little to strengthening the Northern Eurasian carbon sink because of their low baseline carbon uptake.

Professor Sungjin Ahn's team in the School of Computing has developed NEO (Neural Theorizer), an AI model that learns executable theories of how the world works from observation alone, rather than simply predicting what happens next

Selected for an oral presentation at ICML 2026 — the top 0.7% of 23,918 submissions — and for the Best Paper Award at the award of Compositional Learning Workshop

Like a child learning how the world works, the model explains unfamiliar phenomena by recombining rules it has already learned, pointing toward a potential foundation for intelligent robots and autonomous agents

A new study published in Big Earth Data proposes an operational framework that integrates multiple machine learning and deep learning models with high-resolution Sentinel-2 data to estimate agricultural drought conditions in irrigated and non-irrigated maize fields in Eastern Austria. Results from 2018–2023 show that deep learning models, particularly the Deep Neural Network (DNN), achieved the best performance, while non-irrigated fields exhibited higher prediction accuracy due to clearer drought signals, demonstrating the value of Sentinel-2–based deep learning for operational drought monitoring.

This review provides a comprehensive overview of recent advances in bioactive peptide development for cosmetic applications. It summarizes peptide classification, biological mechanisms, production strategies, stabilization technologies, and emerging discovery platforms. By integrating synthetic biology, high-throughput screening, and artificial intelligence-driven design, this work highlights new opportunities for developing multifunctional, efficient, and next-generation peptide-based cosmetic ingredients.