Explorer Mike Horn gave a keynote lecture at KAUST's Winter Enrichment Program (WEP) about his pioneering expeditions. Horn recounted his solo journey around the Equator and his circumnavigation of the Arctic Circle. He also shared experiences from trekking to the North Pole during the Arctic winter. Why it matters: While not directly about AI, the talk highlights KAUST's broader mission to inspire innovation and exploration across diverse fields, which can indirectly foster a culture of creativity relevant to AI research.
British author and explorer Alastair Humphreys visited KAUST as part of the Enrichment in the Spring program. Humphreys, known for trekking across the Empty Quarter, shared his adventures with the KAUST community. The event aimed to bring a sense of adventure to the university. Why it matters: Such events enhance the cultural and intellectual environment at KAUST, fostering a broader perspective among students and faculty.
Victor Vescovo and the Caladan Oceanic crew, in cooperation with KAUST, made multiple manned dives into the Red Sea. They reached the deepest point, the Suakin Trough, for the first time. The team also dove the Kebrit Deep, which is shallower but scientifically important. Why it matters: This exploration provides an opportunity to study and protect the unique resources of the Red Sea's deepest regions, furthering scientific understanding of these previously inaccessible environments.
KAUST Professor Raquel Peixoto has been named one of "50 People Changing the World" by The Explorers Club for her pioneering work on coral probiotics. Her research demonstrates that probiotics can mitigate coral bleaching and prevent coral mortality. Peixoto's work bridges microbial ecology with applied innovation, influencing conservation strategies and international policy. Why it matters: This recognition highlights the importance of nature-based solutions developed in the region for addressing global environmental challenges like climate-driven reef degradation.
This paper addresses exploration in reinforcement learning (RL) in unknown environments with sparse rewards, focusing on maximum entropy exploration. It introduces a game-theoretic algorithm for visitation entropy maximization with improved sample complexity of O(H^3S^2A/ε^2). For trajectory entropy, the paper presents an algorithm with O(poly(S, A, H)/ε) complexity, showing the statistical advantage of regularized MDPs for exploration. Why it matters: The research offers new techniques to reduce the sample complexity of RL, potentially enhancing the efficiency of AI agents in complex environments.
Stanford's Robotics Laboratory, in collaboration with KAUST professors Khaled Nabil Salama and Christian Voolstra and MEKA Robotics, developed OceanOne, a bimanual underwater humanoid robot avatar with haptic feedback. OceanOne allows human pilots to explore ocean depths with high fidelity by relaying instantaneous images. The robot has two fully articulated arms and a tail section with batteries, computers, and thrusters. Why it matters: This collaboration between KAUST and Stanford highlights the increasing role of robotics and AI in deep-sea exploration, with potential applications in underwater research and resource discovery in the Red Sea and beyond.
KAUST will host its first annual Enrichment in the Fall program starting October 17, featuring lectures, films, workshops, concerts, and artworks. The program focuses on the biodiversity and wildlife in the KAUST community, with events aimed at all ages and interests. The enrichment program includes community activities on the weekends of October 17–18 and October 24–25. Why it matters: This program signals KAUST's ongoing commitment to community engagement and education in areas like biodiversity, complementing its established Winter Enrichment Program.
This paper introduces Diffusion-BBO, a new online black-box optimization (BBO) framework that uses a conditional diffusion model as an inverse surrogate model. The framework employs an Uncertainty-aware Exploration (UaE) acquisition function to propose scores in the objective space for conditional sampling. The approach is shown theoretically to achieve a near-optimal solution and empirically outperforms existing online BBO baselines across 6 scientific discovery tasks.