Explore the insights presented in this talk from Upper Bound, featuring Kateryna Nekhomiazh, Solutions Architect at NVIDIA, presenting "Pioneering Deep Learning Acceleration: Latest Advancements on the NVIDIA Platform." This session dives into how NVIDIA's platform and tools accelerate deep learning workflows, emphasizing the shift towards physical AI. Kateryna discusses NVIDIA's comprehensive approach, extending beyond just hardware to include robust software, libraries, and open-source models, highlighting the CUDA platform's evolution into a powerful ecosystem for AI development. The talk also introduces NVIDIA Cosmos, a leading-edge World Foundation Model development platform designed to advance physical AI through synthetic data generation, diverse world foundation models (Predict, Transfer, Reason), and advanced video tokenizers for efficient data curation and post-training. The speaker underscores the importance of physically grounded synthetic data for training robots and autonomous vehicles, showcasing how NVIDIA Omniverse combined with Cosmos Transfer models can create photorealistic and physics-based simulations, addressing the challenges of acquiring real-world edge case data.

Upper Bound 2025 is Amii's annual artificial intelligence conference, held in Edmonton, Alberta, Canada, bringing together researchers, industry, and policymakers. The conference focuses on accelerating AI excellence and innovation for good, emphasizing AI for critical infrastructure, health, industrial operations, responsible AI, and AI Literacy.