“In high volume manufacturing, the accuracy needed from these ML models exceeds 99.9%. Decreasing our error rate from 4% to 0.02%, as we saw with our pilot project, clearly showed the value of synthetic data enabled quickly with Falcon.”
—
Steve Varga, Principal Scientist, P&G AI/ML Systems
"In the past, KEF often faced a trade-off between simulating expansive environments and maintaining high fidelity. With Duality's Falcon simulator, this compromise is no longer a concern. Falcon enables us to rigorously test our autonomy algorithms in photorealistic environments an order of magnitude larger than any we’ve previously used."
—
Fraser Kitchell, CEO, KEF Robotics
“Integrated with AWS SimSpace Weaver, Falcon-based simulations are solving extremely hard customer problems, enabling millions of interactions among multiple simulated digital twins in large spatial environments over a distributed AWS infrastructure.”
—
Rahul Thakkar, Director and General Manager of Simulation Technologies at AWS
"A robust and modular 3D simulation environment, among other tools, is critical to efficiently develop next-generation solutions and enable algorithms and systems to be rapidly designed, tested, and validated."
—
Jeff Radke, Engineering Director, Advanced Technology, Honeywell Aerospace
"The future of ROV autonomy training is in simulation. We recently built a deep sea environment in UE5 and the realism is truly stunning. This simulation sets the stage for synthetic data generation, reinforcement learning, and operator training."
—
Casey Sapp, CEO Blue Ring Imaging
Safely bringing AI into the real world
Robots, including self-driving vehicles, are struggling in environments made for people. They lack adaptability, resilience, and communication paradigms to seamlessly work in the dynamic nature of our world.
Internet AI models are making leaps in generative techniques for text prompting images, video, and audio. And now Embodied AI holds a similar promise for robotics.
Problem
To operate in the real world, these models must be safe and predictable. How can Embodied AI crossover into the physical world where there is simply no room for AI halucinations?
Solution
Duality's high-fidelity digital twin simulation enables our customers to test and fine-tune their models — achieving real-world results on accelerated timelines and within budget constraints.
What can Duality's Digital Twin platform do for you?
Quickly setup diverse scenarios using Unreal Engine based Editor and a massive catalog of simulation ready system, item and space twins
Run Open-Loop simulation for synthetic data generation or Closed-Loop for autonomous systems validation and process automation - collaboratively and in the cloud
Python API and Falcon middleware integrates with proprietary and open source messaging protocols like ROS/ROS2 and the latest generation of AI models
Secure end-to-end cloud based workflow eliminates DevOps headaches and special hardware requirements
Tune system twins and virtual sensors to match observed data and proven data permeability from simulation output as verified by demanding Fortune 100 customers
Leverage first party support of Duality's multidisciplinary team of specialized engineers, simulation pros, AI/ML experts and award-winning technical artists
Integrity Threats in AI: When Data Poisoning Undermines Model Effectiveness
Data poisoning: what is it? And why does it matter? In this blog we explore best practices to secure your data, and how high-quality synthetic data can bring powerful protection to your AI pipeline.
