Niyeo Research
January 5, 2025
Before Niyeo became a platform, there was a long period of exploration into the possibilities of AI. Our work grew from a series of projects that rethought traditional challenges in new ways.
In 2017, research into SD focused on generating synthetic data to train machine learning models more effectively. In 2019, the STC project—developed as part of Apance—brought generative AI effects to a collaborative illustration app for the iPad, transforming simple sketches into fully colored images. This work offered an early glimpse of how AI could enhance creative expression.
By 2022, efforts shifted toward understanding language at a deeper level with LG1, a project that mapped natural language to logical statements. This concept was further refined in 2023 with LGG, which explored tracing logic through graph networks to guide reasoning. That same year, GD-Flow was introduced as a cross-platform dataflow compiler for generative AI, and LGDB reached a proof-of-concept stage, demonstrating a new way to store and query logical constructs.
These early experiments have culminated in n1, our 2025 system for training self-learning AI agents with natural language instructions. n1 now forms the foundation of the Niyeo platform, allowing users to create agents that learn from real-world interactions and adapt over time.
While many details remain confidential to ensure responsible use, our journey reflects a commitment to exploring innovative ideas with care and ethical responsibility. This evolution—from early experiments to the development of n1—illustrates our belief that intelligent machines can be both powerful and responsibly designed, serving as valuable partners in creative and practical endeavors.
