Authors: include Ole-Christoffer Granmo
Published: 2024, Github
TMU serves as a unified codebase, a one-stop repository offering diverse Tsetlin Machine implementations suitable for a variety of AI tasks, making it a treasure trove for researchers and AI enthusiasts.
The Tsetlin Machine Unified (TMU) repository stands out as a comprehensive platform hosting multiple implementations of Tsetlin Machines, each tailored for different aspects of machine learning. From basic to advanced structures like the Convolutional and Regression Tsetlin Machines, the repository is well-equipped to support both standard and cutting-edge AI research. Upcoming features include a Multi-task Classifier and One-vs-one Multi-class Classifier, alongside innovative methods like Focused Negative Sampling and Type III Feedback.
TMU supports continuous feature handling and offers specialized constructs like TMComposites for collaboration between different Tsetlin Machine types. For those looking to dive deep into machine learning at a hardware-accelerated pace, TMU provides wrappers for both C and CUDA, ensuring high-performance computations across platforms.
With its broad spectrum of features and extensions, alongside robust development support and detailed guides for setup and usage, TMU is positioned as an indispensable resource for advancing research and development in the field of machine learning with Tsetlin Machines.
View the Tsetlin Machine Unified - Github Code Base repository at Github.