Fast
Most of our projects have C++/Java implementations with an option to run on GPUs
Computers are great at simple number manipulation if instructed by a human. However, when it comes to learning on their own, to understanding complex relationships or reason about the future they often fall short.
A consequence of this is that many organizations today avoid advanced machine learning algorithms and instead rely on simple models such as linear regression. Of course this is not ideal, as this often to fail to spot non-trivial relationships.
Despite being complex models, there's no reason why machine learning algorithms cannot be abstracted into simple to use libraries and open APIs. There have been several attempts to do this through various open-source projects, but there's still a long way to go before machine learning becomes usable for most applications.
NXT AI will concentrate on time-dependent patterns, such as language modeling, which is useful for prediction.
Most of our projects have C++/Java implementations with an option to run on GPUs
We seek to open-source most of our projects. Happy coding!
Built with modern technologies the projects can easily be ported to the cloud
Open APIs and simple integration with your custom data format
An open prediction framework
Java framework for various prediction algorithms accessible via a JSON API.
→ read more
A C++ implementation of Hessian-free optimization of recurrent neural networks for statistical language modeling.
→ read more
A small toolkit useful for preparing large language modeling data sets for experiments.
→ read more
Machine learning is
too complicated to use
for most organizations.
This needs changing.
Jeppe Hallgren, founder
