Familiarize yourself with Lumina’s products by reviewing our technical documentation.
Request technical support for all Lumina’s products here.
View Lumina’s RCL APIs and access your tokens for model training.
Frequently Asked Questions
How is RCL different from neural networks?
In neural networks, every node of one layer is connected to many nodes of the next. In contrast, RCL employs minimal connections. Each node within a layer has no more than one parent. By moving away from the connectionist paradigm of neural networks, RCL makes significant gains in speed and transparency.
Where neural networks imitate the architecture of the brain, RCL has been designed to imitate the mind, specifically the function of drawing distinctions, as described in the tradition of Husserlian phenomenology. While neural networks tend to gravitate toward common patterns, RCL identifies patterns by distinctiveness and therefore illuminates patterns of greater subtlety and interest.
How do I train an RCL model?
Specifications for our API will be provided when you open an account. RCL models come with sequence prediction, context sensitive search, and translation between parallel corpora. If you have questions about which of these you should apply to your workflow, please contact us.
To which datatypes can RCL be applied?
At present RCL can be applied to text, tabular, time series and image (PNG) data. Future releases will allow for application to additional formats.
How is data structured prior to training?
Prior to training an RCL model, your data must be shaped into a list of sequences.
- In language for example, each item in your list may be a sentence or passage, which abstractly is just a sequence of words separated by spaces.
- In more complex datatypes, each item in a sequence may be represented as a numerical or hexadecimal sequence to represent information from your problem domain.
- For problems that can be represented in time, sequencing is straightforward; Each item in each sequence occurs one after the other.
- For problems that are less clearly represented in a time sequence, we suggest adding each relevant item individually to your sequence with the desired prediction as the last element(s) in the sequence.
Examples will be provided on our documentation page.