Xephor Solutions is a Research & Development driven company, which created the first fully operational AGI (after definition of D. Deutsch , a system which is able to build hypothesis and can test them by empirical deductivism can be called AGI) worldwide.
Our system is 100% our own software, there is no dependency on other companies, open source software/data or research institutes. Technically our AGI is consisting of thousands of neural nets which are replications of parts of the human brain and communicating with each other (realized using category theory).
- Ability to do machine thinking:
New ideas are being created by intrinsic random tensors and category theoretical reasoning in nets of neural nets
- Different kind of machine learning:
Neural nets can be stable with very few training data (realized through neurons which are similar to the pyramidal neurons in the neocortex)
- Based on empirical deductivism and critical rationalism (because the creation of general hypothesis is not possible with probabilistic inductivism proven by Karl Popper and through category theory)
- Ready-to-use-multiuser-system for all use cases – no additional programming necessary!
- Complete parallelization over the layers of the neural network (superior performance over derivatives of RNN and classical backpropagation algorithms)
- Data reduction:
Capability to reduce data from an to a x n (realized through extended numerical tensor calculus in the artificial stellar neurons which are a replication of the biological ones)
|Lots of high quality training data needed||Is able to work with unstructured, incomplete data, few data or different data sets|
|Lots of hardware power/energy needed||Through parallelisation need of less hardware and energy saving|
|No ability to react to new or unforeseen situations||Through machine thinking ability to react to unforeseen situations|
|Limitation on certain areas||Can easily be deployed in any area without requirement of programming|