Artificial Brain-as-a-Service

 

Automated transcription with a new generation bio-inspired neural networks

KanjuTech transcription features

Cloud-based
solution

Distributed through the API and Telegram bot

Multiple
speaker detection

Our brain-inspired AI recognizes each speaker

Accurate
meeting notes

The whole conversation with labeled speakers

The core of our technology is spiking neural networks

Brain-inspired spiking neurons

Transfer the binary values (1/0) and have excitable nature mimicking biological neural cells

Advantages of spiking neural networks

 
  • adaptive self-learning

  • reduced energy consumption

  • important signals recognition

  • better sequential data processing

  • online time learning

  • ability to do pro-active actions

Our transcribing technology advantages

Unsupervised learning

No need for a labeled dataset

Noise cancelation

Works even in noisy environments

Simultaneous speech recognition

Accurately transcribes recordings
with multiple participants

Our technology

Our technology is based on modern advances in neuroscience and state of the art of Deep Learning

KanjuTech technology strengths

Active pattern search

The cortical engine is constantly looking for new knowledge and notions

Selective attention

KanjuTech recognizes the most important part of the data

Source separation

Spiking neurons pick the most valuable information among the noise

Multiple modalities

KanjuTech is able to understand sound, images, text, and other types of data

Our research

Partners and support