learning and working with language models. our models are able to generate phrases, translate text and audio recordings, conduct conversations and determine the parameters of the speaker.
dynamic recognition of meaning (recognition before receiving a complete sentence/paragraph)
recognition of sounds and speech in difficult conditions
taking into account the context of history
creating emotion management and semantic constructions
computer vision department
the research areas of the cv department include:
creation of datasets and training of classifiers for error recognition, slang and abbreviations
autonomous semantic segmentation, classification and identification, partitioning into objects and recognition of small details, including real-time
summarization of information and context
researches timeseries
time series technology predicts the behavior of stocks in the market. we have achieved high forecast accuracy in trading cryptocurrencies, stocks and other securities.
neuro-symbolic architectures
general artificial intelligence will become a priority task of global corporations and states. we create advanced tools for the development of neural networks and artificial intelligence elements.
dialog models
hybrid neuro-symbolic dialog systems with the function of user intent orchestration, knowledge graph, dynamically augmented databases, built-in interface for programming skills and behavior correction.