Machine Learning solutions for researchers and business
From data to knowledge.
OUR EXPERTISE
Information Extraction
explore your textual data structure via our entities, facts, topics extractors, summarizers and complexity evaluators
Custom Search Engines
search in your corpora using our technology stack extended by semantic search with large queries support and entities recognition
Recommender Systems
recommend your documents with self-training algorithm, based on both content and user interests
OUR MACHINE LEARNING TECHNOLOGIES
Deep Learning
advanced language models with SoTA network architectures
General NLP & Linguistics
embeddings, morphological, syntactic and semantic models for English and Russian
Topic Modeling
additively regularized hierarchical models with high variety of extensions
General ML
best practises for classification, regression and ranking
OUR R&D DIRECTIONS
We develop custom high-effective ML-solutions for the analysis of banking transaction data, customer support automation, dialogue systems, information retrieval, predictive analytics and recommendation systems.
Main features:
Banking and online services
- Interpretable profiling (individuals). Advanced technology for building interpretable client profiles structure using transaction data (MCC, timestamp, amount, POS-id, geolocation, product info, etc).
- Interpretable profiling of corporate clients. Building profiles of similar businesses using multidirectional transation data.
Customer Support Automation
- Usupervised latent topic identification. We use advanced topic modeling tools to highlight dataset semantic structure.
- Smart routing. Simple and scalable mechanism for building efficient routing rules.
Information Retrieval
- Aggregator. We accumulate public research papers from numerous approved sources to cover as many scientific domains as possible.
- Search. We build fast, accurate and highly customizable search engine to make sure you find articles you are looking for.
- Libraries. Form collections out of your favourite documents to get personalized up-to-date recommendations.
- Summarization. We automatically generate annotations for single or multiple documents, helping customers to get key ideas «on the go» and obtain full vision of the collection without reading it through.
- Reading Order. Automatic complexity estimation methods are used to generate document reading order sequences helping our customers to explore wide scientific areas efficiently or understand a single document in details.
- Topic Maps. Topic Maps of science provide organized interactive view on the complete our document collection we aggregate or on a custom library.
- Trend Detection. We constantly analyse research papers flow to keep you abreast of all developing scientific areas and papers worth to explore.
Reserve your spot in line for an invite to get started
BOARD
Konstantin Vorontsov
D. Sc., professor: CIO
Victor Safronov
CEO
… and a dozen talented professional developers, analysts and engineers.
CONTACT US
info@aithea.com