We are teaching computers to understand the world

Meanotek develops Artificial Intelligence solutions that can take on human work. We train AI to solve important tasks for you and offer turnkey solutions for managing your business

Artificial intelligence already helps our customers to work more efficiently

Machine Vision

Object detection, Neural superresolution, image transformation

Solutions for medicine

Extracting key insights from medical literature using AI

Conversational AL

Help desk automation. Computer sales assistant for the site. Answers to questions on the knowledge base. Training new employees.

Voice controlled CRM/BPM

CRM / BPM systems with intelligence, includes speech recognition, task planning, communication with the buyer, organization of the work process.

Natural Language Understanding

Extracting entities and relations from text to get buisness insigths

Analysis of team correspondence

Most discussed topics

Speech recognition and synthesis

Custom ASR and TTS solutions for your needs

Machine translation

Custom machine translation models

Natural Language Generation

Descriptions of goods and services on the sites of online stores and aggregators. Generators of letters with an individual appeal to the client.

Custom solutions for your buisness needs

Tell us about your buisness probelm and we will suggest a solution for you!

Projects

We developed custom AI solutions for hudrends of customers all over the world

  • Conversational AI
  • Science, education, medicine
  • Online shops and marketing
  • News analytics

Mindy personal assistant

A system for managing employee tasks that understands voice commands and works on any device. Go to website>

Taxi ordering voice bot

The bot asks questions about the trip, saves the answers and passes the data to the server that assigns the car.

Medical decision support

A system for recommending cancer treatment has been developed, taking into account the genetic profile of the tumor.

Classification of histological slides

In the course of the work, a program was developed to classify carcinoma, a benign tumor according to tissue photos.

X-ray images classification

X-ray pathology classification system using a publicly available set of medical image data.

Reed-Kellog diagrams

Web service for constructing flowcharts from sentences in English according to the method of Reed and Kellogg

Ad generator

Automatic generation of online advertisements from keywords Читать дальше >

Automaitic generation of reports, and online articles

Is it possible to generate new meaningful article using AI?

Information extraction system for a CenterCredit Bank JSC (Kazakhstan)

A system was developed to extract certain information from a large array of natural language data.

Our customers:

Technology

The company is doing research and develops new methods in the fields of:

  • Relations mining
  • Sentiment polarity detection
  • Decision support
  • Low-shot learning
  • Paraphrase Generation
  • Support for multiple languages
  • Denis Tarasov, CTO
  • Topic: Neural network model for solving the problem of answering user questions in an arbitrary subject area
XVIII International Scientific Conference, "Neuroinformatics-2016

Publications

  • Izotova E.D. Extraction of pharmaceutically significant aspect terms by a model of a recurrent neural network from texts in natural language with small samples / E.D. Izotova, D.S. Tarasov // XVIII "Neuroinformatics-2016" .- Sat. n.t. - M.: NRNU MEPHI. - 2016.- T. 1 - S.65-74
  • Tarasov D.S. Neural network model for solving the problem of answering user questions in an arbitrary subject area / D.S. Tarasov // XVIII "Neuroinformatics-2016" .- Sat. n.t. - M.: NRNU MEPHI. - 2016.- T. 3 - S.205-212
  • Tarasov D.S, Izotova E.D. (2016) Deep recurrent neural networks for extracting pharmacological terms from Russian texts" Neuroinformatics (2016) V.2. This paper, produced with collaboration with Institute of Medicine and Biology of Kazan Federal University considers extraction of information from online medical forums for monitoring drug side effects (paper in Russian, for English abstract see conf. program http://neuroinfo.ru/index.php/en/schedule/sections?year=2016 paper text available by request)
Show all...
  • D.S. TARASOV, E.D. IZOTOVA. COMMON SENSE KNOWLEDGE IN LARGE SCALE NEURAL CONVERSATIONAL MODELS//Принято к печати, Нейронформатика-2017 (в сборнике Studies in Computational Intelligence, из-ва Springer) In this work we analyze behavior of a number of neural network architectures, trained on Russian conversations corpus, containing 20 million dialog turns. We found that small to medium neural networks do not really learn any noticeable common-sense knowledge, operating pure on the level of syntactic features, while large very deep networks shows do posses some common-sense knowledge.
  • Tarasov D.S. (2016) Preserving personal conversational style and diversity in neural conversational models//Proceedings of International Conference on Artificial Neural Networks - Neuroinformatics (2016) V.2. This paper presents a method to apply specific personal conversation traits to neural dialog models, trained on large heterogeneous dialog corpus (Abstract is avaliable from conference program url - http://neuroinfo.ru/index.php/en/schedule/sections?year=2016#Stend5).
  • Tarasov D.S. (2016) Neural network model for general domain question answering//Proceedings of International Conference on Artificial Neural Networks - Neuroinformatics (2016) V.3. In this work we propose novel neural network model, capable of answering questions without topic restriction by reading natural language documents provided by simple information retrieval methods. (paper in Russian, for English abstract see conf. program http://neuroinfo.ru/index.php/en/schedule/sections?year=2016 paper text available by request))
  • Tarasov D.S. (2015) Natural Language Generation, Paraphrasing and Summarization of User Reviews with Recurrent Neural Networks // Computational Linguistics and Intellectual Technologies: Proceedings of Annual International Conference “Dialogue”, Issue 14(21), V.1, pp. 571-579 (PDF link: http://www.dialog-21.ru/digests/dialog2015/materials/pdf/TarasovDS2.pdf , presentation slides: http://www.meanotek.io/files/natgen.pdf). This paper, reports first successful application of deep learning to generate abstractive multi-document summaries.
  • Tarasov D.S. (2015) Deep Recurrent Neural Networks for Multiple Language Aspect-Based Sentiment Analysis // Computational Linguistics and Intellectual Technologies: Proceedings of Annual International Conference “Dialogue”, Issue 14(21), V.2, pp. 65-74 (PDF link: http://www.meanotek.io/files/TarasovDS2015-Dialogue.pdf) - this paper describes aspect-based sentiment analysis system that achieved top results on SentiRuEval-2015 competition (http://www.dialog-21.ru/evaluation/2015/sentiment/)

BLOG

Articles (in Russian):

Russian neural network chatbot

Читать дальше >

Articles (in Russian)

Automatically generate unique texts for SEO

In recent years, unique text has become a common tool that SEO experts recommend for website promotion. The article explains how automatic text generation can help

Читать дальше >

Articles (in Russian)

Algorithmic text generation

In this article, we will talk about algorithmic text generation methods and talk about our experience working with them. Suppose we work with phone reviews.

Читать дальше >

Articles (in Russian):

Text analysis system

Читать дальше >