Current baseline for a variety of NLP tasks. Evaluation, training, and calibration of case-specific language models with machine learning and deep learning methods.

Natural Language Understanding is the process of analyzing text or speech, which is often the first step in solving complex NLP tasks. It involves linguistic analysis on different levels, including morphological and syntactic analyses, as well as the analysis of semantic and discourse relations.

Natural language generation can be used to accelerate content creation as well as for the automation of various communication and documentation tasks and processes. Retresco textengine platform enables users to easily generte texts of varying complexity in serveral languages.

NER (named entity recognition) is used to identify important data points, such as names of people and organizations, dates and numbers, as well as predefined concepts in a given text. It is important for such tasks as information retrieval and question answering.

Content Classification is one of the most important practical use cases in NLP today. Specifying features and topics of a given content (emails, advertises, intends, unstructured texts) is highly useful in many business scenarios.

Information extraction and knowledge retrieval is important for internet applications like, chatbots, voice bots, customer service and so on.
Retrieve, extract and return information for a given piece of content is the foundation to solve many automation tasks (tagging, labelling, classification, knowledge retrieval, question & answering, chatbots).

Common method to extract relevant information from an input text in many different usage scenarios (legal & insurance business workflows, healthcare, etc.)

Sentiment detection using machine learning methods for social media platforms and in other scenarios (reputation analysis, analysis of user-generated content, offensive-language and hate-speech filtering, marketing and communications).

Automated question answering enables and enhances real-time scalable customer interaction. Structured and unstructured data sources can be used for answering world-knowledge or domain-specific questions.