A.N.I.T.A. (Artificial Neural-Intelligence Translator Assistant) is a vertically integrated artificial intelligence / machine learning solution developed by Stratis, primarily designed to facilitate, partially or completely replace the document-based tasks of a corporate administrative employee.
We regularly expand and develop its functions, while the AI solution and continuously learns throughout the daily collaboration with our customers.
Avoiding expensive and long-term assessment of operational rules, corporate procedures, and programming, A.N.I.T.A. learns them based on patterns so that it can solve complex tasks that could not be solved from a realistic time and financial framework.
Key Features
Machine translation
A.N.I.T.A.’s machine translation capabilities can be utilized primarily for mass translation needs, or for translation of languages that are rare, making it difficult to provide human translation. Even the most advanced machine-translated solutions cannot be compared to a translation by someone who speaks a particular language on a high level but it is not their aim, neither. It is able to translate massive amounts of documents in such a high quality that their topics and key concepts can be understood, so that they can be searched and the essential parts can be easily selected and then translated with human power in a compelling way.
A.N.I.T.A. is able to continuously execute the translation of thousands of documents without interruption on daily basis, sparing the human workforce and multiplying the amount of material that can be processed in a single day with human power.
Language detection
Based on words, A.N.I.T.A. can recognize the language of a given sentence or a the entire document with good efficacy, even if the document contains words in other languages (for example, English terms). This is a useful feature for several reasons when processing written documents. On the one hand, typically different workflows apply to documents in different languages (for example, it needs to be taken to a different colleague, or some languages has to be given priority, while others not, etc.). On the other hand, one of the key information in the documents is the language of the document itself. For many processing processes, when translated, or even searched, it is worth knowing the language of the given sentence or document.
Named entity recognition
A.N.I.T.A. can extract important key-expressions from a sentence or document, not (only) based on pre-written dictionaries, but instead on the fundamental artificial understanding of the written text, therefore A.N.I.T.A. can identify a name, location, organization etc. in a text even, if nobody programmed or told her to find that exact person, location, organization etc. before (no need to list all persons, locations, organizations etc.)
Document (text) categorization
A.N.I.T.A. learns on examples, and understands the texts she sees, so it can efficiently determine if a text, document is a particular category. It can also invoke the corresponding processes in the organization and notify the suitable person.
Search
After A.N.I.T.A. mined all the above information of the documents and texts she can scan, all of these are stored in a searchable database. This provides the possibility to search in multiple languages, fuzzy search, search named entities, search for language and category etc.
Technological features
- Completely flexibly scalable
- A closed system corresponding to the most strictest security principles
- Local installation on demand, not requiring external resources
- Its intelligence can be taught for the needs of our customers, fully customizable