NLP by Trask Helps Banks Analyse Documents Automatically
Trask’s natural language processing solution helps banks automatically analyse documents.
Natural language processing is another modern technology based on artificial intelligence. It helps computers understand human language – spoken or written – and extract relevant information. Full comprehension of human language starts with words, but there’s syntax and semantics as well. Speech recognition systems are integrated in popular applications such as Google Translator, MS Office grammar checkers, intelligent conversational UIs (so-called chatbots) or in digital assistants (Cortana, Siri, Alexa, etc.). Trask has something to offer too. Our unique solution eliminates and improves manual work, untying personnel’s hands for higher level client care operations.
Based mostly on machine learning, natural language processing technologies (NLP) decipher human language (speech or text), understand the meaning and extract actionable information. The utmost goal is to achieve understanding at the same level as humans. Computer systems will then understand written and spoken human discourse, draw conclusions, sum up, translate, and generate natural language output.
Natural language processing can be approached from various angles:
- Symbolic approach: is based on generally accepted grammatical rules and lexicons prepared by a human for use by the system.
- Statistical approach: is based on machine learning. It works with a vast corpus of texts and uses mathematical methods to analyse linguistic phenomena found in the corpus. It develops its own rules and applies them to next inputs.
- Hybrid approaches: are a combination of the two. Systems use pre-defined, generally applicable rules for statistical analysis of each input and can adapt the rules for a specific purpose.
Genuine NLP Solution
Our Trask Semantic Tool is an example of such approach. We have worked on its development with the Institute of Formal and Applied Linguistics (ÚFAL) at the Faculty of Mathematics and Physics, Charles University, Prague. The ÚFAL is a centre of excellence for computerised text processing and it generates big corpora of linguistic data for massive use in machine learning. We have built our NLP solution on our extensive knowledge of banking environment and on many years’ experience in the field of banking systems development, implementation and integration. Our NLP integrates top security standards and complies with the applicable legislation (GDPR) and with internal bank rules for data processing, including full audit trail.
We have built our NLP solution on our extensive knowledge of banking environment and on many years’ experience in the field of banking systems development, implementation and integration. Our NLP integrates top security standards and complies with the applicable legislation (GDPR) and with internal bank rules for data processing, including full audit trail.
Trask Semantic Tool in Practice
Trask NLP system serves a large bank for notarial deeds processing. It takes care of the tedious and time-con- suming tasks that follow from the mounting legislative requirements and bring no added value for the bank’s clients. The system automatically captures key information from notarial deeds – deceased persons’ data (name, surname, date of birth, birth-certificate number, date of death), notary’s name, application reference number and recipient – and checks the outcome for correctness against pre-defined rules. It also allows to search documents. When in doubt, it asks a human operator for assistance and passes any perceived flaws over for human analysis.
You might be interested
Modern information technologies for banking services enhance digitalisation, improve client comfort, support legislative compliance and, most importantly, offer ready-made solutions that are available right now.
With the progressive digitalisation, stronger competition and the fragmentation of the banking market banks need to come up with new methods to attract clients.