bigRing Machine Learning module allows your computer to see, hear and understand for you. Module integrates the machine learning senses such as Machine Vision (OCR, ICR, OMR and other), Machine Speech (Speech-to-Text) and Machine Conversation Understanding (NLP). Coupled with machine understanding capabilities, cognitive AI capabilities of bigRing technology allows you the deliver unprecedented optimization and automation advantages.
bigRing Machine Vision
- Allows the organization to implement Vision understanding of your computer infrastructure such as OCR, ICR, OBR, OMR, Business Card or MRZ – personal information from ID documents recognition
- Process broad range of input formats like PDFs, images such as photos, screenshots or scans, as well as documents in different Office formats, that can be acquired directly from memory, uploaded from storage or scanned via the TWAIN or WIA interface
- Create searchable and editable documents that exactly match scanned or photographed originals
- Export to variety of saving formats or export options, such as TXT, RTF, DOCX, XLS(X), CSV, HTML, HTML5, ODT and PPTX that can be directly edited, E-book formats EPUB and FB2, XML, the library standard ALTO XML, XPS, vCard for business card data, as well as many types of PDF and PDF/A formats
- Export business card data directly into contact management systems
- Extract information from machine readable zones (MRZ) in ID documents and use it in customer onboarding systems for fast entry of personal information
bigRing Machine Speech
- the technology offers fast and accurate speech-to-text automatic transcription of recordings and dictated texts using deep neural networks
- bigRing Machine Speech can help you create documents up to two times faster and significanlty improve your efficiency for writing longer texts
- bigRing Machine Speech module allows your computer to „hear“ and „understand“ speech for you by integrating external technologies and applications that can automatically recognize human speech with high accuracy for all of the major languages spoken in regions of Western, Central and Eastern Europe, including Slavic languages
bigRing Machine Understanding
- bigRing Machine Understanding is the module replying on the external open-source libraries built for deep learning models for Natural Language Processing (NLP), the technique used for tasks such as language detection, key words or phrase extraction, topic recognition, sentiment analysis, and/or document categorization
- bigRing Machine Understanding uses the technique of matching the key words and/or phrases to bigRing Knowledge Elements, for unstructured data analysis and parsing of written text (Word documents, PDF texts, emails,…) and its mapping into the bigRing Causality Graph
- Components typically used for the solution creation are rich-text editor (for copy / paste of text), text-file upload and transformation into rich-text format, for custom list build-up for key words/phrases occurrence results, causality relationships as well as custom tree views
- bigRing Machine Understanding based on NLP can be used to classify and label documents (e.g. labelling as sensitive, classified or spam) that can further be used for subsequent processing, search or advanced analysis
- Furthermore, it can be used to summarize text by identifying the entities present in the document, tagging documents with keywords for subsequent search and retrieval based on content, parsing into summary that describe the important topics present in document, categorizing documents for further navigation
- bigRing Machine Understanding can use various techniques for Natural Language Processing, such as:
- Tokenization: splitting the text into words or phrases
- Lemmatization and Stemming: “normalizing” words in a way that their different forms can be mapped to the key word with the same meaning For example, “speaking” and “spoke” is mapped to “speak”
- Entity extraction: identifying subjects in the text
- Phrases and speech detection: identifying text as verb, noun, participle, verb phrase, and similar
- Sentence boundary detection: Detecting complete sentences within paragraphs of text.