SAGE search – Natural language understanding
SAGE search – Natural language understanding
SAGE computers (operating NLU and semantic search) learn to read and understand (humanly) documents, data, content, images, natural language, scanned documents (OCR), handwriting (ICR, also in digital pens), PDF, HTML, meaning of words by virtue of organization In the sentence and paragraph, FREE TEXT, requirements of seekers (conducting a short dialogue when necessary, refining the question and providing an accurate, intelligent and unambiguous answer), emails, databases, messages and rejects (from RPA).
SAGE Computers learn to search for information in the organization and on the Internet according to business rules and grammatical rules, Separate the wheat from the chaff , capture content in documents, photos, emails, FREE TEXT and forms (if an undeciphered field remains, an upgraded mouse (Sage Robot) is placed on the word, the next time a similar document arrives it is automatically decoded) and run automatically (without any typing) any software (mouse actions, keyboard, digital pen, touch screen, Icons, buttons, text boxes, scroll lists) and any process (automatically or through natural language interaction with the computer – AI + Robotic Process Automation), operate Sage Robots ( Virtual Employees) for Procurement, Customer Service, Sales, Marketing, HR, Inventory, Maintenance, Automated Warehouse, Import, Export, Treasury, Hedging, Investment, Knowledge Management and Documents,Operation, production, quality control, software development and testing).
SAGE decipher and understand sound and images (from all types of cameras, including IR, terahertz, X-ray and scanners); search for images by their parts and content related to images; automatically translate texts and simplify them linguistically and conceptually (ATS – Individuality of the language, selection of appropriate vocabulary, wording of short sentences in simple, clear, unambiguous and easy language Understanding for computers and each person) ; Automatically understand and translate natural language definitions of rules (business, mathematical, logical and linguistic – vocabulary, word relations, homonymy, morphology) into software language, use them for the benefit of all applications in the organization and network, allow anyone to develop software without code (including queries, interfaces , RPA and AI) and replace human intuition with artificial intelligence and mathematical and statistical algorithms. SAGE includes a huge library of ready-made components and interfaces that automatically run search engines ERP, CRM, OCR, ASR, EXCEL, ML, BI, DM, GOOGLE AI, WINDOWS AI, WhatsAPP, SP, OneNote, Power Automate, Noise removal, Image recognition, Robots for transport and collection, Image repair & enhancement, dictionaries.
SAGE NLU enables the automatic construction of unstructured data, semantic and accurate search of structured and unstructured data and optimization of databases (in the ability to search and query and retrieve time); Automatically cleans noise, optimizes and upgrades any application (reads and understands the fields on screens and DB and allows the user to add screens, queries, RPA and AI – in natural language and without code), optimizes all NLU and OCR software and creates the most accurate Natural language understanding and OCR in the world , Makes information and knowledge accessible to computers and people (from millions of applications and billions of documents on the web), optimizes the knowledge and learning of computers (Artificial Intelligence) and humans and creates the most intelligent robots and computers in the world.
SAGE mimics the way the human mind learns. Uses artificial neural networks, which include a large number of “neurons” arranged in layers. Each neuron can communicate with a number of other neurons in the system. Each neuron is able to perform simple computational operations and in turn pass on the information it has learned to the other neurons. In this way, as the learning progresses, SAGE turns raw information into valuable information and learns to accurately identify images and content. The first neurons knew how to recognize straight lines, the next in line knew how to recognize simple corners or patterns, the ones after them already knew how to recognize the contours of the letters and finally the last neurons knew how to understand the text or image. In order to derive optimal value from learning, a maximum amount of examples and information must be revealed to SAGE. The training is performed automatically, without code, repeatedly. At each stage SAGE compares its results to existing information and checks whether it was indeed right to decipher. If not, it slightly changes the parameters in the relevant neurons until optimal accuracy is obtained.