AI Suite Summary

AI SUITE

AI recognition and technology paradigm change

AI (Artificial Intelligence) technologies have been studied and developed for a long time. While research on AI technology had started since 1950s, the early goal was to explicitly establish knowledge and define logic (rules) in the machine just as human beings understand and think. Research was focused on enabling machines to make more complex inferences and explorations, and by the 1980s, it was further researched and developed to become a form of expert system, which was further studied and developed.

By the 2000s, research turned to the field of machine learning, where too much time and effort had to be spent on building the knowledge and logic required for the system. Thus, researchers tried to overcome the limitations by letting the machines learn by themselves. Since then, there have been significant developments in machine learning and hardware. As a result, research on deep learning technology has grown rapidly since 2012. Machines have learned by themselves without human intervention, bringing about excellent results in various fields.

This led to the change in the approach in the field of artificial intelligence, in which development through deep learning technology was deemed important.

At present, artificial intelligence technology has been applied to real life, leading to the expansion of services, and artificial intelligence services are already being established in various industries such as financial, medical, legal, public and private sectors. To achieve a satisfactory services level, machines need to be more intelligent through interaction with the service environments and services in which AI is fused with other technologies are needed. So what is needed is not only technical growth, but development from a service perspective

Saltlux AI Suite offers the following 10 engines.

Insight on product innovation based on the changes in the market

As we have seen in the development of artificial intelligence, technology is being developed rapidly based on market demand. Thus, in order to provide viable services, product development must be in line with technological development.

Since its establishment, Saltlux has made efforts to secure its matchless original technology of artificial intelligence including natural language processing, semantics, and inferences. Through semantic and knowledge graph technologies, an intelligent system is built by accumulating explicit knowledge and logic. In addition, we have achieved a technological accomplishment in which the system grows on its own by learning statistical patterns through big data-based machine learning and deep learning. In other words, Saltlux’s AI technology is turned into a commercial Ensemble AI in which the system is made intelligent through an interaction in which knowledge-based reasoning and data-based are integrated. In addition, to make AI technology available to real-life services, the system was trained through such processes as cognition, understanding, knowledge, reasoning, and prediction with data such as language, voice, and vision. We are integrating the required features in each process, and making it better to ensure superior performance. As such, Saltlux’s AI Suite products are evolving into an innovative AI platform that provides integrated intelligent software solutions and services through the convergence of AI technology.

Introduction

AI Suite is the culmination of Saltlux’s investment for the last 20 years. It is divided into four distinct areas, i.e. language intelligence, speech intelligence, learning / reasoning intelligence, and visual intelligence, and a unique technology has been developed for each field. Korea’s first commercially available AI platform, ADAM, has the largest knowledge graph in Asia and shows 94% accuracy in daily Q&A using knowledge learning and reasoning, making it able to answer common sense questions. Each engine can combine the components needed for each applicable domain or use the 60 APIs provided in RESTful form by ADAMS.ai, which allows the innovative AI services to be implemented in such areas as virtual assistants, AI agents, and intelligent robots.

ai01
ai01

< AI Suite block diagram >

As shown in the following figure, AI Suite includes different engines in different areas, so that various AI-based technologies can be utilized to perform necessary functions and is configured to operate as an intelligent knowledge processing platform. The platform can be used to analyze and understand data such as text, voice, and video, etc. and a knowledge graph is established to be utilized in intelligent knowledge processing such as dialogue, question and answer, reading comprehension , and translation. In addition, each engine is designed to be applied independently and selectively according to the functions needed in services, so it can be effectively used in implementing various kinds of AI services.

ai02

< AI Suite block diagram >

AI Suite can be used for AI-based real-time virtual counseling and counseling support services. It can also be widely used in areas requiring immediate knowledge-based customer service, such as finance, telecommunications / broadcasting, legal/patents, healthcare, shopping/travel, public and private, etc.

ai03

< Scope of main application of AI Suite >

Main Functions

AI Suite categorizes individual engines for major technologies in each AI field as shown in the table below, and provides the best performance features and services for each engine. In addition, it provides management functions such as dictionary and data establishment, and model training to provide optimal results through customization according to each application field.

ai34

Main Features

Untitled web 1
Untitled web 2
Untitled web 3

Main competitiveness

Unlike other products that provide only natural language processing and dialogue model-oriented functions, AI Suite continuously manufactures products for which the following differentiated and competitive functional elements are under constant development.

Untitled web 4
Untitled web 5