A lightweight AI inference model developed with "C"
Data Analysis of Daton's Own Data Using Retention Technology

• 01

  Development of its own engine for predicting and
  analyzing financial/traffic big data (RBM)

• 02

  Promote internalization of S/W
  by expanding GS1 grade list

• 03

  Development of a Lightweight Analysis Engine
  for Mass Data Analysis (RBM)

• 04

  Development of Data Anomaly Detection
  Cause Analysis/Prediction Module

RBM System Configuration

3 Clusters

MI1Real-time prediction and judgment of events/signals that occur in real-time, such as management, customer behavior, and product sales of enterprises

MI2can support decision-making such as management decision-making and service operation

MI3System Fault Diagnosis/Prediction Neural Networking

a commercial model

Large scale implementation of basic functionality (predictive/situation judgment) quickly

Increase resource efficiency and processing speed through lightweight modules

Research and maintenance for quality control, service expansion, and evolution

Real-time lightweight anomaly detection
(Unsupervised Learning RBM : DBN)

• 

  Development of a model that allows hundreds
  of AI engines to operate simultaneously
  with node-specific anomaly detection
  and hierarchical models

• 

  A lightweight engine that can be applied to
  a large-scale data model that is difficult
  with algorithms, and after modeling,
  a large-scale model is developed through
  interworking between image analysis systems

• 

  Perform data verification on abstraction
  accuracy with Tiks (data minimum unit)
  test model