Unique AI analysis technology to predict future abnormal behavior

1H3A, four exclusive technologies

  • HTD³ automatically identify abnormalities without the intervention of experts in specific fields,

  • AFS according to the intelligent analysis of indicators value correlation, "key indicators" can be found from a large number of system indicators,

  • AEC link system abnormality warning and alarm to locate abnormality location,

  • AUI with AFS, make users pay more attention to important key information.

    • Monitor multiple systems and accurately locate abnormalities

    1H3A technology constitutes three unique and effective applications, Artificial Intelligence for IT Operations (AIOPs), Artificial Intelligence of Things (AIoT), and Business operation. Provide a complete solution "Osprey", to find a needle in a haystack to find abnormal data points and locate the cause in the daily data generated by massive information systems. We continue to carry out a number of PoC projects with the telecommunications industry and have solved the detection and prediction of large-scale information systems and user behavior anomalies. In Business operation application, We focus on monitoring/forecasting/alarming/root cause analysis of End-to-End AI, and provide possible causes to prevent problems before they happen, reduce operational risks/improve operational efficiency.

    Automated processing saves cost and time

    We can automatically solve the work that required the participation of experts and scholars. Such as extracting behavior from training data, identifying whether the extracted behavior is normal or abnormal, and automatically marking the training data. Our Core technology can achieve: can detect abnormal sensing values, can detect abnormal system behavior, without required the participation of experts in specific fields, can process high-dimensional data. It only requires normal behavior during training processing. The most important is to save time and labor costs.













    The application of 1H3A in cross-fields and related research results has also been recognized internationally. It also published papers in international journals with high impact factors.

  • Applied in sleep apnea detection。 Collect brainwave EEG signals to judge the abnormality of the patient's sleep behavior and give an abnormal score to determine whether the patient suffers from sleep apnea. The application results have been published in high-impact factor IEEE international journals. RAPIDEST: A Framework for Obstructive Sleep Apnea Detection. IEEE Transactions on Neural Systems and Rehabilitation Engineering, Early Access, November 2022. (5-year Impact Factor (IF)) 5.345, SCI Rank = 6/68, Q1: REHABILITATION)











  • Applied in the abnormal detection of people flow. When an unexpected event occurs, we can find areas with abnormal population distribution in a short period of time, which will help decision makers quickly locate the area affected by the unexpected event and provide quick rescue. The application results have been published in high-impact factor IEEE international journals. ADPD: Anomaly Detection for Population Distribution in Geo-Space using Mobile Networks Data. IEEE Internet of Things Journal, 9(22): 22774-22784, November 2022. (5-year Impact Factor (IF) 10.127, SCI Rank = 5/162, Q1: COMPUTER SCIENCE, INFORMATION SYSTEMS)














  • Applied in large-scale business support systems. Extract important features from raw data, predict upcoming system abnormal events in advance, and use root cause analysis technology to find the cause of abnormalities to rule out upcoming system abnormalities early and avoid losses caused by system abnormal events. The application results have been published in high-impact factor IEEE international journals. System Error Prediction for Business Support Systems in Telecommunications Networks. IEEE Transactions on Parallel and Distributed Systems, 31(11): 2723-2733, November 2020. (5-year Impact Factor (IF) 3.215, SCI Rank = 27/110, Q1:COMPUTER SCIENCE, THEORY & METHODS)。














  • Anomaly detection and prediction are the key technologies for detecting the Internet of Things (IoT), and 1H3A is the core technical process for anomaly detection without expert intervention. The application results have been published in high-impact factor IEEE international journals. Anomaly Detection/Prediction for Internet of Things: State-of-the-Art and the Future. IEEE Network, 35(1), 212-218, January/February 2021. (5-year Impact Factor (IF) 8.876, SCI Rank = 5/91, Q1: TELECOMMUNICATIONS).





    • The patents extended by 1H3A.

    R.O.C. Patents

  • 2022.11.11, R.O.C. Patent: I783229, Title: ANOMALY FLOW DETECTION DEVICE AND ANOMALY FLOW DETECTION METHOD.

  • 2022.06.21, R.O.C. Patent: I768588, Title: PREDICTION METHOD FOR SYSTEM ERRORS.

  • 2022.01.11, R.O.C. Patent: I752638, Title: METHOD AND SYSTEM FOR DETECTION OF DRIVING ANOMALY.

  • 2021.11.21, R.O.C. Patent: I746914, Title: DETECTIVE METHOD AND SYSTEM FOR ACTIVITY-OR-BEHAVIOR MODEL CONSTRUCTION AND AUTOMATIC DETECTION OF THE ABNORMAL ACTIVITIES OR BEHAVIORS OF A SUBJECT SYSTEM WITHOUT REQUIRING PRIOR DOMAIN KNOWLEDGE.

  • 2016.10.21, R.O.C. Patent: I554873, Title: PERSONAL MOBILE COMMUNICATION DEVICE AND METHOD THEREOF EXECUTING POWER SAVING SETTING BY USER BEHAVIOR.


    • United States Patents

  • 2022.11.08, App. No.: US20210367885 A1, Title: ANOMALY FLOW DETECTION DEVICE AND ANOMALY FLOW DETECTION METHOD.

  • 2022.04.12, United States Patent: US 11,301,759 B2, Title: DETECTIVE METHOD AND SYSTEM FOR ACTIVITY-OR-BEHAVIOR MODEL CONSTRUCTION AND AUTOMATIC DETECTION OF THE ABNORMAL ACTIVITIES OR BEHAVIORS OF A SUBJECT SYSTEM WITHOUT REQUIRING PRIOR DOMAIN KNOWLEDGE


    • R.O.C. Patents (Patent Pending)

  • 2022.05.01, App. No.: 202217639, Title: PREDICTION METHOD BASED ON UNSTRUCTURED DATA. Patent Pending. May 1, 2022.


    • United States Patents (Patent Pending)

  • 2022.06.16, App. No.: US20220188669 A1, Title: PREDICTION METHOD FOR SYSTEM ERRORS.

  • 2022.04.28, App. No.: US20220129490 A1, Title: PREDICTION METHOD BASED ON UNSTRUCTURED DATA.

  • 2022.03.17, App. No.: US20220080988 A1, Title: METHOD AND SYSTEM FOR DETECTING DRIVING ANOMALIES.



    • Cooperating with us, possible major system vulnerabilities can be prevented in advance.

    View application cases