Research

2025

Federated learning

Federated learning is a distributed approach to train machine learning models. Each node in a distributed network trains a global model using its own local data, while a central server improves the global model by aggregating node updates.

A Study on the Life Cycle Management of Federated Learning

Establish FeDOps platform of Federated Learning + Ops for federated learning life cycle management

Privacy protection artificial intelligence technology

Provides AI learning solutions in data sensitive domains with AI learning method without data sharing

Data diversity

Unrestrict data usage to secure data from various organizations and customers to perform advanced AI learning

Building an Ecosystem for Decentralized Artificial Intelligence

Establishment of a cooperative learning ecosystem through website, community, etc. and technology exchange with other overseas organizations to realize a cooperative learning ecosystem

Spain Telefonica federated learning-based

MyHealthMart Technology Presents

Best institution for joint learning posted on the Flower website

AI Agent

AI agents are autonomous systems that use AI to pursue user-defined goals through reasoning, planning, memory, and decision-making. Powered by generative AI and multimodal models, they can process text, speech, video, audio, and code while learning and adapting.

AI Agent System Based on

Federated Learning

A federated learning-based AI system that leverages user-specific health data to integrate a local AI model from the client into a global model via training FedOps

Utilize global models

and secure data

Integrated global models are used for a variety of AI functions such as health forecasting and data valuation Securely call and process user data in conjunction with MyData systems

Custom Response

Generation Mechanism

User queries are communicated to the Medical LLM via the prompt module Create reliable, customized answers with personalized memory, RAG-based discovery, tool calls, and action plans

Perform tool-based complexity

AI Agent proactively utilizes a variety of tools and functional modules Perform complex artificial intelligence functions such as question-and-answer, data analysis, tool execution, and assetization

GCCL Medical LLM

Healthcare domain-specific data and federated learning-based fine tuning strategies work effectively. High accuracy and consistent performance on healthcare QA benchmarks, improved understanding and reasoning skills for complex healthcare contexts

Rank 1 in the NLP leaderboard

Rank 3 on the Medical leaderboard

TAVR Medical Twin

an existing Post-TAVR Navigator platform is restructured into a federated learning solution, FL Post-TAVR, based on the proposed federated learning approach through lifecycle management by using FedOps.

TAVR AI MODEL / XAI