Fostering AI-Fuelled Technologies to Fuse and Analyse Sensorial Knowledge for Causal Civilian Situational Awareness and Decision-Making

The vision of CYGNUS is to develop a bouquet of big data and AI services that put the human-in-the-loop to sense, harvest, harmonize, and aggregate, sensor data, train AI models to detect entities and patterns, and reduce the analysis load performed by human operators through scalable big data processing methods and Explainable AI (XAI).

Key points

Real-Time Analysis

Real-time object segmentation, detection and classification using Convolutional Neural Networks (CNNs) and YOLO Models with Confidence Score.

AI Algorithms

Behavioural pattern mining and prediction using Reinforcement Learning (RL) designed Long Short-Term Memory (LSTM), RL-CNN and Recurrent Neural Network (RNN) models and Variational Autoencoders.

XAI methods

Justified predictions on objects, patterns and events with XAI methods extended with a Probability Score (+2 Likert scale) to let human operators interpret the AI models decision paths.

M/O Optimisation Solver

Multi-Objectives (M/O) Optimisation Solver for Causal AI to derive cause-and-effect relationships and enable autonomous actions.

Pillars

Human-in-the-Loop with boosted AI explainability, trustworthiness and transparency

Flexible edge and cloud deployments for “almost-zero latency”, smart placement and high-computational capabilities

Accelerated AI automation concept thorough greener and more precise AI models with Probability Scores, increased true positives, adversarial robustness, autonomous and scalable federated learning

AI-specific KPIs validation concept via the M/O Optimisation Solver to derive cause-and-effect relationships, prescribe actions more effectively and act more autonomously

News

𝐈𝐧𝐟𝐨𝐫𝐦𝐚𝐭𝐢𝐨𝐧 𝐒𝐞𝐦𝐢𝐧𝐚𝐫 𝐨𝐧 𝐭𝐡𝐞 𝐂𝐲𝐩𝐫𝐮𝐬 𝐌𝐢𝐧𝐢𝐬𝐭𝐫𝐲 𝐨𝐟 𝐃𝐞𝐟𝐞𝐧𝐜𝐞’𝐬 𝐀𝐜𝐭𝐢𝐨𝐧𝐬 𝐢𝐧 𝐑𝐞𝐬𝐞𝐚𝐫𝐜𝐡, 𝐈𝐧𝐧𝐨𝐯𝐚𝐭𝐢𝐨𝐧, 𝐚𝐧𝐝 𝐭𝐡𝐞 𝐃𝐞𝐟𝐞𝐧𝐜𝐞 𝐈𝐧𝐝𝐮𝐬𝐭𝐫𝐲

28 May 2025

Partners

Funded by