About
Research & Development
Unlocking Time Series Model Decisions with PHAR: A Framework for Interpretable Rules
TLDR: PHAR is a new framework that converts numerical explanations from AI models (like LIME, SHAP) for time series data into clear, human-readable "IF-THEN"...
Research & Development
Advancing Human Activity Recognition: A Multimodal and Privacy-Preserving Approach
TLDR: A new AI framework, FedTime-MAGNET, significantly improves Human Activity Recognition (HAR) by combining diverse sensor data (depth cameras, pressure mats, accelerometers) using a...
Research & Development
HateClipSeg: Advancing the Detection of Hate Speech in Videos
TLDR: HateClipSeg is a new large-scale, multimodal dataset with over 11,714 video segments annotated for fine-grained hate speech detection, including categories like Hateful, Insulting,...
Research & Development
Navigating Complex Data: How AI Agents Are Transforming Knowledge Graph Question Answering for a Sustainable Future
TLDR: AGENTIC T2S is a new multi-agent AI framework designed to answer complex questions by querying across diverse and distributed knowledge graphs, particularly in...
Research & Development
Accelerating Microstructure Prediction with Deep Learning and Latent Space Interpolation
TLDR: A new deep learning framework uses a Conditional Variational Autoencoder (CVAE) combined with Cubic Spline and Spherical Linear Interpolation (SLERP) to rapidly predict...
Research & Development
Neural Predictive Control: Enhancing Time-Series Analysis with Coordinated Discrete and Continuous Models
TLDR: Neural Predictive Control (NPC) is a new method for time-series analysis that combines discrete-time and continuous-time models using Model Predictive Control (MPC). It...