Hybrid LLM and SHAP framework explains cyberbullying predictions
A study of 5,000 Turkish tweets found that SHAP offered detailed feature attributions while local LLM explanations were clearer, with ...
Tag: Explainable AI
Research report: A population representative XAI framework for transparent dysglycemia screening in Indonesia
An Android screening app trained on Indonesian survey data used XGBoost and SHAP to explain dysglycemia risk, achieving a 0.72 ...
Diabetes AI combines knowledge graphs with language models for traceable answers
A diabetes question-answering prototype combines a structured medical knowledge graph, guideline retrieval and a large language model while exposing the ...
Explainable AI system predicts IT outsourcing success before contracts are signed in Iraq
An explainable machine learning system trained on 336 Iraqi IT outsourcing projects predicts success from pre-contractual data, finding that fixed-price ...
Frequency-aware transformer spots AI-generated fake voices and shows its reasoning
A new Frequency-Aware Transformer achieved 94.6% accuracy and a 6.3% equal error rate on the ASVspoof 2019 Logical Access benchmark ...
Explainable AI yields a five-parameter formula for nonlinear concrete creep
Researchers combined an optimized gradient-boosting model, SHAP explainability, and symbolic regression to derive an explicit, accurate equation for concrete creep ...
Trust-aware graph network shows promise and limits in skin lesion classification
A new graph neural network framework makes skin lesion classification more transparent, but candid reporting reveals serious weaknesses in melanoma ...
Explainable Volterra Networks Bring Transparent Nonlinear Dynamics to Time Series Forecasting
A new neural architecture converts the classical Volterra series into learnable coefficient tensors, recovering known chaotic dynamics, matching transformer physics, ...
Explainable AI framework blends calibrated models to predict heart disease risk
Researchers report a feature-weighted, calibrated explainable AI framework that reached 91.07% accuracy on a cleaned heart disease dataset while exposing ...
How Image Quality and Scaling Shape the Accuracy of Future AI Systems
A peer-reviewed experimental review in Artificial Intelligence Review examines how image enhancement, scaling, and evaluation metrics jointly determine the reliability ...
Tiny AI Model Reads Diseased Crop Leaves With Big-Model Accuracy and Explains Itself
A 79,717-parameter attention-fused neural network called LeafXNet matches or beats far larger deep learning models at detecting watermelon, grape, and ...
AI scoring tool shows promise in flagging cancerous skin lesions
A retrospective Mayo Clinic study found that the DEXI dermoscopy AI tool, using a threshold score of 3.0, correctly identified ...
Uncertainty-aware Bayesian neural network brings confidence scores to breast cancer AI
A variational Bayesian neural network trained on the Wisconsin Diagnostic Breast Cancer dataset matches conventional classifiers while flagging which predictions ...
AI and a Chat Interface Team Up to Predict Wind Turbine Failures in Real Time
Researchers in Türkiye have combined IoT sensors, gradient-boosted machine learning, and a safely grounded large language model into a prototype ...
AI Explanations Collapse When Supply Networks Fracture, Spectral Study Warns
A new spectrally grounded framework proves that collapsing network connectivity, measured by the Fiedler eigenvalue, is a sufficient condition for ...
Machine learning reveals buckling prediction limits in cold-formed steel columns
A study of 148 built-up cold-formed steel column tests found that machine-learning corrections performed well within the database but transferred ...
Machine learning predicts self-reported anger improvement in teens
A comparative machine-learning study found that meditation, behavioral and lifestyle survey data can predict self-reported anger-management improvement in adolescents, but ...
Federated ensemble framework predicts chronic kidney disease without sharing records
A simulated three-client framework combined stacked machine-learning models, differential privacy, federated SHAP explanations and low-bandwidth aggregation to classify chronic kidney ...
Machine learning predicts severe lung disease risk in extremely preterm infants
A single-center retrospective study found that a calibrated random forest using respiratory monitoring, chest radiographs and perinatal data predicted severe ...
Prospective Vietnamese cohort tests explainable machine learning for acute heart failure mortality
A prospective Vietnamese cohort found that explainable machine learning could rank in-hospital mortality risk in acute heart failure, but it ...
POPULAR NEWS
-
Two C3 Gene Variants Reveal Functional Independence Between Extracellular and Intracellular Forms of Complement C3
-
Nearly Half of Stroke Patients in Africa and Asia Show Low HDL Cholesterol
-
German heat-health plans rarely evaluated due to political and structural barriers, study finds
-
Carb Quality Beats Glycemic Index in Sperm DNA Study