Popular repositories Loading
-
-
explainable-physiological-stress-ai
explainable-physiological-stress-ai PublicEnd-to-end CPU-only physiological ML pipeline using WESAD wrist sensors (EDA, BVP, TEMP, ACC). Features feature extraction via NeuroKit2, Leave-One-Subject-Out (LOSO) evaluation, SHAP explainabilit…
Jupyter Notebook
-
cps-skab-adversarial-security
cps-skab-adversarial-security PublicAdversarial evasion attacks and ensemble defenses for industrial anomaly detection (LSTM Autoencoder & Isolation Forest) on the SKAB water testbed.
HTML
-
honest-calibrated-small-llms
honest-calibrated-small-llms PublicTruthfulQA MC1 evaluation harness for small LLMs: conditional log-likelihood scoring, 10-bin ECE, Brier score, prompt-flip heatmaps, and post-hoc temperature scaling.
Jupyter Notebook
-
cifar10-robustness-report-card
cifar10-robustness-report-card PublicCPU-only PyTorch benchmark evaluating CIFAR-10 CNN robustness under FGSM & PGD attacks omparing standard vs. adversarially trained models with ASR, confidence, and t-SNE analysis.
Jupyter Notebook
-
distilbert-adversarial-xai
distilbert-adversarial-xai PublicNLP robustness & explainability study: breaking DistilBERT on SST-2 with TextFooler, DeepWordBug, and PWWS, explaining flips via Captum Integrated Gradients, and hardening with 1-epoch adversarial …
Jupyter Notebook
If the problem persists, check the GitHub status page or contact support.