A CNN-LSTM model trained on simulated thermal and power data reaches 94% accuracy on a private synthetic dataset, but the practical claim is not supported by real-world validation or a rule-based baseline.
How transferable are features in deep neural networks?
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Predictive Failure Detection in Network Hardware Using Thermal Imaging and Deep Learning with Sensor Fusion
A CNN-LSTM model trained on simulated thermal and power data reaches 94% accuracy on a private synthetic dataset, but the practical claim is not supported by real-world validation or a rule-based baseline.