Quantum Urban Heat LSTM Fix
Budget: ₹1,500 – ₹12,500 INR
I have a Quantum-IoT pipeline that predicts urban-heat islands with a hybrid LSTM / Quantum-LSTM network. The raw sensor feed is undermining performance: inconsistent readings are slipping in right at the data-collection stage, and the model accuracy drops sharply after training.
I need someone who can:
• Inspect the current Python/TensorFlow + PennyLane codebase and pinpoint where inconsistent values, duplicate timestamps, or unit mismatches enter the dataset.
• Build a lightweight data-validation layer that flags or corrects those inconsistencies before the preprocessing step.
• Retrain the classical LSTM and the Q-LSTM branches after the data fix and show the improvement through comparative metrics (MAE, RMSE, R²).
• Deliver clean, well-commented scripts and a short README outlining what was changed, why, and the results.
Familiarity with IoT time-series cleaning, TensorFlow/Keras, and at least one quantum framework (PennyLane or Qiskit) is essential. If you can get the pipeline back on track and demonstrate better predictive accuracy, that’s exactly what I’m after.
I need someone who can:
• Inspect the current Python/TensorFlow + PennyLane codebase and pinpoint where inconsistent values, duplicate timestamps, or unit mismatches enter the dataset.
• Build a lightweight data-validation layer that flags or corrects those inconsistencies before the preprocessing step.
• Retrain the classical LSTM and the Q-LSTM branches after the data fix and show the improvement through comparative metrics (MAE, RMSE, R²).
• Deliver clean, well-commented scripts and a short README outlining what was changed, why, and the results.
Familiarity with IoT time-series cleaning, TensorFlow/Keras, and at least one quantum framework (PennyLane or Qiskit) is essential. If you can get the pipeline back on track and demonstrate better predictive accuracy, that’s exactly what I’m after.