Multi-Modal Machine Learning for Airbnb Listings (Text + Images)

Description:

Built a multi-modal ML model combining text (GRU/LSTM) and images (Conv2D) to predict price ranges and listing types of Airbnb properties.

Highlights:

Preprocessed text & image data into meaningful features.

Used multi-objective learning for simultaneous predictions.

Applied grid/random/Bayesian search for hyperparameter tuning.

Explored transfer learning to boost performance.

Tech: Python, PyTorch, TensorFlow, GRU, LSTM, Conv2D

Impact: Improved property pricing and categorization using combined text + image insights.