Zero-Shot Image Captioning Method Based on the Hamiltonian Monte Carlo

Key Points

• To explore a zero-shot learning approach in image captioning using Hamiltonian Monte Carlo techniques.
• Implemented a zero-shot learning framework for image captioning.
• Utilized Hamiltonian Monte Carlo for efficient sampling.
• Analyzed performance across various image datasets.
• Achieved competitive performance in generating captions without prior category training.
• Demonstrated improved flexibility in image captioning tasks.