Scene 1 (0s)
1. Learning Profitable NFT Image Diffusions via Multiple Visual-Policy Guided Reinforcement Learning.
Scene 2 (10s)
Background. 2. Far Beyond an Image. Society Community Culture.
Scene 3 (22s)
Diffusion-MVP. Motivation. Challenge 1: Lack of large-scale high-quality NFT dataset Challenge 2: Mining market value-related visual features Challenge 3: Incorporating market value into generation.
Scene 4 (34s)
Resolution Filter. Property Filter. Visual Content Filter.
Scene 5 (51s)
Market Price-based Reward Ours Rarity-based Reward Noise Affected by Infrequent and fluctuate transactions Market noise Less affected by market noise Correlated with image visual features Prediction accuracy 39.60% 85.62%.
Scene 6 (1m 8s)
Lion, Golden Mane, Standard Eyes, Pearl Nose, Green Background,.
Scene 7 (1m 35s)
Results: Overall Comparison Results. Evaluation Criteria.
Scene 8 (1m 48s)
Cyan Background. Driller Forehead. More results compared to the SOTA.
Scene 9 (2m 15s)
Thank You!. 9.