BeamVQ: Aligning Space-Time Forecasting Model via Self-training on Physics-aware Metrics.
Hao WuXingjian ShiZiyue HuangPenghao ZhaoWei XiongJinbao XueYangyu TaoXiaomeng HuangWeiyan WangPublished in: CoRR (2024)
Keyphrases
- space time
- forecasting model
- short term
- spatial and temporal
- video sequences
- prediction model
- spatio temporal
- bp neural network
- support vector regression
- stock index futures
- long term
- dynamic scenes
- rainfall runoff
- training set
- artificial intelligence
- image processing
- back propagation
- active learning
- feature vectors
- ls svm
- machine learning