Global ETD Search
Search theses and dissertations gathered from participating repositories worldwide. Every result links back to the library that holds it. No account is needed.
Results
Showing 1 to 4 of 4 for “"Stage-wise Training"”.
-
Conformal Prediction for Time-Series and Flow-Based Generative Models
… based on the Jordan-Kinderlehrer-Otto scheme for stage-wise training. Finally, Chapter 4 enhances the scalability of ODE-based models by introducing a local flow matching approach, improving training efficiency and distillation performance.
-
Solving Visual Species Recognition with Open-World Resources
… and ecology and ecosystem management. However, training such a machine learning model typically requires a large amount of species-level annotated data, which is costly in terms of time, budget, and domain expertise. In addition, observations of species naturally follow a long-tailed …
-
Solving Visual Species Recognition with Open-World Resources
… and ecology and ecosystem management. However, training such a machine learning model typically requires a large amount of species-level annotated data, which is costly in terms of time, budget, and domain expertise. In addition, observations of species naturally follow a long-tailed …
-
Regularizing Deep Models for Visual Recognition
… representational power of deep networks, training these models is computationally expensive. In addition, considering the lack of enough labeled training data in many applications, over-fitting is a serious threat for deep models with large number of free parameters. Also, there are innate …