Meta-learning using privileged information for dynamics
This work addresses the problem of enhancing meta-learning in physical sciences by leveraging privileged information, though it appears incremental as it builds on existing frameworks.
The paper tackles meta-learning for dynamics by extending Neural ODE Processes to incorporate structured knowledge, such as conserved quantities, leading to improved accuracy and calibration on simulated tasks.
Neural ODE Processes approach the problem of meta-learning for dynamics using a latent variable model, which permits a flexible aggregation of contextual information. This flexibility is inherited from the Neural Process framework and allows the model to aggregate sets of context observations of arbitrary size into a fixed-length representation. In the physical sciences, we often have access to structured knowledge in addition to raw observations of a system, such as the value of a conserved quantity or a description of an understood component. Taking advantage of the aggregation flexibility, we extend the Neural ODE Process model to use additional information within the Learning Using Privileged Information setting, and we validate our extension with experiments showing improved accuracy and calibration on simulated dynamics tasks.