Learning from Snapshots of Discrete and Continuous Data Streams
This work addresses fundamental learnability challenges in online learning scenarios like smart camera traps, providing theoretical insights but is incremental in extending existing learning theory to snapshot-based streams.
The paper tackles the problem of learning from adaptively sampled snapshots of discrete and continuous data streams, showing that uniform sampling can learn concept classes with finite Littlestone dimension in an update-and-deploy setting, while non-trivial classes are unlearnable in a blind-prediction setting, with adaptive algorithms necessary for pattern classes.
Imagine a smart camera trap selectively clicking pictures to understand animal movement patterns within a particular habitat. These "snapshots", or pieces of data captured from a data stream at adaptively chosen times, provide a glimpse of different animal movements unfolding through time. Learning a continuous-time process through snapshots, such as smart camera traps, is a central theme governing a wide array of online learning situations. In this paper, we adopt a learning-theoretic perspective in understanding the fundamental nature of learning different classes of functions from both discrete data streams and continuous data streams. In our first framework, the \textit{update-and-deploy} setting, a learning algorithm discretely queries from a process to update a predictor designed to make predictions given as input the data stream. We construct a uniform sampling algorithm that can learn with bounded error any concept class with finite Littlestone dimension. Our second framework, known as the \emph{blind-prediction} setting, consists of a learning algorithm generating predictions independently of observing the process, only engaging with the process when it chooses to make queries. Interestingly, we show a stark contrast in learnability where non-trivial concept classes are unlearnable. However, we show that adaptive learning algorithms are necessary to learn sets of time-dependent and data-dependent functions, called pattern classes, in either framework. Finally, we develop a theory of pattern classes under discrete data streams for the blind-prediction setting.