Polyra Swarms: A Shape-Based Approach to Machine Learning
This work addresses the need for low-bias, transparent machine learning methods, potentially benefiting researchers and practitioners in fields like anomaly detection, though it appears incremental as it builds on shape-based approximations rather than introducing a completely new paradigm.
The authors tackled the problem of general-purpose machine learning by proposing Polyra Swarms, a shape-based approach that approximates shapes instead of functions, showing it can be preferable to neural networks for tasks like anomaly detection and introducing an automated abstraction mechanism to simplify complexity and enhance generalization and transparency.
We propose Polyra Swarms, a novel machine-learning approach that approximates shapes instead of functions. Our method enables general-purpose learning with very low bias. In particular, we show that depending on the task, Polyra Swarms can be preferable compared to neural networks, especially for tasks like anomaly detection. We further introduce an automated abstraction mechanism that simplifies the complexity of a Polyra Swarm significantly, enhancing both their generalization and transparency. Since Polyra Swarms operate on fundamentally different principles than neural networks, they open up new research directions with distinct strengths and limitations.