Accelerating Machine Learning via the Weber-Fechner Law
This work addresses the challenge of efficient learning in machine learning, particularly for tasks involving human-like concepts, though it appears incremental as it adapts a known psychological law to existing neural network methods.
The paper tackled the problem of accelerating machine learning for human concepts by applying the Weber-Fechner Law, which models human perception as logarithmic scaling, to neural networks, resulting in improved performance and accuracy on the MNIST dataset with few training iterations and limited resources.
The Weber-Fechner Law observes that human perception scales as the logarithm of the stimulus. We argue that learning algorithms for human concepts could benefit from the Weber-Fechner Law. Specifically, we impose Weber-Fechner on simple neural networks, with or without convolution, via the logarithmic power series of their sorted output. Our experiments show surprising performance and accuracy on the MNIST data set within a few training iterations and limited computational resources, suggesting that Weber-Fechner can accelerate machine learning of human concepts.