Selectivity and Shape in the Design of Forward-Forward Goodness Functions
For researchers in biologically plausible or local learning, this work provides a principled improvement over the default sum-of-squares goodness function, showing consistent gains across multiple benchmarks.
The paper systematically explores goodness-function design for the Forward-Forward algorithm, identifying that functions sensitive to shape (e.g., burstiness) outperform sum-of-squares, achieving 89.0% on Fashion-MNIST and 98.2% on MNIST, with gains up to +72pp on USPS and +52pp on SVHN.
The Forward-Forward (FF) algorithm trains networks layer-by-layer using a local "goodness function," yet sum-of-squares (SoS) has remained the only choice studied. We systematically explore the goodness-function design space and identify a unifying principle: the goodness function must be sensitive to the shape of neural activity, not its total energy. This principle is motivated by the observation that deep network activations follow heavy-tailed distributions and that discriminative information is often concentrated in peak activities. We propose two complementary families: selective functions (top-k, entmax-weighted energy) that measure only peak activity, and shape-sensitive functions (excess kurtosis / "burstiness" and higher-order moments) that reward heavy-tailed distributions via scale-invariant statistics. Combined with separate label-feature forwarding (FFCL), controlled experiments across 13 goodness functions, 5 activations, 6 datasets, and three continuous sweeps, each tracing a characteristic inverted-U, yield 89.0% on Fashion-MNIST and 98.2+-0.1% on MNIST (4x2000), a +32.6pp gain over SoS, with consistent improvements across all benchmarks (+72pp USPS, +52pp SVHN). The scale-invariant nature of burstiness makes it particularly robust to magnitude shifts across layers and datasets.