Adaptive Multi-Scale Goodness Aggregation for Forward-Forward Learning
For researchers in biologically plausible neural networks, this work incrementally improves the Forward-Forward algorithm on small benchmarks.
The paper proposes Adaptive Multi-Scale Goodness Aggregation (AMSGA) to improve the Forward-Forward algorithm, achieving up to +1.45% on MNIST and +1.50% on Fashion-MNIST over the baseline.
We propose Adaptive Multi-Scale Goodness Aggregation (AMSGA), a novel extension of the Forward-Forward (FF) algorithm designed to improve stability, robustness, and generalization in local-learning neural networks. AMSGA addresses several limitations of the original FF framework by introducing multi-scale goodness aggregation across local, intermediate, and global representations; adaptive curriculum-guided hard negative mining; layer-dependent adaptive thresholds; and a warm-up cosine annealing learning-rate schedule for improved optimization stability. Together, these modifications strengthen the FF paradigm while preserving its biologically plausible and memory-efficient properties. Experiments on MNIST and Fashion-MNIST demonstrate consistent performance improvements over the baseline FF algorithm, achieving up to +1.45% improvement on MNIST and +1.50% improvement on Fashion-MNIST without significant computational overhead. Our results suggest that local learning methods can become substantially more competitive when goodness estimation and training dynamics are carefully designed.