An energy-based recurrent state-space model with a continuous attractor memory predicts sensory observations after actions using local Hebbian learning, matching several ML baselines.
For the experiments depicted in our figures, each one takes 5-20 minutes, with the best performance on the DeepLab and Google Street datasets requiring about 10 hours
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Predictive Learning in Energy-based Models with Attractor Structures
An energy-based recurrent state-space model with a continuous attractor memory predicts sensory observations after actions using local Hebbian learning, matching several ML baselines.