Bidirectional predictive coding
This work addresses the problem of degraded performance in bidirectional processing tasks for computational neuroscience and AI, offering a more biologically plausible model, though it is incremental in extending existing predictive coding frameworks.
The authors tackled the limitation of unidirectional predictive coding models by proposing bidirectional predictive coding (bPC), which integrates both generative and discriminative inference, and demonstrated that it matches or outperforms unidirectional models in specialized tasks and excels in multimodal learning and inference with missing information.
Predictive coding (PC) is an influential computational model of visual learning and inference in the brain. Classical PC was proposed as a top-down generative model, where the brain actively predicts upcoming visual inputs, and inference minimises the prediction errors. Recent studies have also shown that PC can be formulated as a discriminative model, where sensory inputs predict neural activities in a feedforward manner. However, experimental evidence suggests that the brain employs both generative and discriminative inference, while unidirectional PC models show degraded performance in tasks requiring bidirectional processing. In this work, we propose bidirectional PC (bPC), a PC model that incorporates both generative and discriminative inference while maintaining a biologically plausible circuit implementation. We show that bPC matches or outperforms unidirectional models in their specialised generative or discriminative tasks, by developing an energy landscape that simultaneously suits both tasks. We also demonstrate bPC's superior performance in two biologically relevant tasks including multimodal learning and inference with missing information, suggesting that bPC resembles biological visual inference more closely.