Research that learns like the brain
Our mission: translate neural computation into practical AI systems that are efficient, robust, and aligned.
Research areas
Cortical priors
Predictive Coding Models
Hierarchical inference and error-driven learning architectures.
Silicon brains
Neuromorphic Learning
Local learning rules on spiking substrates for efficiency.
Human-like
Cognitive Constraints
Memory, attention, and rational bounds for safer AI behavior.
Interfaces
Brain–Computer Interfaces
Decoding and stimulation with adaptive, interpretable models.
Lifelong
Continual Learning
Avoiding catastrophic forgetting via synaptic consolidation.
Insight
Mechanistic Interpretability
Linking circuits to capabilities through causal analysis.
Publications
-
2025Cognitive Constraints
Bounded Rationality for Safer LLMs
Integrates cognitive cost into decoding to reduce hallucinations.
Read on arXiv -
2024Neuromorphic Learning
Local Plasticity Rules on Spiking Accelerators
Demonstrates energy-accuracy trade-offs for on-device continual learning.
Read preprint -
2023Predictive Coding
Hierarchical Error Feedback in Transformers
Introduces cortical feedback pathways to improve sample efficiency.
Read paper
Team highlights
Dr. Ada Jensen
Chief Scientist — Predictive Coding
Ken Alvarez
Lead — Neuromorphic Systems
Dr. Mira Shah
Research — Cognitive Modeling
Leo Park
BCI Engineer
Collaborate with WaxBrain
We partner with labs and industry teams to accelerate brain-inspired AI.