FAQ
We’ve compiled a list of frequently asked questions along with their answers.
What is Ephapsys?
Ephapsys is a platform that uses ephaptic coupling technology to empower and govern AI agents.
What is ephaptic coupling?
Ephaptic coupling is a biological phenomenon where neurons communicate via electric fields rather than synapses. Ephapsys adapts this concept into AI models, enabling new forms of learning and interaction beyond traditional neural networks.
What is the difference between synaptic and ephaptic learning?
Synaptic learning propagates signals through weighted connections (rows in a matrix). Ephaptic learning influences neural activations through neighboring cross-talk effects or local field interactions (columns), offering new dimensions of computation.
How different is ephaptic coupling from LoRA?
Ephaptic coupling and LoRA differ fundamentally in how they modify neural computation and learning. LoRA (Low-Rank Adaptation) is a parameter-efficient fine-tuning technique that inserts small, trainable low-rank matrices (A and B) into existing weight layers, effectively learning linear adjustments to the model’s synaptic weights while keeping the base model frozen. In contrast, ephaptic coupling introduces non-synaptic, field-based interactions between neurons, inspired by how real biological neurons influence each other through local electric fields rather than direct synaptic connections. In an AI context, ephaptic coupling alters how activations co-vary across neurons (columns) instead of just along synaptic weight paths (rows), enabling collective, context-sensitive modulation of activity without adding explicit trainable parameters like LoRA does. While LoRA is linear, modular, and purely synaptic, ephaptic coupling adds a nonlinear, global influence layer called Ephaptic Coupling Matrix (ECM) that can dynamically reshape learning and inference behavior based on emergent field effects, more closely resembling how biological brains self-organize and adapt under shared electric environments.
How does Secure Inference Enforcement work?
Secure Inference Enforcement ensures that every inference made by an AI agent is verified and bound to cryptographic identity and runtime policies. This prevents tampering, unauthorized use, or unverified outputs.
How can I register my organization?
You can register your organization through our sign-up page. Once registered, you’ll get access to a dashboard, model modulation and agent management tools, and API keys for using the SDK.
Do you provide customer support?
Yes — our support engineers are available 24/7 for critical issues. You can reach us via phone, email, or Discord.
How secure is the Ephapsys platform?
Every agent is cryptographically anchored with decentralized identity and PKI. Policies such as revocation, kill-switches, and audit logs ensure full lifecycle governance.
Can agents run offline?
Yes. Agents can operate offline with locally cached policies and secure elements. Emergency revocation is still possible via pre-provisioned kill-switch mechanisms.
How does billing work?
Billing is agent-based, with flexible tiers. Revoked agents are not charged, and enterprises can scale usage with predictable costs.
What industries can benefit from Ephapsys?
From finance and healthcare to telecom and robotics, any industry that requires trusted AI agents can leverage Ephapsys.
