CAMELEON: Cognitively Adaptive Memory with Emotion-weighted Layered Expert Orchestration Network
Year: 2026
Type: preprint
Abstract: We present CAMELEON — a conceptual architecture for AI agents that grow more personalised through lived interaction rather than retraining. The architecture combines emotion-weighted context compression, interpretable domain-expert routing via keyword saliency scoring, speculative context pre-activation, and time-decay relevance scoring...
Author: Arumugam, Rajkumar
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7165046933
Cites:
Cited by:
Related to: