Search papers, labs, and topics across Lattice.
This paper investigates representational stability in a minimal continual learning agent with a persistent state vector, focusing on how internal representations evolve over time. The agent incrementally updates its state vector based on incoming textual data, and representational change is quantified using cosine similarity between successive normalized state vectors. Results show a transition from a plastic to a stable regime under consistent input, and a bounded similarity decrease followed by recovery after semantic perturbation, indicating inherent stability-plasticity tradeoffs.
A surprisingly simple agent reveals that continual learning can exhibit meaningful stability-plasticity tradeoffs without complex regularization or replay mechanisms.
Continual learning systems are increasingly deployed in environments where retraining or reset is infeasible, yet many approaches emphasize task performance rather than the evolution of internal representations over time. In this work, we study a minimal continual learning agent designed to isolate representational dynamics from architectural complexity and optimization objectives. The agent maintains a persistent state vector across executions and incrementally updates it as new textual data is introduced. We quantify representational change using cosine similarity between successive normalized state vectors and define a stability metric over time intervals. Longitudinal experiments across eight executions reveal a transition from an initial plastic regime to a stable representational regime under consistent input. A deliberately introduced semantic perturbation produces a bounded decrease in similarity, followed by recovery and restabilization under subsequent coherent input. These results demonstrate that meaningful stability plasticity tradeoffs can emerge in a minimal, stateful learning system without explicit regularization, replay, or architectural complexity. The work establishes a transparent empirical baseline for studying representational accumulation and adaptation in continual learning systems.
