StatePlane: A Cognitive State Plane for Long-Horizon AI Systems Under Bounded Context
This addresses a fundamental bottleneck for AI systems like LLMs and SLMs in multi-session tasks, offering a novel solution rather than an incremental improvement.
The paper tackles the problem of limited context windows in language models hindering long-horizon reasoning by introducing StatePlane, a cognitive state plane that manages episodic, semantic, and procedural state without expanding context or retraining, achieving long-horizon intelligence as demonstrated in six domain-specific benchmarks.
Large language models (LLMs) and small language models (SLMs) operate under strict context window and key-value (KV) cache constraints, fundamentally limiting their ability to reason coherently over long interaction horizons. Existing approaches -- extended context windows, retrieval-augmented generation, summarization, or static documentation -- treat memory as static storage and fail to preserve decision-relevant state under long-running, multi-session tasks. We introduce StatePlane, a model-agnostic cognitive state plane that governs the formation, evolution, retrieval, and decay of episodic, semantic, and procedural state for AI systems operating under bounded context. Grounded in cognitive psychology and systems design, StatePlane formalizes episodic segmentation, selective encoding via information-theoretic constraints, goal-conditioned retrieval with intent routing, reconstructive state synthesis, and adaptive forgetting. We present a formal state model, KV-aware algorithms, security and governance mechanisms including write-path anti-poisoning, enterprise integration pathways, and an evaluation framework with six domain-specific benchmarks. StatePlane demonstrates that long-horizon intelligence can be achieved without expanding context windows or retraining models.