A formal architecture with typed provenance graphs, a resolver, and a generate-verify-revise mediator enforces autobiographical assertion boundedness for persistent AI agents.
MemTX: Transactional Belief Commit for Stateful Agent Memory
6 Pith papers cite this work. Polarity classification is still indexing.
6
Pith papers citing it
abstract
LLM agents increasingly coordinate through persistent shared memory: one agent's write becomes another agent's premise, and eventually a tool call with real side effects. Current agent memory systems treat every accepted write as immediately actionable truth, so a polluted tool result, a stale update, or a teammate's half-finished note can silently drive an irreversible action. We argue that a memory write is not a belief commit. We present MemTX, a transactional belief-commit protocol. Each record carries evidence, permissions, provenance, and validity. Writes are staged inside snapshot-isolated transactions and admitted by a validate-and-commit pipeline, irreversible tool calls are gated on in-flight belief state, and retracting a belief triggers typed cascading repair of its derived records and tool side effects. Two invariants, action-safety gating and cascade-repair completeness, are machine-checked by property-based testing and bounded exhaustive enumeration of 5.5 million protocol states, with zero violations. Across five backbones from three model families, MemTX leads all eight baselines with paired-McNemar significance on four backbones and statistically ties the best baseline on the fifth and strongest, while remaining the only method with zero downstream harm on every backbone. Backbone capability does not substitute for commit discipline.
citation-role summary
background 2
citation-polarity summary
years
2026 6roles
background 2representative citing papers
On three native tool-calling endpoints, BFCL rerun noise is under 1.1 percentage points at temperature 0, while matched prompt rewrites shift outcomes by 5 to 21 points and change failure types.
Adding a mandatory memory-state output field does not significantly improve memory-policy classification in two large language models, while exposing the state definitions alone does improve it.
A deterministic activation contract called the Continuity Kernel records one of four decisions (Commit, Reject, Quarantine, Defer) inside an atomic transaction, and only Commit advances an agent's authoritative branch head.
The paper proposes Graph Engineering, a framework in which tasks, agents, and runtime states are represented as explicit, evolving graphs to support system-level intelligence in LLM agent systems.
citing papers explorer
-
Stored Is Not Supported: Typed Provenance and Assertion Guardrails for Persistent AI Agents
A formal architecture with typed provenance graphs, a resolver, and a generate-verify-revise mediator enforces autobiographical assertion boundedness for persistent AI agents.
-
Noise Floor Audit for Agent Benchmarks
On three native tool-calling endpoints, BFCL rerun noise is under 1.1 percentage points at temperature 0, while matched prompt rewrites shift outcomes by 5 to 21 points and change failure types.
-
Explicit State Elicitation Is Not Enough: A Controlled Audit of Memory-Policy Classification
Adding a mandatory memory-state output field does not significantly improve memory-policy classification in two large language models, while exposing the state definitions alone does improve it.
-
Beyond Memory: A Transactional Continuity Kernel for Long-Lived AI Agents
A deterministic activation contract called the Continuity Kernel records one of four decisions (Commit, Reject, Quarantine, Defer) inside an atomic transaction, and only Commit advances an agent's authoritative branch head.
-
Graph Engineering in the Era of LLM Agents: From Individual Intelligence to System Intelligence
The paper proposes Graph Engineering, a framework in which tasks, agents, and runtime states are represented as explicit, evolving graphs to support system-level intelligence in LLM agent systems.
- SuperLocalMemory 4.0: The Governed Memory Operating System for AI Agents