ANAMNESIS: Forecasting and Self-Healing of Memory Corruption in LLM Agents
Skip to content
Projects<br>ANAMNESIS: Forecasting and Self-Healing of Memory Corruption in LLM Agents
August 3, 2026<br>A living document<br>This is not a finished paper. It is the working record of a final-year project that runs to April 2027, and I update it as the work moves rather than writing it once at the end. Section 8 is the changelog, and the status table below is the fastest way to see what is actually true today.<br>Two disciplines carry over from the thesis into this page. Nothing is claimed as a result until it has been measured, and anything still pending says so in plain words. Where the work has already proved me wrong, that is on the page too, because those are the parts worth reading.<br>1. The problem<br>A memory system reads your conversation and distils it into short facts it can recall later. That distilling step is done by a language model, and a language model can write down something you never said. This much is well known.<br>The harder problem is the one that arrives afterwards. A fact can be recorded perfectly and still stop being true. You move city. You change jobs. You sell the laptop. Nothing was hallucinated and nothing was attacked; the world simply moved, and the memory did not.<br>I hit this in my own system before I read about it anywhere. I told the assistant I had switched to a MacBook. Consolidation admitted the new fact thirteen seconds before my next question. Retrieval pulled it into context. The answer still said ThinkPad.
Press enter or space to select a node. You can then use the arrow keys to move the node around. Press delete to remove it and escape to cancel.<br>Press enter or space to select an edge. You can then press delete to remove it or escape to cancel.
Nothing in that sequence is a bug in the ordinary sense. Every component did its job. The gap is that no component's job was to notice that an old fact had just been outlived, and that is the gap this project sits in.<br>2. What ANAMNESIS is<br>A local-first memory system for long-running conversations with a language model. Memory is organised in four tiers, loosely mapped to how cognitive neuroscience decomposes human long-term memory: a working tier for the live turn, an append-only episodic record, a distilled semantic tier built by out-of-band consolidation, and a procedural tier for response patterns. The whole store is a single SQLite file. The only network call in the write path is to the chat model itself; retrieval, entailment checking, and risk scoring all run locally on CPU.<br>Every candidate fact travels the same short path before it becomes memory.
Press enter or space to select a node. You can then use the arrow keys to move the node around. Press delete to remove it and escape to cancel.<br>Press enter or space to select an edge. You can then press delete to remove it or escape to cancel.
The load-bearing property is that nothing enters semantic memory unexamined, and nothing is ever deleted. A rejected fact is quarantined with its audit record intact, so a refusal can be inspected rather than silently swallowed.<br>3. Why the obvious version is not the contribution<br>This is the section I would want to read first if someone else had written this page.<br>I started in May 2026 with a claim I was confident about: that nobody was measuring or gating whether memory consolidation is faithful. A literature scan the following day dismantled it. Brain-inspired tiered memory had been built several times over. Persistent single-conversation memory had shipped. A benchmark purpose-built for hallucination in memory systems, HaluMem, had been published in November 2025. Write-time gates had been specified.<br>So the claim moved: nobody had integrated these pieces and tested whether they help each other. That survived about three weeks. In June I ran an adversarial audit of the 2024 to 2026 field, specifically trying to kill my own narrowed claim, and it worked. ProMem had already built write-time verification, benchmarked it on HaluMem, and beaten Mem0 doing it. Several other systems had shipped gated or trained memory operations in the same window.<br>What survived the third pass is the part the audit could not find anywhere. Every one of those systems checks a memory in the present tense: is this supported right now. None of them forecast which currently-clean memory is going to rot, and none close a repair loop driven by that forecast.<br>That is what the project is, and the narrowness is deliberate.
Press enter or space to select a node. You can then use the arrow keys to move the node around. Press delete to remove it and escape to cancel.<br>Press enter or space to select an edge. You can then press delete to remove it or escape to cancel.
The three mechanisms and their honest novelty scoping:<br>MechanismThe question it asksNovelGATE, past tenseWas this supported when it was writtenNo. ProMem got there first.PREDICT, future tenseWill it corrupt within the next N...