Gradient-free persistent learning through hidden-state episodic recall. @@ -91,8 +91,8 @@ models/tokenizer/ ← Qwen 2.5 tokenizer files ## Citation ```bibtex -@article{niemi2026epimem, - title={One-Shot Learning on Frozen Transformers via Hidden-State Episodic Memory}, +@article{niemi2026clivewearing, + title={Solving the Clive Wearing Problem: One-Shot Episodic Memory for Frozen Transformers}, author={Tommi Niemi}, year={2026}, organization={Rotko Networks}, diff --git a/paper.md b/paper.md index 01b680e..535db8f 100644 --- a/paper.md +++ b/paper.md @@ -1,4 +1,4 @@ -# Clive Wearing: One-Shot Learning on Frozen Transformers via Hidden-State Episodic Memory +# Solving the Clive Wearing Problem: One-Shot Episodic Memory for Frozen Transformers *Named for Clive Wearing, the musician who lost the ability to form new long-term memories but retained his procedural skills. Like Wearing, our frozen backbone retains all its trained capabilities but cannot form new memories through its own weights. We give it an external hippocampus.* @@ -250,7 +250,7 @@ Hidden-state episodic memory enables one-shot, gradient-free, persistent learnin The system demonstrates that biologically-inspired learning mechanisms — Hebbian association, episodic binding, sleep consolidation with optimality certificates — can practically augment modern language models. The frozen backbone retains all capabilities. The synthetic hippocampus gives it new memories. -Like Clive Wearing, the backbone remembers everything it learned during training. Unlike Wearing, we gave it a working hippocampus. +The backbone remembers everything it learned during training. We gave it a working hippocampus. ## References