NuoNuo: Hippocampal memory module prototype

Hopfield + Hebbian hybrid memory system for LLMs.
Two nights of experiments (16 iterations), validated on LongMemEval (ICLR 2025).

Architecture:
- Single-hop: Two-Stage Hopfield (NN top-20 → softmax settle)
- Multi-hop: Hebbian W matrix with WTA pattern separation
- 64% on LongMemEval (500 questions), retrieval-only, no LLM dependency
- 4ms latency @ 20K memories, ~1GB VRAM

Key findings:
- Hopfield attention solved noise tolerance (20% → 100% vs flat Hebbian)
- WTA pattern separation enables 20K+ capacity
- Multi-hop associative chains (6 hops, CosSim=1.0) — RAG can't do this
- MiniLM-L6 is optimal (discrimination gap > absolute similarity)
- Paraphrase cue augmentation: 55% → 100% on synthetic, 36% → 64% on benchmark
- SNN encoder viable (CosSim 0.99) but not needed for current architecture
This commit is contained in:
2026-04-07 10:37:24 +01:00
commit d923aa1e31
65 changed files with 13148 additions and 0 deletions

37
doc/p3_scale_ceiling.md Normal file
View File

@@ -0,0 +1,37 @@
# P3: 突破 20K 80% 天花板
## 结论:天花板来自 embedding 模型,不是架构
### Top-K Coverage 分析
| K | N=20K |
|---|-------|
| 5 | 80% |
| 50 | 80% |
| 200 | 80% |
K 从 5 增加到 200coverage 不变。那 2 个 failure 的 paraphrase 在 embedding 空间里根本不是正确 cue 的最近邻——即使只有 10 条记忆也找不到。
### 架构优化无效
| 方法 | bg=20K |
|------|--------|
| Two-stage K=5 | 60% |
| Two-stage K=200 | 30% (更大 K 更差!) |
| Hierarchical clustering | 40% |
更大的 K 引入更多噪声Hopfield attention 被分散。Hierarchical 也没帮助。
### 根因
失败的 paraphrase 对embedding cosine similarity:
- "Need observability" ↔ "Let's set up monitoring" = 0.257
- "When's the standup?" ↔ "Team meeting schedule" = 0.375
这些在 MiniLM 的 embedding 空间里根本不算"相似"。任何基于 embedding 距离的检索方法都无法找到它们。
### 解法 = P2
**Paraphrase augmentation 是唯一解法**(已验证 55% → 100%)。
不需要改架构。不需要换 K。不需要 hierarchical memory。只需要在存储时覆盖更多的表达方式。