跳到正文
𝕏 elvis(x.com)·· 5 天前品質評分56
重點導讀

AI 助理要持續學習,就得知道哪些資訊該記下來、怎麼記,以及之後何時取用。

  • Elvis Saravia 整理了 Meta 與 RIME 的不同做法:Meta 先找出較好的記憶提示詞,再用相關資料微調模型。
  • RIME 則用強化學習,讓模型一邊解題、一邊寫下記憶,並獎勵日後能幫上忙的筆記。這篇整理也列出 Google 等相關研究連結,呈現這個研究方向的不同路徑。

AI 伴隨問答

自備 Key 深度分析

針對這篇報導有疑問?可直接點選快捷問題或向 AI 追問背景脈絡與技術細節:

elvis
elvis@omarsar0
·5 天前
𝕏

Highly recommended read.

Self-improving memory is starting to look like an important component for continual-learning agents.

So pay close attention to this research direction if you are studying recursive self-improvement.

Just last week, I featured a few papers on the topic, and this week both Meta and Google published on it.

(find more references towards the end)

Meta lets a model design its own memory habits.

Then they train those habits into the model weights.

The goal is to allow the system to learn the memory design (what to memorize, how to memorize, and what to recall).

This work is actually in the same vein as RIME, which I wrote some notes on last week: https://t.co/wLJy6ntdaJ

Meta first searches for better memory instructions (prompts the model to rewrite itself), then fine-tunes the model (SFT with LoRA) on runs that used the best ones.

RIME, on the other hand, skips the search and uses RL to train one model to solve tasks and write memories, rewarding notes that help on later tasks. It also trains the model to explore so it collects useful information.

Different ways to address the problem.

More useful links below:

Google Self-Improving Memory for Continual Learning - https://t.co/ons3yOcSkW

How to Train Memory Writing for Continual Learning - https://t.co/wLJy6ntdaJ

Can LLM Agents Organize Their Own Memory - https://t.co/uibQW9oKmD

Continual-ARC - https://t.co/BPAO8Cb0s3

Adaptive Memory Selection for LLM Agents - https://t.co/qxLyIr0DyN

MemCredit - https://t.co/AlrGHCvZ8O

推文來源 · X (Twitter)在 X 參與討論
來源: elvis (x.com)
在 X 查看推文

讀者討論

理性交流 · 即時討論