46 ai与人的对话模式

毕苏林
Join to follow...
Follow/Unfollow Writer: 毕苏林
By following, you’ll receive notifications when this author publishes new articles.
Don't wait! Sign up to follow this writer.
WriterShelf is a privacy-oriented writing platform. Unleash the power of your voice. It's free!
Sign up. Join WriterShelf now! Already a member. Login to WriterShelf.
爱科学,也爱文艺;重逻辑,也重情感。以最硬核的科幻为壳,写最柔软的人间故事。愿以文字为桥,结识品味相投的读友。
31   0  
·
2026/04/17
·
2 mins read


ai与人对话系统不是静态分类器,

而是动态识人系统。

看错人、归错类,是正常且必然的,

关键在于:能否像人类一样,及时修正、重新归类。

 

传统AI(包括大模型)的缺陷:

 

- 要么一次性分类,永不更新

- 要么黑盒模糊,永远不承认自己错了

- 完全没有**“我看错了,我重新判断你”**的机制

 

而我提出的:

归类 → 映射 → 检测错误 → 重新归类 → 新映射

 

这是人类级社交智能,不是机器问答。

动态硬归类-自适应映射(DHC-AM)对话模型

 

一、基础符号

X:用户特征空间

C={c1,c2,...,cn}:用户硬分类集合

Q:用户请求

Y:AI响应

τ:置信度阈值

D:错配检测函数,D=1需重归类,D=0保持

t:时间步

 

二、初始硬归类 t=0

ĉ₀ =

若满足优先级规则 → c_k

否则 maxP(ci|x)≥τ → argmax P(ci|x)

否则为空类别

 

三、映射

y₀ = g(ĉ₀, q)

 

四、错配检测

D(ĉ_{t-1}, q_t, y_{t-1}, user_feedback) = 1 或 0

 

五、动态重归类 t≥1

ĉ_t =

新规则→c_k

否则 maxP(ci|x_hist+x_t)≥τ→argmax P(ci|x_hist+x_t)

否则为空

 

六、自适应映射

y_t = g(ĉ_t, q_t)

 

七、整体闭环

y_t = g( C(X_t), q_t )

 

完整LaTeX公式:

 

\hat{c}0 =

\begin{cases}

c_k, & \text{规则匹配} \

\arg\max\limits{c_i \in C} P(c_i|x), & \max P\geq\tau \

\varnothing, & \text{else}

\end{cases}

 

 

 


WriterShelf™ is a unique multiple pen name blogging and forum platform. Protect relationships and your privacy. Take your writing in new directions. ** Join WriterShelf**
WriterShelf™ is an open writing platform. The views, information and opinions in this article are those of the author.


Article info

This article is part of:
分類於:

合計:335字


Share this article:



Join the discussion now!
Don't wait! Sign up to join the discussion.
WriterShelf is a privacy-oriented writing platform. Unleash the power of your voice. It's free!
Sign up. Join WriterShelf now! Already a member. Login to WriterShelf.