20260514:增加新内容
This commit is contained in:
42
concepts/certainty-based-rewards.md
Normal file
42
concepts/certainty-based-rewards.md
Normal file
@@ -0,0 +1,42 @@
|
||||
---
|
||||
title: 确定性奖励 (Certainty-Based Rewards)
|
||||
created: 2025-04-15
|
||||
updated: 2026-05-01
|
||||
type: concept
|
||||
tags: []
|
||||
sources: []
|
||||
---
|
||||
|
||||
# 确定性奖励 (Certainty-Based Rewards)
|
||||
|
||||
**URLVR 的内在奖励范式之一**,从策略的置信度(logits/概率分布)推导奖励,假设更高置信度 = 更正确。
|
||||
|
||||
## 代表方法
|
||||
|
||||
| 方法 | 奖励函数 | 核心思想 |
|
||||
|------|---------|---------|
|
||||
| EM-RL | 轨迹级平均对数概率 | 鼓励低熵(高置信)轨迹 |
|
||||
| RENT | 序列级熵最小化 | 同上,不同归一化 |
|
||||
| RLIF | 自确定性 (KL 散度) | 鼓励输出分布偏离均匀 |
|
||||
| RLSC | 概率自我一致性 | 高概率采样点的自我一致性 |
|
||||
| RLSF | 概率差异 | 交叉样本概率对比 |
|
||||
|
||||
## 理论局限
|
||||
|
||||
[[intrinsic-rewards-sharpening|Sharpening 理论]] 揭示了确定性奖励的根本问题:置信度是模型内部状态——它只反映"模型认为什么是对的",而非"什么客观上是对的"。当模型自信但错误时,确定性奖励在积极强化错误。
|
||||
|
||||
## 对比 Ensemble-Based
|
||||
|
||||
| 确定性奖励 | [[ensemble-based-rewards|集成奖励]] |
|
||||
|-----------|------|
|
||||
| 单次前向传播 | 需多次采样 |
|
||||
| 计算成本低 | 计算成本高 |
|
||||
| 完全依赖模型内部状态 | 通过多样本交叉验证 |
|
||||
| 同样受 Sharpening 限制 | 同样受 Sharpening 限制 |
|
||||
|
||||
## 相关概念
|
||||
|
||||
- [[ensemble-based-rewards]] — 另一内在范式
|
||||
- [[intrinsic-rewards-sharpening]] — 统一理论
|
||||
- [[unsupervised-rlvr]] — URLVR 全景
|
||||
- [[he-urlvr-sharpening-2026]] — 综述参考
|
||||
Reference in New Issue
Block a user