ai-agent-book 精选快照（<2MB 代码与文档，来自 github.com/bojieli/ai-agent-book）
Build latest book artifacts / build (push) Canceled after 0s
dependency resolution / resolve (3.11) (push) Canceled after 0s
dependency resolution / resolve (3.13) (push) Canceled after 0s
deploy-pages / build (push) Canceled after 0s
deploy-pages / deploy (push) Canceled after 0s
i18n consistency check / check (push) Canceled after 0s
provider adoption tests / test (chapter2/context-compression) (push) Canceled after 0s
provider adoption tests / test (chapter2/prompt-injection) (push) Canceled after 0s
provider adoption tests / test (chapter2/system-hint) (push) Canceled after 0s
provider adoption tests / test (chapter3/log-sanitization) (push) Canceled after 0s
web-search-agent tests / test (push) Canceled after 0s
web-search-agent tests / agentbook (push) Canceled after 0s
Build latest book artifacts / build (push) Canceled after 0s
dependency resolution / resolve (3.11) (push) Canceled after 0s
dependency resolution / resolve (3.13) (push) Canceled after 0s
deploy-pages / build (push) Canceled after 0s
deploy-pages / deploy (push) Canceled after 0s
i18n consistency check / check (push) Canceled after 0s
provider adoption tests / test (chapter2/context-compression) (push) Canceled after 0s
provider adoption tests / test (chapter2/prompt-injection) (push) Canceled after 0s
provider adoption tests / test (chapter2/system-hint) (push) Canceled after 0s
provider adoption tests / test (chapter3/log-sanitization) (push) Canceled after 0s
web-search-agent tests / test (push) Canceled after 0s
web-search-agent tests / agentbook (push) Canceled after 0s
This commit is contained in:
@@ -0,0 +1,257 @@
|
||||
---
|
||||
theme: seriph
|
||||
title: "Lesson 32 — How Do Failed Trajectories Become Learning Signals?"
|
||||
info: "English video course for AI Agents in Depth"
|
||||
author: Bojie Li
|
||||
transition: slide-left
|
||||
mdc: true
|
||||
lineNumbers: false
|
||||
monaco: false
|
||||
aspectRatio: 16/9
|
||||
canvasWidth: 980
|
||||
layout: cover
|
||||
class: cover
|
||||
---
|
||||
|
||||
<div class="course-kicker">Improve · Chapter 8 · Continual Evolution</div>
|
||||
|
||||
# How Do Failed Trajectories Become Learning Signals?
|
||||
|
||||
<p class="course-subtitle">Outcome verification, process rules, Rubrics, and cross-trajectory experience</p>
|
||||
|
||||
<div class="course-cover-meta">Lesson 32 of 42 · 19 minutes · Deriving Learning Signals from Operational Trajectories; Consolidating Experience into Knowledge</div>
|
||||
|
||||
<!-- Presenter cue: Open with the concrete problem. Add personal context in your own words; do not read the slide. -->
|
||||
|
||||
---
|
||||
|
||||
<div class="course-kicker">Improve · Chapter 8 · Continual Evolution</div>
|
||||
|
||||
# Problems this chapter will solve
|
||||
|
||||
<div class="grid grid-cols-3 gap-5 mt-6">
|
||||
<div class="course-card blue">
|
||||
<h3>Lesson 32</h3>
|
||||
<p>How Do Failed Trajectories Become Learning Signals?</p>
|
||||
</div>
|
||||
<div class="course-card green">
|
||||
<h3>Lesson 33</h3>
|
||||
<p>Where Should an Agent Store What It Learns?</p>
|
||||
</div>
|
||||
<div class="course-card purple">
|
||||
<h3>Lesson 34</h3>
|
||||
<p>How Can a Self-Modifying Agent Change Without Drifting?</p>
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<!-- Presenter cue: Orient viewers to the chapter. Name the progression, then highlight today's first problem. -->
|
||||
|
||||
---
|
||||
|
||||
# Why this problem matters
|
||||
|
||||
<div class="grid grid-cols-3 gap-5 mt-6">
|
||||
<div class="course-card blue">
|
||||
<h3>Outcome</h3>
|
||||
<p>Read what changed in the environment.</p>
|
||||
</div>
|
||||
<div class="course-card green">
|
||||
<h3>Process</h3>
|
||||
<p>Locate rule violations and ineffective decisions.</p>
|
||||
</div>
|
||||
<div class="course-card orange">
|
||||
<h3>Meaning</h3>
|
||||
<p>Use a Rubric for dimensions that code cannot settle.</p>
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<!-- Presenter cue: Connect each card to a product or experiment consequence. -->
|
||||
|
||||
---
|
||||
|
||||
# Three ideas to keep in view
|
||||
|
||||
<div class="grid grid-cols-3 gap-5 mt-6">
|
||||
<div class="course-card purple">
|
||||
<h3>Trajectory verifier</h3>
|
||||
<p>Outcome checks + process rules + language Rubric</p>
|
||||
</div>
|
||||
<div class="course-card blue">
|
||||
<h3>Contrastive evidence</h3>
|
||||
<p>Compare success, partial success, and failure</p>
|
||||
</div>
|
||||
<div class="course-card green">
|
||||
<h3>Experience document</h3>
|
||||
<p>Mechanism + conditions + evidence + exceptions</p>
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<!-- Presenter cue: Define unfamiliar terms in plain language; the audience is new to ML training and RL. -->
|
||||
|
||||
---
|
||||
|
||||
# The book's visual model
|
||||
|
||||
<img class="course-figure" src="/images/fig8-2.svg" alt="Three-layer trajectory verification from outcomes to an LLM Rubric">
|
||||
|
||||
<div class="course-caption">Three-layer trajectory verification from outcomes to an LLM Rubric</div>
|
||||
|
||||
<!-- Presenter cue: Trace the diagram in one direction and name the mechanism that matters for this lesson. -->
|
||||
|
||||
---
|
||||
|
||||
# Save the trajectory vs. Consolidate experience
|
||||
|
||||
<div class="grid grid-cols-2 gap-6 mt-5">
|
||||
<div class="course-card orange"><h3>Save the trajectory</h3><ul><li>High detail</li><li>Hard to retrieve</li><li>Incidental actions become noise</li></ul></div>
|
||||
<div class="course-card green"><h3>Consolidate experience</h3><ul><li>Cross-run pattern</li><li>Explicit applicability</li><li>Evidence and counterexamples</li></ul></div>
|
||||
</div>
|
||||
|
||||
<div class="course-caption course-caption-strong">A trajectory is evidence; it is not yet a lesson.</div>
|
||||
|
||||
<!-- Presenter cue: Explain the trade-off; avoid presenting the right column as universally superior. -->
|
||||
|
||||
---
|
||||
|
||||
# Diagnose before updating
|
||||
|
||||
~~~python
|
||||
outcome = environment_verifier(trajectory)
|
||||
violations = process_verifier(trajectory)
|
||||
rubric = semantic_judge(trajectory, outcome)
|
||||
diagnosis = triangulate(outcome, violations, rubric)
|
||||
experience = consolidate(similar_diagnoses)
|
||||
~~~
|
||||
|
||||
<!-- Presenter cue: Walk through the executable idea line by line; keep implementation details for the terminal. -->
|
||||
|
||||
---
|
||||
|
||||
# Test the claim
|
||||
|
||||
<div class="grid grid-cols-2 gap-4 mt-5">
|
||||
<div class="course-card blue">
|
||||
<div class="course-demo-head"><span>8-1</span><span>2 min</span></div>
|
||||
<h3>Diagnose customer-service trajectories with three evidence layers</h3>
|
||||
<p><strong>Observe:</strong> False promises, privacy violations, over-refusal, and cited evidence</p>
|
||||
</div>
|
||||
<div class="course-card blue">
|
||||
<div class="course-demo-head"><span>8-2</span><span>2 min</span></div>
|
||||
<h3>Consolidate several trajectories into experience documents</h3>
|
||||
<p><strong>Observe:</strong> Transfer gain, retrieval cost, negative transfer, and applicability conditions</p>
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<div class="course-caption course-caption-strong">Demo budget: 4 minutes · one contiguous terminal block</div>
|
||||
|
||||
<!-- Presenter cue: State the prediction before running anything. Name the observation that could disconfirm it. -->
|
||||
|
||||
---
|
||||
class: course-terminal
|
||||
---
|
||||
|
||||
<div class="course-kicker">Live demo</div>
|
||||
|
||||
# Switching to the terminal
|
||||
|
||||
~~~bash
|
||||
$ cd chapter8/trajectory-verifier && python demo.py
|
||||
|
||||
$ cd chapter8/gaia-experience && python demo_documents.py
|
||||
~~~
|
||||
|
||||
<div class="course-terminal-watch">Run the command(s), narrate decisions, and point to the observation—not just the output.</div>
|
||||
|
||||
<!-- Presenter cue: Switch windows now. Return to the next slide after every listed experiment is complete. -->
|
||||
|
||||
---
|
||||
|
||||
# What the evidence supports
|
||||
|
||||
<div class="grid grid-cols-3 gap-5 mt-6">
|
||||
<div class="course-card green">
|
||||
<h3>Finding 1</h3>
|
||||
<p>Environment outcomes constrain what a language judge may claim.</p>
|
||||
</div>
|
||||
<div class="course-card blue">
|
||||
<h3>Finding 2</h3>
|
||||
<p>Failures and partial successes reveal conditions hidden by successful runs.</p>
|
||||
</div>
|
||||
<div class="course-card purple">
|
||||
<h3>Finding 3</h3>
|
||||
<p>Cross-trajectory documents can transfer while using fewer tokens than raw history.</p>
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<!-- Presenter cue: Tie each finding to something viewers just observed; distinguish evidence from interpretation. -->
|
||||
|
||||
---
|
||||
layout: center
|
||||
---
|
||||
|
||||
<div class="course-kicker course-kicker-red">Where the claim stops</div>
|
||||
|
||||
# Boundary condition
|
||||
|
||||
<div class="course-boundary">A pattern supported by past trajectories may become obsolete after an API, policy, or environment change.</div>
|
||||
|
||||
<!-- Presenter cue: Say explicitly what this experiment does not establish. -->
|
||||
|
||||
---
|
||||
layout: center
|
||||
---
|
||||
|
||||
<div class="course-kicker">Engineering takeaway</div>
|
||||
|
||||
# Design rule
|
||||
|
||||
<div class="course-rule">Promote experience only with provenance, applicability conditions, counterevidence, and a revalidation trigger.</div>
|
||||
|
||||
<!-- Presenter cue: Present this as a reusable decision rule, then give one counterexample or trade-off. -->
|
||||
|
||||
---
|
||||
|
||||
# Continue the experiment
|
||||
|
||||
<div class="grid grid-cols-2 gap-4 mt-6">
|
||||
<a class="course-link" href="../chapter8/trajectory-verifier/test_verifier.py">
|
||||
<span class="course-link-title">Trajectory verifier tests</span>
|
||||
<span class="course-link-path">chapter8/trajectory-verifier/test_verifier.py</span>
|
||||
</a>
|
||||
<a class="course-link" href="../chapter8/gaia-experience/">
|
||||
<span class="course-link-title">Experience-document implementation</span>
|
||||
<span class="course-link-path">chapter8/gaia-experience/</span>
|
||||
</a>
|
||||
<a class="course-link" href="../book-en/images/fig8-4.svg">
|
||||
<span class="course-link-title">Knowledge-document architecture</span>
|
||||
<span class="course-link-path">book-en/images/fig8-4.svg</span>
|
||||
</a>
|
||||
</div>
|
||||
|
||||
<!-- Presenter cue: Point viewers to the companion paths; do not walk through every extension. -->
|
||||
|
||||
---
|
||||
layout: center
|
||||
class: text-center
|
||||
---
|
||||
|
||||
<div class="course-kicker">Pause and apply</div>
|
||||
|
||||
# Your turn
|
||||
|
||||
<div class="course-big course-reflection">Which detail in a successful trajectory was causal, and how would you distinguish it from coincidence?</div>
|
||||
|
||||
<!-- Presenter cue: Invite viewers to pause the video. Offer your own answer after a short beat. -->
|
||||
|
||||
---
|
||||
layout: center
|
||||
class: text-center
|
||||
---
|
||||
|
||||
<div class="course-kicker">Next · Lesson 33</div>
|
||||
|
||||
<div class="course-next">Choose the artifact that should change: knowledge, instructions, programs, or parameters.</div>
|
||||
|
||||
<div class="course-next-arrow">→</div>
|
||||
|
||||
<!-- Presenter cue: Use this transition to make the course feel continuous rather than episodic. -->
|
||||
Reference in New Issue
Block a user