d953030747
* feat: add v1/v2 versioning and language selector for mdbook - Copy current content to v1/ directory (1st Edition) - Create v2/ directory with new TOC structure (2nd Edition) and placeholder chapters - Add version selector (V1/V2) and language toggle (EN/ZH) in top-right nav bar - Add build scripts: build_mdbook_v1.sh, build_mdbook_v2.sh - Update assemble_docs_publish_tree.py to support v1/v2 deployment layout - Fix mdbook preprocessor to use 'sections' key (v0.4.43 compatibility) - Update .gitignore for new build artifact directories - Deployment layout: / = v2 EN, /cn/ = v2 ZH, /v1/ = v1 EN, /v1/cn/ = v1 ZH Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * build: update CI to build and verify all four books (v1/v2 x EN/ZH) - Clarify step names: "Build v2 (EN + ZH)" and "Build v1 (EN + ZH)" - Add verification step to check all four index.html outputs exist - Deploy workflow assembles: / = v2 EN, /cn/ = v2 ZH, /v1/ = v1 EN, /v1/cn/ = v1 ZH Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> * fix: gracefully skip missing TOC entries instead of crashing resolve_toc_target() now returns None for missing files instead of raising FileNotFoundError. This fixes v1 EN build where chapter index files reference TOC entry names that don't match actual filenames. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2.8 KiB
Explainable AI Systems
Over the past decade, driven by the cost-performance ratio of computational power and data scale surpassing critical thresholds, connectionist model architectures represented by deep neural networks and statistical learning paradigms (hereinafter referred to as deep learning) have achieved breakthrough advances in feature representation capabilities, greatly advancing the development of artificial intelligence and achieving remarkable results in many scenarios. For example, face recognition accuracy has reached over 97%, and Google's intelligent voice assistant achieved a 92.9% correct response rate in 2019 tests. In these typical scenarios, deep learning's intelligent performance has surpassed that of ordinary humans (and even experts), reaching a tipping point for technology replacement. In recent years, in domains where business logic is technology-friendly or where ethical regulations are temporarily sparse---such as security, real-time scheduling, process optimization, competitive gaming, and information feed distribution---artificial intelligence and deep learning have achieved rapid technical and commercial breakthroughs.
Having tasted success, no domain wants to miss out on the benefits of technological progress. However, when the commercial application of deep learning enters domains that are technology-sensitive and closely related to human survival or safety---such as autonomous driving, finance, healthcare, and judicial high-risk application scenarios---the existing business logic encounters resistance during technology replacement, leading to slowdowns or even failures in commercialization. The root cause is that the business logic and underlying ethical regulations of these scenarios center on stable, traceable accountability and responsibility distribution; yet the models produced by deep learning are black boxes from which we cannot extract any information about model behavior from the model's structure or weights, rendering the accountability and distribution mechanisms in these scenarios inoperative and causing technical and structural difficulties for AI in business applications. Moreover, model interpretability has attracted national-level attention, with relevant institutions issuing related policies and regulations.
Therefore, from both the commercial promotion and regulatory perspectives, we need to open up the black box model and provide explanations for models. Explainable AI is precisely the technology that addresses this class of problems.
The learning objectives of this chapter include:
-
Understand the goals and application scenarios of explainable AI
-
Master the common types of explainable AI methods and their representative techniques
-
Reflect on the future development of explainable AI methods
:maxdepth: 2
explainable_ai