Authoritative Creators
Authoritative Creators (Individual sources certified within a defined professional scope) holds 3 entries from 3 sources. Authority tiers in this category: B.
Andrej Karpathy
Teaching on neural networks, language models and engineering (1 entry)
Neural Networks:Zero to Hero
Builds neural networks, language models and GPT from scratch in code; good for turning abstract mechanisms into running implementations.
Evidence use: Supports reproducible coding instruction from the course; theoretical and historical claims still need cross-checking against textbooks or papers.
Limitations: Authority is limited to teaching neural networks and language model engineering. The course implementation is a simplified teaching version.
Related concepts: neural-network · transformer
Link last checked: 2026-07-24
Jeremy Howard
Practical deep learning and fast.ai teaching (1 entry)
Practical Deep Learning for Coders
Evidence grade: for discovery only; must not by itself support a formal conclusion.
A practical course for readers with coding experience, approaching deep learning through tasks and code.
Evidence use: Use to discover source material and reproducible examples; verify claims against the specific version and original text.
Limitations: The teaching stack varies by course version; follow the original setup instructions.
Related concepts: neural-network
Link last checked: 2026-11-01
Lilian Weng
Reviews of deep learning and agent research (1 entry)
LLM Powered Autonomous Agents
Evidence grade: for discovery only; must not by itself support a formal conclusion.
A 2023 overview of agent systems, organized around planning, memory and tool use.
Evidence use: Use to discover source material and reproducible examples; verify claims against the specific version and original text.
Limitations: This is not a current product manual; interpret examples and claims in their publication context.
Related concepts: agent
Link last checked: 2026-11-01