← Agent Architectures: Advanced Strategies for Intelligent LLM Systems

References and Supporting Papers

Collect supporting papers and references for the agent architecture concepts discussed throughout the book.

Agent Architectures: Advanced Strategies for Intelligent LLM Systems Chapter 90 of 16

01 02 03 04 05 06 07 08 09 10 11 12 80 81 82 90

Concepts

WHAT YOU NEED TO KNOW

STARTING POINT NOT PROOF

A reference collection opens further reading rather than closing an argument. Papers listed beside a chapter suggest where an idea came from or where it can be explored next. Presence on the list certifies relevance to a theme, never correctness of the book’s claim.

VERIFICATION BEFORE PUBLICATION

Several entries carry explicit warnings that titles, authors, dates, or relevance still need checking. That flag is a publication boundary: unverified material may guide exploration but must not ship as established fact. The honest move is to mark uncertainty in the open and resolve it before a formal pass.

candidate reference
  ↓
check title, authors, date
  ↓
check relevance to the claim
  ↓
keep, correct, or remove

ANCHORS PER CHAPTER

Each chapter pairs with papers that anchor its mechanisms, from transformers and few-shot learning behind the introduction to reasoning-and-acting loops behind methodology. Reflection chapters lean on self-critique work while systems chapters lean on multi-agent coordination studies. The mapping shows which intellectual neighbor each design idea lives beside.

RELEVANCE IS NOT SUPPORT

A paper can be relevant to tool interfaces, declarative programs, or reflection loops without proving the book’s specific application of it. Filtering lenses built on tool-use research, for example, extrapolate beyond what the cited work demonstrated. Naming the gap between inspiration and evidence keeps a reading list from becoming false authority.

CLAIMS NEED SOURCES

Some entries are flagged because a cited title may not exist as described or its connection to the chapter is unconfirmed. Automated reasoning, retrieval planning, and optimization papers fall in this group. Until title and relevance are confirmed, such candidates must be treated as leads to investigate rather than foundations to build on.

LIVING COLLECTION

A reference list for a fast-moving field decays from the day it is written. New systems for conversation frameworks, role teams, and prototyping appear while older links shift meaning. Treating the collection as a maintained starting point, corrected by readers and builders who use it, matches the spirit the chapter requests.

CANDIDATES STAY QUARANTINED

The freestyle chapter holds its references apart as candidates rather than mixing them into the main list. Each awaits confirmation of title, author list, date, and relevance before it can anchor a claim. Quarantine is a kindness to the reader: possible leads stay visible without lending them authority they have not earned.

READ THEN BUILD

The list points outward to essays on media, programming craft, alignment, and interaction at scale alongside technical papers. Its purpose is fulfilled when a reader follows one thread far enough to test an idea in a small system. References serve the book the way tools serve an agent: they supply material while judgment about what to trust stays with the human.

This section collects the papers, essays, and project references behind the themes in the book. Each work appears once, with its verified identifier. Works are listed as supporting reading: where a chapter relies on a source directly, the source is cited in the text.

Canonical entries

Vaswani et al. (2017). Attention Is All You Need. NeurIPS 2017. arXiv:1706.03762 — Introduced the transformer architecture foundational to modern LLMs.
Brown et al. (2020). Language Models are Few-Shot Learners (GPT-3). NeurIPS 2020. arXiv:2005.14165 — Demonstrates zero/few-shot learning without fine-tuning.
OpenAI (2023, Jun 13). Function calling and other API updates. openai.com — Product note (not a paper) describing tool-calling interfaces and structured model outputs.
Yao et al. (2022). ReAct: Synergizing Reasoning and Acting in Language Models. ICLR 2023. arXiv:2210.03629 — Reasoning traces interleaved with actions; relevant to reasoning-and-acting loops.
Schick et al. (2023). Toolformer: Language Models Can Teach Themselves to Use Tools. arXiv:2302.04761 — Self-supervised API-call learning; relevant to tool-use interfaces.
Shinn et al. (2023). Reflexion: Language Agents with Verbal Reinforcement Learning. NeurIPS 2023. arXiv:2303.11366 — Verbal self-reflection over trials; relevant to reflection and feedback loops. Does not by itself establish companion-agent memory or content filtering.
Madaan et al. (2023). Self-Refine: Iterative Refinement with Self-Feedback. NeurIPS 2023. arXiv:2303.17651 — Same-model generator/feedback/refiner loop; the model-as-critic structure.
Qin et al. (2023). ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs. ICLR 2024 Spotlight. arXiv:2307.16789 — Basis for tool-augmented capabilities; the ToolBench dataset and evaluation are introduced inside this work (cited as ToolBench therein, not as an independent paper).
Patil et al. (2023). Gorilla: Large Language Model Connected with Massive APIs. arXiv:2305.15334 — Fine-tuned API-call generation; relevant to tool and API use.
Wu et al. (2023). AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation. arXiv:2308.08155 — Conversation-first multi-agent framework; practical implementation of roles and coordination.
Li et al. (2023). CAMEL: Communicative Agents for “Mind” Exploration of Large Language Model Society. NeurIPS 2023. arXiv:2303.17760 — Role-played multi-agent dialogue; relevant to role assignment in simulations.
Chen et al. (2023). AgentVerse: Facilitating Multi-Agent Collaboration and Exploring Emergent Behaviors. arXiv:2308.10848 — Multi-agent collaboration; relevant to centralized vs decentralized architectures.
Khattab et al. (2023). DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines. ICLR 2024. arXiv:2310.03714 — Declarative language-model programs; relevant to modular program design.
Rajani et al. (2019). Explain Yourself! Leveraging Language Models for Commonsense Reasoning. ACL 2019 (P19-1487). arXiv:1906.02361 — Background on explanation and reasoning traces.
Bai et al. (2022). Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback. Anthropic. arXiv:2204.05862 — Reward feedback loops and alignment strategies.
Bai et al. (2022). Constitutional AI: Harmlessness from AI Feedback. Anthropic. arXiv:2212.08073 — Building AI systems aligned with stated principles.
McLuhan, M. (1964). Understanding Media: The Extensions of Man. McGraw-Hill, New York. — “The medium is the message”; the AI-as-medium discussion.
Sutton, R. (2019, Mar 13). The Bitter Lesson. Essay. incompleteideas.net — Argues that general methods that scale with computation, search and learning, have outperformed hand-built knowledge.
Norvig, P. Teach Yourself Programming in Ten Years. Essay. norvig.com — Programming as craft; agents change the entry point, not the need for judgment.
Gottweis et al. (2025). Accelerating scientific discovery with Co-Scientist (first posted as Towards an AI co-scientist). arXiv:2502.18864 — Multi-agent generate, reflect, rank, and evolve pipeline for hypothesis generation; the paper used as the worked example in Chapter 10.
Wooldridge & Jennings (1995). Intelligent Agents: Theory and Practice. Knowledge Engineering Review 10(2). doi:10.1017/S0269888900008122 — Agent theory, architectures, and languages.
Maes, P. (1994). Agents that Reduce Work and Information Overload. CACM 37(7). doi:10.1145/176789.176792 — Software agents acting on behalf of users.
OpenAI (2023). GPT-4 System Card (pdf) and GPT-4 Technical Report (arXiv:2303.08774); Introducing GPTs (openai.com, Nov 2023). — Browse/code/tool behavior and custom GPTs.