From 281af1183e1c4b214130253e197359c6624403e3 Mon Sep 17 00:00:00 2001 From: Yuchen Wang <93700456+wangyuchen333@users.noreply.github.com> Date: Sat, 5 Apr 2025 10:35:54 +0800 Subject: [PATCH] =?UTF-8?q?=E6=B7=BB=E5=8A=A0=20CS294=20=E8=AF=BE=E7=A8=8B?= =?UTF-8?q?=E7=9A=84=E8=AF=A6=E7=BB=86=E6=96=87=E6=A1=A3=E5=92=8C=E8=B5=84?= =?UTF-8?q?=E6=BA=90?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- .../CS294.en_20250405103123.md | 0 .../CS294.en_20250405103315.md | 82 +++++++++++++++++++ docs/大语言模型与智能体/CS294.en.md | 82 +++++++++++++++++++ .../大语言模型与智能体/llm_zero_to_hero.en.md | 0 4 files changed, 164 insertions(+) create mode 100644 .history/docs/大语言模型与智能体/CS294.en_20250405103123.md create mode 100644 .history/docs/大语言模型与智能体/CS294.en_20250405103315.md create mode 100644 docs/大语言模型与智能体/CS294.en.md create mode 100644 docs/大语言模型与智能体/llm_zero_to_hero.en.md diff --git a/.history/docs/大语言模型与智能体/CS294.en_20250405103123.md b/.history/docs/大语言模型与智能体/CS294.en_20250405103123.md new file mode 100644 index 00000000..e69de29b diff --git a/.history/docs/大语言模型与智能体/CS294.en_20250405103315.md b/.history/docs/大语言模型与智能体/CS294.en_20250405103315.md new file mode 100644 index 00000000..e57bb419 --- /dev/null +++ b/.history/docs/大语言模型与智能体/CS294.en_20250405103315.md @@ -0,0 +1,82 @@ +# CS294/194-196 Large Language Model (LLM) Agents Course + +## Course Overview + +- **University**: UC Berkeley +- **Prerequisites**: None +- **Programming Language**: Python +- **Course Difficulty**: 🌟🌟🌟 +- **Estimated Study Hours**: 100 hours +- The course covers LLM reasoning, tool usage, multi-agent collaboration, and application areas such as code generation and robotics. +- It provides abundant resources, including the course website [CS294/194-196 Course Homepage](http://rdi.berkeley.edu/llm-agents/f24), MOOC site [Large Language Model Agents MOOC](https://llmagents-learning.org/f24), YouTube lecture videos, and GitHub notes. +- Assignments include reading summaries, experiments, and projects such as participating in the [LLM Agents Hackathon](http://rdi.berkeley.edu/llm-agents-hackathon/). +- Extended learning includes the spring course [CS294/194-280 Advanced Large Language Model Agents](http://rdi.berkeley.edu/adv-llm-agents/sp25) and related GitHub resources [Awesome LLM Agents](https://github.com/kaushikb11/awesome-llm-agents). + +## Course Structure and Content + +The course explores the core content of LLM agents, including: + +- **LLM Basics and Reasoning**: Covers techniques like Chain-of-Thought and Self-Consistency to enhance model reasoning through intermediate steps. +- **Agent Infrastructure**: Includes Retrieval-Augmented Generation (RAG), Tool Use, and multi-agent collaboration frameworks (e.g., AutoGen). + +While the course has no formal prerequisites, students are encouraged to have a background in machine learning and deep learning, such as courses CS182, CS188, or CS189. + +## Course Resources + +To support self-study, the course offers abundant resources: + +- **Course Website**: [CS294/194-196 Course Homepage](http://rdi.berkeley.edu/llm-agents/f24), containing the syllabus, registration info, and contact details. +- **MOOC Website**: [Large Language Model Agents MOOC](https://llmagents-learning.org/f24), providing additional labs, certifications, and interactive communities for those unable to attend the formal course. +- **Lecture Videos**: All 12 lectures are recorded on YouTube. The schedule and links are as follows: + +| Date | Topic | Guest Lecturer | Video Link | +|------------|-------------------------------------------------|---------------------------------------|-------------------------------------------------| +| Sept 9 | LLM Reasoning | Denny Zhou, Google DeepMind | [Video](https://www.youtube.com/live/QL-FS_Zcmyo) | +| Sept 16 | LLM Agents: History and Overview | Shunyu Yao, OpenAI | [Video](https://www.youtube.com/watch?v=RM6ZArd2nVc) | +| Sept 23 | Agent AI Framework & AutoGen, Building Multimodal Knowledge Assistants | Chi Wang, AutoGen-AI; Jerry Liu, LlamaIndex | [Video](https://www.youtube.com/live/OOdtmCMSOo4) | +| Sept 30 | Generative AI Trends in Business and Key Components for Building Successful Agents/Applications | Burak Gokturk, Google | [Video](https://www.youtube.com/live/Sy1psHS3w3I) | +| Oct 7 | Composite AI Systems & DSPy Framework | Omar Khattab, Databricks | [Video](https://www.youtube.com/live/JEMYuzrKLUw) | +| Oct 14 | Software Development Agents | Graham Neubig, Carnegie Mellon University | [Video](https://www.youtube.com/live/f9L9Fkq-8K4) | +| Oct 21 | Enterprise Workflow AI Agents | Nicolas Chapados, ServiceNow | [Video](https://www.youtube.com/live/-yf-e-9FvOc) | +| Oct 28 | A Unified Framework for Neuro-Symbolic Decision-Making | Yuandong Tian, Meta AI (FAIR) | [Video](https://www.youtube.com/live/wm9-7VBpdEo) | +| Nov 4 | Project GR00T: The Universal Robotics Blueprint | Jim Fan, NVIDIA | [Video](https://www.youtube.com/live/Qhxr0uVT2zs) | +| Nov 18 | Open-Source and Science in the Era of Foundation Models | Percy Liang, Stanford University | [Video](https://www.youtube.com/live/f3KKx9LWntQ) | +| Nov 25 | Measuring Agent Abilities and Anthropic's RSP | Ben Mann, Anthropic | [Video](https://www.youtube.com/live/6y2AnWol7oo) | +| Dec 2 | Building Safe and Trustworthy AI Agents, and AI Policies Based on Science and Evidence | Dawn Song, UC Berkeley | [Video](https://www.youtube.com/live/QAgR4uQ15rc) | + +Each lecture has corresponding reading materials; see the course website [CS294/194-196 Course Homepage](http://rdi.berkeley.edu/llm-agents/f24#syllabus). + +- **Course Notes**: [GitHub](https://github.com/rajdeepmondaldotcom/CS294_LLM_Agents_Notes_Fall2024) + +## Relevant Papers and Resources + +The course focuses on the following papers and frameworks, categorized by technical direction: + +### Reasoning and Planning + +- **ReAct**: A framework combining reasoning and action to enhance task-solving capabilities. Paper: [ReAct: Synergizing Reasoning and Acting in Language Models](https://arxiv.org/abs/2210.03629). +- **Chain-of-Thought**: Using intermediate steps to stimulate reasoning in models. Paper: [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://arxiv.org/abs/2201.11903). +- **Chain-of-Thought Reasoning Without Prompting**: Exploring reasoning through chain-of-thought without prompting. Paper: [Chain-of-Thought Reasoning Without Prompting](https://arxiv.org/abs/2402.10200). + +### Agent Frameworks + +- **AutoGen**: A framework for multi-agent conversation development. Paper: [AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation](https://arxiv.org/abs/2308.08155). +- **DSPy**: A programming framework for composite AI systems. Paper: [DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines](https://arxiv.org/abs/2310.03714). + +### Application Scenarios + +- **Code Generation**: + - **SWE-agent**: An agent interface for automating software engineering. Paper: [SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering](https://arxiv.org/abs/2402.01030). +- **Robotics**: + - **Voyager**: An embodied agent based on LLMs. Paper: [Voyager: An Open-Ended Embodied Agent with Large Language Models](https://arxiv.org/abs/2305.16291). + +## Extended Learning Suggestions + +For further study, we recommend: + +- **Practical Projects**: Participate in the [LLM Agents Hackathon](http://rdi.berkeley.edu/llm-agents-hackathon/), aiming to build agents supporting tool usage (e.g., web automation). The Hackathon provides tracks such as applications, benchmarks, and foundational research. + +- **Advanced Courses**: Explore the spring course [CS294/194-280 Advanced Large Language Model Agents](http://rdi.berkeley.edu/adv-llm-agents/sp25), focusing on advanced topics like LLM reasoning, mathematical proofs, and code generation. +- **Online Resources**: + - [Large Language Model Agents MOOC](https://llmagents-learning.org/f24): Offers course materials, labs, and certifications, ideal for self-learners. + - [Awesome LLM Agents](https://github.com/kaushikb11/awesome-llm-agents): A collection of frameworks, papers, and projects on LLM agents, ideal for deep research. \ No newline at end of file diff --git a/docs/大语言模型与智能体/CS294.en.md b/docs/大语言模型与智能体/CS294.en.md new file mode 100644 index 00000000..e57bb419 --- /dev/null +++ b/docs/大语言模型与智能体/CS294.en.md @@ -0,0 +1,82 @@ +# CS294/194-196 Large Language Model (LLM) Agents Course + +## Course Overview + +- **University**: UC Berkeley +- **Prerequisites**: None +- **Programming Language**: Python +- **Course Difficulty**: 🌟🌟🌟 +- **Estimated Study Hours**: 100 hours +- The course covers LLM reasoning, tool usage, multi-agent collaboration, and application areas such as code generation and robotics. +- It provides abundant resources, including the course website [CS294/194-196 Course Homepage](http://rdi.berkeley.edu/llm-agents/f24), MOOC site [Large Language Model Agents MOOC](https://llmagents-learning.org/f24), YouTube lecture videos, and GitHub notes. +- Assignments include reading summaries, experiments, and projects such as participating in the [LLM Agents Hackathon](http://rdi.berkeley.edu/llm-agents-hackathon/). +- Extended learning includes the spring course [CS294/194-280 Advanced Large Language Model Agents](http://rdi.berkeley.edu/adv-llm-agents/sp25) and related GitHub resources [Awesome LLM Agents](https://github.com/kaushikb11/awesome-llm-agents). + +## Course Structure and Content + +The course explores the core content of LLM agents, including: + +- **LLM Basics and Reasoning**: Covers techniques like Chain-of-Thought and Self-Consistency to enhance model reasoning through intermediate steps. +- **Agent Infrastructure**: Includes Retrieval-Augmented Generation (RAG), Tool Use, and multi-agent collaboration frameworks (e.g., AutoGen). + +While the course has no formal prerequisites, students are encouraged to have a background in machine learning and deep learning, such as courses CS182, CS188, or CS189. + +## Course Resources + +To support self-study, the course offers abundant resources: + +- **Course Website**: [CS294/194-196 Course Homepage](http://rdi.berkeley.edu/llm-agents/f24), containing the syllabus, registration info, and contact details. +- **MOOC Website**: [Large Language Model Agents MOOC](https://llmagents-learning.org/f24), providing additional labs, certifications, and interactive communities for those unable to attend the formal course. +- **Lecture Videos**: All 12 lectures are recorded on YouTube. The schedule and links are as follows: + +| Date | Topic | Guest Lecturer | Video Link | +|------------|-------------------------------------------------|---------------------------------------|-------------------------------------------------| +| Sept 9 | LLM Reasoning | Denny Zhou, Google DeepMind | [Video](https://www.youtube.com/live/QL-FS_Zcmyo) | +| Sept 16 | LLM Agents: History and Overview | Shunyu Yao, OpenAI | [Video](https://www.youtube.com/watch?v=RM6ZArd2nVc) | +| Sept 23 | Agent AI Framework & AutoGen, Building Multimodal Knowledge Assistants | Chi Wang, AutoGen-AI; Jerry Liu, LlamaIndex | [Video](https://www.youtube.com/live/OOdtmCMSOo4) | +| Sept 30 | Generative AI Trends in Business and Key Components for Building Successful Agents/Applications | Burak Gokturk, Google | [Video](https://www.youtube.com/live/Sy1psHS3w3I) | +| Oct 7 | Composite AI Systems & DSPy Framework | Omar Khattab, Databricks | [Video](https://www.youtube.com/live/JEMYuzrKLUw) | +| Oct 14 | Software Development Agents | Graham Neubig, Carnegie Mellon University | [Video](https://www.youtube.com/live/f9L9Fkq-8K4) | +| Oct 21 | Enterprise Workflow AI Agents | Nicolas Chapados, ServiceNow | [Video](https://www.youtube.com/live/-yf-e-9FvOc) | +| Oct 28 | A Unified Framework for Neuro-Symbolic Decision-Making | Yuandong Tian, Meta AI (FAIR) | [Video](https://www.youtube.com/live/wm9-7VBpdEo) | +| Nov 4 | Project GR00T: The Universal Robotics Blueprint | Jim Fan, NVIDIA | [Video](https://www.youtube.com/live/Qhxr0uVT2zs) | +| Nov 18 | Open-Source and Science in the Era of Foundation Models | Percy Liang, Stanford University | [Video](https://www.youtube.com/live/f3KKx9LWntQ) | +| Nov 25 | Measuring Agent Abilities and Anthropic's RSP | Ben Mann, Anthropic | [Video](https://www.youtube.com/live/6y2AnWol7oo) | +| Dec 2 | Building Safe and Trustworthy AI Agents, and AI Policies Based on Science and Evidence | Dawn Song, UC Berkeley | [Video](https://www.youtube.com/live/QAgR4uQ15rc) | + +Each lecture has corresponding reading materials; see the course website [CS294/194-196 Course Homepage](http://rdi.berkeley.edu/llm-agents/f24#syllabus). + +- **Course Notes**: [GitHub](https://github.com/rajdeepmondaldotcom/CS294_LLM_Agents_Notes_Fall2024) + +## Relevant Papers and Resources + +The course focuses on the following papers and frameworks, categorized by technical direction: + +### Reasoning and Planning + +- **ReAct**: A framework combining reasoning and action to enhance task-solving capabilities. Paper: [ReAct: Synergizing Reasoning and Acting in Language Models](https://arxiv.org/abs/2210.03629). +- **Chain-of-Thought**: Using intermediate steps to stimulate reasoning in models. Paper: [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://arxiv.org/abs/2201.11903). +- **Chain-of-Thought Reasoning Without Prompting**: Exploring reasoning through chain-of-thought without prompting. Paper: [Chain-of-Thought Reasoning Without Prompting](https://arxiv.org/abs/2402.10200). + +### Agent Frameworks + +- **AutoGen**: A framework for multi-agent conversation development. Paper: [AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation](https://arxiv.org/abs/2308.08155). +- **DSPy**: A programming framework for composite AI systems. Paper: [DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines](https://arxiv.org/abs/2310.03714). + +### Application Scenarios + +- **Code Generation**: + - **SWE-agent**: An agent interface for automating software engineering. Paper: [SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering](https://arxiv.org/abs/2402.01030). +- **Robotics**: + - **Voyager**: An embodied agent based on LLMs. Paper: [Voyager: An Open-Ended Embodied Agent with Large Language Models](https://arxiv.org/abs/2305.16291). + +## Extended Learning Suggestions + +For further study, we recommend: + +- **Practical Projects**: Participate in the [LLM Agents Hackathon](http://rdi.berkeley.edu/llm-agents-hackathon/), aiming to build agents supporting tool usage (e.g., web automation). The Hackathon provides tracks such as applications, benchmarks, and foundational research. + +- **Advanced Courses**: Explore the spring course [CS294/194-280 Advanced Large Language Model Agents](http://rdi.berkeley.edu/adv-llm-agents/sp25), focusing on advanced topics like LLM reasoning, mathematical proofs, and code generation. +- **Online Resources**: + - [Large Language Model Agents MOOC](https://llmagents-learning.org/f24): Offers course materials, labs, and certifications, ideal for self-learners. + - [Awesome LLM Agents](https://github.com/kaushikb11/awesome-llm-agents): A collection of frameworks, papers, and projects on LLM agents, ideal for deep research. \ No newline at end of file diff --git a/docs/大语言模型与智能体/llm_zero_to_hero.en.md b/docs/大语言模型与智能体/llm_zero_to_hero.en.md new file mode 100644 index 00000000..e69de29b