Coding

Reading

future

graph/network 为辅，LLM-agent为主;

也可以考虑多模态信息；

社交网络认知建模和预测

alignment, reasoning, graph-structure, simulation,

basic

大模型基础教材（赵鑫）

DeepSeek Tech Reports

reasoning

DeepSeek R1 report

Sys1 to Sys2: Survey

alignment

From PG to GRPO