诸神缄默不语-个人技术博文与视频目录legalAI文本摘要算文本摘要legalAI其他算legalAI。Re后面的顺序是我写笔记的顺序这个是我之前写的顺序现在我决定慢慢把这个序号给删掉了论文本身的发表顺序标在论文标题前面了。文章目录GNN通用同质图节点表征通用异质图节点表征同质图链路预测基础模型RNNLLMGPT系Claude系DeepSeek系Qwen系继续预训练部分微调prompt位置编码LM中蕴含的知识RAGagentAI编程文本摘要LegalAILJP分类文本相似性案例匹配案例检索/推荐公平性事件检测信息抽取其他信息抽取意图识别槽填充数值推理MWP向量压缩向量量化交叉学科情报学GNN通用同质图节点表征(2018 ICLR) Re37读论文 G2G Graph2Gauss Deep Gaussian Embedding of Graphs: Unsupervised Inductive Learning via Rank(2019 ICLR) Re0读论文 PPNP/APPNP Predict then Propagate: Graph Neural Networks meet Personalized PageRank端到端先transform再propagate(2019 KDD) Re3读论文 PGE A Representation Learning Framework for Property Graphs(2020 KDD) Re46读论文 DAGNN Towards Deeper Graph Neural Networks(2020 ICLR) Re2: 读论文 CS-GNN Measuring and Improving the Use of Graph Information in Graph Neural Networks(2021 ICLR) Re1读论文 CS (Correct and Smooth) Combining Label Propagation and Simple Models Out-performs Graph Ne解耦transform和propagate再加一个correct通用异质图节点表征(2017 KDD) Re31读论文 metapath2vec: Scalable Representation Learning for Heterogeneous Networks异质图版的node2vec(2020 AAAI 滴滴北大) Re22读论文 HetSANN An Attention-based Graph Neural Network for Heterogeneous Structural Learning(2021 KDD 清华) Re10读论文 Are we really making much progress? Revisiting, benchmarking, and refining heterogeneous gr喷了一圈各种HGNN算法最后提出了一个简单的HGNN模型然后发现这个新模型表现最好了同质图链路预测(2020 IJCAI) 论文阅读笔记DEAL_inductive链路预测_分别表征节点特征和拓扑结构对比学习对齐拓扑表征和特征表征分开建模论文全名Inductive Link Prediction for Nodes Having Only Attribute Information基础模型RNN(2014 NIPS) Re71读论文 Sequence to Sequence Learning with Neural NetworksLLM(2018 ACL) Re73 读论文ULMFiT Universal Language Model Fine-tuning for Text Classification(2019 NAACL) Re63读论文 BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding(2019 NeurIPS) Re72读论文 XLM Cross-lingual Language Model Pretraining(2020 EMNLP) Re55读论文 Entities as Experts: Sparse Memory Access with Entity Supervision将实体表征结合到LLM中(2020 JMLR) Re70读论文 T5 Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer(2022 谷歌) Re69读论文 LaMDA: Language Models for Dialog Applications(2022 ICLR 谷歌) Re68读论文 instruction tuning FLAN Finetuned Language Models Are Zero-Shot Learners(2023 Meta) Re75 读论文Toolformer: Language Models Can Teach Themselves to Use ToolsGPT系(2018 OpenAI) GPT-1论文阅读笔记_Improving Language Understanding by Generative Pre-Training(2019 OpenAI) GPT-2论文阅读笔记_Language Models are Unsupervised Multitask Learners(2020 NeurIPS OpenAI) Re65读论文 GPT-3 Language Models are Few-Shot Learners(2023) Re78 读论文GPT-4 Technical ReportClaude系论文阅读笔记Claude如何思考(2025) On the Biology of a Large Language Model(2025) Circuit Tracing: Revealing Computational Graphs in Language ModelsDeepSeek系(2025) Re 80 读论文DeepSeek-V32025年初最强大模型几天前还是的(2025) Re 83 读论文DeepSeek-R12025年初最强大模型的推理时候Qwen系(2025) Re 82读论文qwen 3继续预训练(2020 ACL) Re26读论文 Don’t Stop Pretraining: Adapt Language Models to Domains and Tasks继续预训练能提升下游任务表现效果部分微调Re77 读论文LoRA: Low-Rank Adaptation of Large Language Modelsprompt(2022 ACM Computing Surveys 卡耐基梅隆大学) Re33读论文 Pre-train, Prompt, and Predict: A Systematic Survey of Prompting Methods in Natural Languagprompt综述(2023 EMNLP) Re 84 读论文TELeR: A General Taxonomy of LLM Prompts for Benchmarking Complex Tasksprompt分类位置编码(2021 苏剑林) Re 79 读论文RoPE RoFormer: Enhanced Transformer with Rotary Position EmbeddingLM中蕴含的知识(2019 EMNLP) Re51读论文 Language Models as Knowledge Bases?完形填空(2020 EMNLP) Re52读论文 How Much Knowledge Can You Pack Into the Parameters of a Language Model?QA(2020 TACL) Re53读论文 How Can We Know What Language Models Know?完形填空但是template是通过学习得到的(2023 EMNLP) Re67读论文 Don‘t Trust ChatGPT when Your Question is not in English: A Study of Multilingual AbilitiesRAG叠实体表征那几个真的能算检索增强吗不过既然别人这么分类我就姑且这么算了(2020 ICLR 斯坦福Facebook) Re48读论文 kNN-LMs Generalization through Memorization: Nearest Neighbor Language Models在语言模型计算出的token概率的基础上增加kNN token概率kNN概率的计算首先构建海量文本向量数据库key是上下文表征value是target token。对每个测试样本用FAISS检索得到k个最近的样本其target token的概率与向量距离成反比向量越近概率越大(2020 ICML 谷歌) Re58读论文 REALM: Retrieval-Augmented Language Model Pre-Training从海量维百中检索相关文本然后加到输入文本后面。这个检索器是端到端预训练的(2020 NeurIPS Facebook) Re59读论文 Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks(2020 AKBC) Re54读论文 How Context Affects Language Models‘ Factual Predictions对比了不同检索方案。其实做得挺单薄的也就是拿TF-IDF检索维基百科的检索器加上跟监督算法比较了一下。关于seperation的实验比较有参考价值(2021 NAACL 谷歌) Re60读论文 FILM Adaptable and Interpretable Neural Memory Over Symbolic Knowledge这篇也是叠实体表征(2022 ICLR 谷歌) Re57读论文 Mention Memory: incorporating textual knowledge into Transformers through entity mention atmention memory构建语料中的mention表征向量TOME模型在实现下游任务时结合实体mention表征向量做sparse attention(2023 ACL) Re49读论文 When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parame自适应选择少见实体加检索(2023 ICML) Re50读论文 Large Language Models Struggle to Learn Long-Tail Knowledge检索增强解决LM搞不动长尾实体的问题(2024 谷歌) Re74 读论文DataGemma Knowing When to Ask - Bridging Large Language Models and DataagentRe 81 读论文AlphaEvolve: A Gemini-powered coding agent for designing advanced algorithmsAI编程论文阅读笔记AI编程用的AGENTS.md应该不写或少写论文(2026) Evaluating AGENTS.md: Are Repository-Level Context Files Helpful for Coding Agents?文本摘要(2017 ACL) Re47读论文PGN/Pointer-Generator Netwoks Get To The PointSummarization with Pointer-Generator Networks生成概率指针概率抽取(2018 EMNLP) Re66读论文 Bottom-Up Abstractive Summarization(2019 EMNLP) Re64读论文 BertSum Text Summarization with Pretrained Encoders Fine-tune BERT for Extractive Summari(2021 AAAI) Re4读论文 CGSum: Enhancing Scientific Papers Summarization with Citation Graph结合文献引用关系实现文献摘要(2021 ACL 清华) Re5读论文 TWAG: A Topic-guided Wikipedia Abstract Generator结合维基百科的小标题生成摘要(2022 AAAI) Re12读论文 Se3 Semantic Self-segmentation for Abstractive Summarization of Long Legal Documents in Low分治生成摘要(2022 SIGIR) Re32读论文 Summarizing Legal Regulatory Documents using Transformers(2022 COLING 匹兹堡大学) Re35读论文 ArgLegalSumm: Improving Abstractive Summarization of Legal Documents with Argument Mining识别出argumen然后生成摘要LegalAI(2020 ACL) Re23读论文 How Does NLP Benefit Legal System: A Summary of Legal Artificial Intelligence综述LJP(2017 EMNLP 北大) Re7读论文 FLA/MLAC/FactLaw Learning to Predict Charges for Criminal Cases with Legal Basis结合法条预测罪名(2019 Law in Context) Re56读论文 A Brief History of the Changing Roles of Case Prediction in AI and Law美国LJP传统方法综述(2020 ACL) Re27读论文 LADAN Distinguish Confusing Law Articles for Legal Judgment Prediction结合法条相似关系(2021 ACL) Re16读论文 ILDC for CJPE: Indian Legal Documents Corpus for Court Judgment Prediction and Explanation(2021 NAACL 北大) Re18读论文 GCI Everything Has a Cause: Leveraging Causal Inference in Legal Text Analysis(2021 SIGIR 北大阿里) Re21读论文 MSJudge Legal Judgment Prediction with Multi-Stage Case Representation Learning in the Real(2021 SIGIR) Re38读论文 NeurJudge: A Circumstance-aware Neural Framework for Legal Judgment Prediction结合犯罪情节(2022 AAAI) 论文阅读笔记LeSICiN_以inductive链路预测范式解决多标签文本分类任务法条预测结合案例引用和法条层级异质图用链路预测范式做法条预测任务论文全名LeSICiN: A Heterogeneous Graph-based Approach for Automatic Legal Statute Identification from Indian Legal Documents(2022 AAAI) Re14读论文 ILLSI Interpretable Low-Resource Legal Decision Making(2022 ACL 南大) Re11读论文 EPM Legal Judgment Prediction via Event Extraction with Constraints结合事件抽取(2022 IJCAI 西电) Re28读论文 CECP Charge Prediction by Constitutive Elements Matching of Crimes结合犯罪要素强化学习(2022 IPM) Re36读论文 CEEN Improving legal judgment prediction through reinforced criminal element extraction结合犯罪要素强化学习(2022 COLING) Re 39读论文 CTM Augmenting Legal Judgment Prediction with Contrastive Case Relations结合案例标签相似关系和频率(2022 Artificial Intelligence and Law) Re41NumLJP Judicial knowledge‑enhanced magnitude‑aware reasoning for numerical legal judgment predi结合数值信息分类(2022 NAACL) Re29读论文 D2GCLF: Document-to-Graph Classifier for Legal Document Classification文本相似性(2020 SIGIR) Re8读论文 Hier-SPCNet: A Legal Statute Hierarchy-based Heterogeneous Network for Computing Legal Case结合案例引用和法条层级异质图案例匹配(2022 SIGIR 人大华为) Re24读论文 IOT-Match Explainable Legal Case Matching via Inverse Optimal Transport-based Rationale Ext案例检索/推荐(2022 SIGIR) Re25读论文 LecutJOTR Incorporating Retrieval Information into the Truncation of Ranking Lists in the(2022 ACM Transactions on Information Systems 清华IBM) Re30读论文 LegalGNN: Legal Information Enhanced Graph Neural Network for Recommendation公平性(2022 AAAI) Re13读论文 Gender and Racial Stereotype Detection in Legal Opinion Word Embeddings事件检测(2022 ACL 清华) Re15读论文 LEVEN: A Large-Scale Chinese Legal Event Detection Dataset信息抽取(2021 ACL) Re17读论文 Challenges for Information Extraction from Dialogue in Criminal Law(2021 NAACL) Re19读论文 Paragraph-level Rationale Extraction through Regularization: A case study on European Court其他(2021 NAACL 剑桥) Re20读论文 What About the Precedent: An Information-Theoretic Analysis of Common Law(2022 SIGIR) Re34读论文 Organizing Portuguese Legal Documents through Topic Discovery信息抽取意图识别槽填充(2021 ACL) Re 40读论文 GL-GIN: Fast and Accurate Non-Autoregressive Model for Joint Multiple Intent Detection and数值推理MWP(2014 EMNLP) Re42读论文 ARIS Learning to Solve Arithmetic Word Problems with Verb Categorization识别和分类动词(2017 EMNLP) Re43读论文 DNS Deep Neural Solver for Math Word ProblemsRNN检索(2021 OpenAI) Re44数据集 GSM8K 和 论文 Training Verifiers to Solve Math Word Problemsverifier(2024 AAAI) Re61读论文 PRP Get an A in Math: Progressive Rectification Prompting向量压缩向量量化论文阅读笔记TurboQuant_谷歌整的新活正在攻击你的内存板块美股还有造假和抄袭嫌疑(2026 ICLR 谷歌) TurboQuant: Online Vector Quantization with Near-optimal Distortion Rate涉及争议(2024 SIGMOD) RaBitQ: Quantizing High-Dimensional Vectors with a Theoretical Error Bound for Approximate Nearest Neighbor Search交叉学科情报学(2022 图书情报工作) Re76 读论文新兴技术的多指标量化识别研究——基于向量表征方法的探索