本文前置知识: CNN: 详见卷积神经网络小结. 本文着重介绍深度可分离卷积和分组卷积两种操作. 深度可分离卷积深度可分离卷积(Depthwise Separable Convolution)应用在MobileNet和Xceptio
2020-11-26
Pytorch实现: Transformer本文前置知识: Pytorch基本操作 Transformer: 详见Transformer精讲 2022.04.03: 去掉了Pre Norm比Post Norm效果好的表述. Pytorch实现: Transformer本文是T
2020-11-23
KEPLER: Knowledge Embedding and Pre-trained Language Representation本文前置知识: BERT(详见ELMo, GPT, BERT) KEPLER: A Unified Model for Knowledge Embedding and Pre-trained Language Representat
2020-11-21
KGE预警论文两则本文是两篇KGE方向的预警论文的阅读笔记和个人理解. 预警类的工作其实是比较少见的, 对领域的发展也非常有指导意义. 2020.11.22: 更新Reciprocal Relation. 2021.05.13: 修正Reciprocal
2020-11-20
Pytorch实现: Skip-Gram本文前置知识: Pytorch基本操作 Word2Vec Pytorch实现: Skip-Gram本文用Pytorch实现了Skip - Gram, 它是Word2Vec的其中一种. 本文实现参考PyTorch 实现 Word2Ve
2020-11-19
RoBERTa: A Robustly Optimized BERT Pretraining Approach本文前置知识: BERT(详见ELMo, GPT, BERT) RoBERTa: A Robustly Optimized BERT Pretraining Approach本文是论文RoBERTa: A Robustly Opti
2020-11-18
Integrating Image-Based and Knowledge-Based Representation Learning本文前置知识: AlexNet(详见卷积神经网络发展史) Attention(详见Seq2Seq和Attention) TransE(详见TransE: Translating Embeddings for Modeling Multi
2020-11-13
TransE: Translating Embeddings for Modeling Multi-relational DataTransE: Translating Embeddings for Modeling Multi-relational Data本文是论文Translating Embeddings for Modeling Multi-relation
2020-11-13
CoKE: Contextualized Knowledge Graph Embedding本文前置知识: Self - Attention BERT 2020.11.17: 解决了标签泄露的疑惑. 2021.04.09: 修正可视化实验的描述. CoKE: Contextualized Knowledge Graph E
2020-11-11
ConvKB: A Novel Embedding Model for KB Completion Based on CNN本文前置知识: ConvE Conv1d 2021.03.15: 指出权重共享并没有出现在源码中. A Novel Embedding Model for Knowledge Base Completion Based on Con
2020-11-07
InteractE: Improving Convolution-based KGE by Increasing Feature Interactions本文前置知识: ConvE Depth - wise Convolution 2020.11.14: 对实验进行部分补充. InteractE: Improving Convolution-based Knowledge Graph
2020-11-06
CoLAKE: Contextualized Language and Knowledge EmbeddingCoLAKE: Contextualized Language and Knowledge Embedding 本文前置知识: BERT Self - Attention 2020.11.11: 想通了CoLAKE在训练时最关键的部分.
2020-10-29