自定义博客皮肤VIP专享

*博客头图:

格式为PNG、JPG,宽度*高度大于1920*100像素,不超过2MB,主视觉建议放在右侧,请参照线上博客头图

请上传大于1920*100像素的图片!

博客底图:

图片格式为PNG、JPG,不超过1MB,可上下左右平铺至整个背景

栏目图:

图片格式为PNG、JPG,图片宽度*高度为300*38像素,不超过0.5MB

主标题颜色:

RGB颜色,例如:#AFAFAF

Hover:

RGB颜色,例如:#AFAFAF

副标题颜色:

RGB颜色,例如:#AFAFAF

自定义博客皮肤

-+

John159151的专栏

Things worth sharing

原创 #Paper Reading# A Survey on Deep Hashing Methods

论文题目: A Survey on Deep Hashing Methods论文地址: https://arxiv.org/abs/2003.03369论文发表于: arXiv 2020论文大体内容:本文主要对使用深度学习的哈希方法进行了较为全面综述,是一篇总结概览近几年来学术界提出的深度哈希方法的文章。Motivation:深度哈希是近几年来的热方向,大家都将传统的哈希方法往深度学习方向引入,以获取更好的end2end哈希表示。Contribution:本文对深度哈希的综述,方便对这

2020-07-05 12:03:52 47

原创 #Paper Reading# Implicit Neural Representations with Periodic Activation Functions

论文题目: Implicit Neural Representations with Periodic Activation Functions论文地址: https://arxiv.org/abs/2006.09661论文发表于: arXiv 2020论文大体内容:本文主要提出使用sin来做激活函数,并且在特定任务上比relu好。Motivation:如何更好的处理复杂信号(如音频、图像、3D)是一个问题。Contribution:①本文提出使用周期激活函数来处理复杂信号(音频、图像

2020-06-24 19:22:02 395

原创 #Paper Reading# AUCµ: A Performance Metric for Multi-Class Machine Learning Models

论文题目: AUCµ: A Performance Metric for Multi-Class Machine Learning Models论文地址: http://proceedings.mlr.press/v97/kleiman19a.html论文发表于: PLMR 2019论文大体内容:本文主要提出了在多分类下的AUC计算方法——AUCµ,具有与AUC同样的特性。Motivation:二分类问题往往会使用AUC进行eval,而多分类上往往会使用softmax。本文提出了AUCµ是为

2020-06-19 19:50:03 77

原创 #Paper Reading# TabNet: Attentive Interpretable Tabular Learning

论文题目: TabNet: Attentive Interpretable Tabular Learning论文地址: https://arxiv.org/abs/1908.07442论文发表于: arXiv 2019论文大体内容:本文主要提出了TabNet模型,能够高效地在tabular数据上完成分类/回归的任务,且具可解释性。本文提出的模型是用DNN的方式获得树模型的可解释性,且超越树模型的效果。Motivation:tabular数据一般都使用树模型去处理,怎么用DNN去实现树模型的效

2020-06-11 11:54:01 227

原创 #Paper Reading# Learnable pooling with Context Gating for video classification

论文题目: Learnable pooling with Context Gating for video classification论文地址: https://arxiv.org/abs/1706.06905论文发表于: arXiv论文大体内容:本文主要提出了Gated NetVLAD模型,用于给视频打标,并在kaggle比赛中取得第一名。Motivation:如何更有效的对视...

2020-04-29 01:58:36 88

原创 #Paper Reading# PyTorch-BigGraph: A Large-scale Graph Embedding Framework

论文题目: PyTorch-BigGraph: A Large-scale Graph Embedding Framework论文地址: https://arxiv.org/abs/1903.12287论文发表于: SysML 2019论文大体内容:本文主要实现了大规模数据下的graph embedding,在效果微涨的情况下,训练性能有显著的提升,对于工业界大规模图网络的embeddi...

2020-04-21 15:57:05 143 1

原创 #Paper Reading# EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks

论文题目: EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks论文地址: https://arxiv.org/abs/1905.11946论文发表于: ICML 2019论文大体内容:本文主要提出了EfficientNet模型,来同时考虑网络的deep、width、resolution(分辨率)...

2020-02-23 17:57:24 106

原创 #Paper Reading# Class-Balanced Loss Based on Effective Number of Samples

论文题目: Class-Balanced Loss Based on Effective Number of Samples论文地址: http://openaccess.thecvf.com/content_CVPR_2019/html/Cui_Class-Balanced_Loss_Based_on_Effective_Number_of_Samples_CVPR_2019_paper.ht...

2020-02-20 17:50:59 264

原创 #Paper Reading# On the Measure of Intelligence

论文题目: On the Measure of Intelligence论文地址: https://arxiv.org/pdf/1911.01547.pdf论文发表于: arxiv论文大体内容:这篇论文比较有趣,作者主要讨论了怎么测量AI,什么才是AI。然后作者提出了ARC数据集(Abstraction and Reasoning Corpus),用于更好的测量AI的能力。Motiv...

2020-02-11 13:43:11 126

原创 新的abtest方法: Innovating Faster on Personalization Algorithms at Netflix Using Interleaving

1. 我们在线上做实验的时候,常常会用到abtest,来确定线上效果。而我们的abtest方法,往往就是对用户进行分组,对于相同比例的随机用户,进行效果比较。这种方法有个前提/假定,就是各组随机用户的分布是完全一样的。当然,一般情况下(如用户量特别大),这种方法大概率情况是不太有问题的,但是对于某些场景,如Netflix推荐场景,这种方法会有所局限;2. 在Netflix的推荐场景中,较少部分...

2019-12-30 01:10:21 183

原创 #Paper Reading# DeepGBM: A Deep Learning Framework Distilled by GBDT for Online Prediction Tasks

论文题目: DeepGBM: A Deep Learning Framework Distilled by GBDT for Online Prediction Tasks论文地址: https://dl.acm.org/citation.cfm?id=3330858论文发表于: KDD 2019论文大体内容:本文主要提出了DeepGBM模型,来整合NN与GBDT的优势,支持sparse...

2019-12-21 12:22:29 248

原创 #Paper Reading# Deep Learning Recommendation Model for Personalization and Recommendation Systems

论文题目: Deep Learning Recommendation Model for Personalization and Recommendation Systems论文地址: https://arxiv.org/abs/1906.00091论文发表于: arxiv 2019论文大体内容:本文主要提出了deep learning recommendation model(DLRM...

2019-11-05 20:55:10 169

原创 #Paper Reading# Objects as Points

论文题目: Objects as Points论文地址: https://arxiv.org/abs/1904.07850论文发表于: arxiv 2019论文大体内容:本文主要提出了CenterNet的方法,来解决object detector中效率低的问题。CenterNet不但能够端到端训练,还能在效率和精度上做trade-off,并且在COCO数据集[1]上取得state-of-...

2019-10-10 13:20:10 61

原创 #Paper Reading#Modeling Task Relationships in Multi-task Learning with Multi-gate Mixture-of-Experts

论文题目: Modeling Task Relationships in Multi-task Learning with Multi-gate Mixture-of-Experts论文地址: https://dl.acm.org/citation.cfm?id=3220007论文发表于: KDD 2018(CCF A类会议)论文大体内容:本文主要提出了Multi-gate Mixtur...

2019-10-09 15:59:15 229

原创 #Paper Reading# Gradient Harmonized Single-stage Detector

论文题目: Gradient Harmonized Single-stage Detector论文地址: https://www.aaai.org/ojs/index.php/AAAI/article/view/4877论文发表于: AAAI 2019(CCF A类会议)论文大体内容:本文主要提出了gradient harmonizing mechanism (GHM)的方法,来解决on...

2019-09-27 15:09:21 124

原创 #Paper Reading# Focal Loss for Dense Object Detection

论文题目: Focal Loss for Dense Object Detection论文地址: http://openaccess.thecvf.com/content_iccv_2017/html/Lin_Focal_Loss_for_ICCV_2017_paper.html论文发表于: ICCV 2017(CCF A类会议)论文大体内容:本文主要提出了Focal Loss,用于解决...

2019-09-24 19:32:06 178

原创 #Paper Reading# RSLIME: An Efficient Feature Importance Analysis Approach for Industrial RS

论文题目: RSLIME: An Efficient Feature Importance Analysis Approach for Industrial Recommendation Systems论文地址: 暂无论文发表于: IJCNN 2019(CCF C类会议)论文大体内容:本文主要介绍了爱奇艺团队用一个类似于LIME[1]的算法——RSLIME(Recommen- datio...

2019-07-01 01:18:18 258

原创 #Paper Reading# "Why Should I Trust You?" Explaining the Predictions of Any Classifier

论文题目: "Why Should I Trust You?" Explaining the Predictions of Any Classifier论文地址: https://dl.acm.org/citation.cfm?Id=2939778论文发表于: KDD 2016(CCF A类会议)论文大体内容:本文主要提出了LIME(Local Interpretable Model-a...

2019-07-01 00:04:53 930

原创 #Paper Reading# Seq2Slate: Re-ranking and Slate Optimization with RNNs

论文题目: Seq2Slate: Re-ranking and Slate Optimization with RNNs论文地址: https://negative-dependence-in-ml-workshop.lids.mit.edu/wp-content/uploads/sites/29/2019/06/seq2slate_icml_workshop.pdf论文发表于: ICML 2...

2019-06-26 15:15:31 479

原创 #Paper Reading# Learning a Deep Listwise Context Model for Ranking Refinement

论文题目: Learning a Deep Listwise Context Model for Ranking Refinement论文地址: https://dl.acm.org/citation.cfm?id=3209985论文发表于: SIGIR 2018(CCF A类会议)论文大体内容:本文主要提出了一个Rerank模型——Deep Listwise Context Model...

2019-06-20 20:15:30 1168

原创 #Paper Reading# Deep Interest Network for Click-Through Rate Prediction

论文题目: Deep Interest Network for Click-Through Rate Prediction论文地址: https://dl.acm.org/citation.cfm?id=3219823论文发表于: KDD 2018(CCF A类会议)论文大体内容:本文主要介绍了阿里妈妈团队用于CTR任务的DIN(Deep Interest Network)模型,该模型通...

2019-06-10 15:02:46 204

原创 #Paper Reading# Personalized Context-aware Re-ranking for E-commerce Recommender Systems

论文题目: Personalized Context-aware Re-ranking for E-commerce Recommender Systems论文地址: https://arxiv.org/abs/1904.06813论文发表于: arxiv,2019.04论文大体内容:本文主要提出了PCRM(Personalized Context-aware Re-ranking Mo...

2019-06-04 17:02:50 597

原创 #Paper Reading# Wide & Deep Learning for Recommender Systems

论文题目: Wide & Deep Learning for Recommender Systems论文地址: https://dl.acm.org/citation.cfm?id=2988454论文发表于: DLRS 2016(RecSys 2016的Workshop,CORE B类会议)论文大体内容:本文主要提出了WDL(Wide & Deep Learning)模型...

2019-06-03 14:42:41 190

原创 #Paper Reading# xDeepFM:Combining Explicit and Implicit Feature Interactions for Recommender Systems

论文题目: xDeepFM: Combining Explicit and Implicit Feature Interactions for Recommender Systems论文地址: https://dl.acm.org/citation.cfm?id=3220023论文发表于: KDD 2018(CCF A类会议)论文大体内容:本文主要介绍了DeepFM模型的变种——xDee...

2019-06-02 15:34:12 289

原创 #Paper Reading# DeepFM: a factorization-machine based neural network for CTR prediction

论文题目: DeepFM: a factorization-machine based neural network for CTR prediction论文地址: https://www.ijcai.org/proceedings/2017/0239.pdf论文发表于: IJCAI 2017(CCF A类会议)论文大体内容:本文主要介绍了一种融合FM和DNN的模型——DeepFM模型,...

2019-05-31 16:58:11 232

原创 #Paper Reading# RippleNet: Propagating User Preferences on the KG for Recommender Systems

论文题目: RippleNet: Propagating User Preferences on the Knowledge Graph for Recommender Systems论文地址: https://dl.acm.org/citation.cfm?id=3271739论文发表于: CIKM 2018(CCF B类会议)论文大体内容:本文主要介绍了一种通过引入Knowledge...

2019-05-31 11:55:38 768

原创 #Paper Reading# Field-aware factorization machines for CTR prediction

论文题目: Field-aware factorization machines for CTR prediction论文地址: https://dl.acm.org/citation.cfm?id=2959134论文发表于: RecSys 2016(CORE B类会议)论文大体内容:本文主要介绍了FM模型的变种——FFM(Field-aware Factorization Machin...

2019-05-29 18:38:46 164

原创 #Paper Reading# Factorization Machines

论文题目: Factorization Machines论文地址: https://ieeexplore.ieee.org/abstract/document/5694074论文发表于: ICDM 2010(CCF B类会议)论文大体内容:本文主要介绍了FM模型,该模型相比与SVD++,PITF,FPMC等有明显的优势,同时FM模型也能学习出这几个模型的效果,是一种高效、通用、不惧稀疏的...

2019-05-28 19:38:14 116

原创 #Paper Reading# Ad Click Prediction: a View from the Trenches

论文题目: Ad Click Prediction: a View from the Trenches论文地址: https://dl.acm.org/citation.cfm?id=2488200论文发表于: KDD 2013(CCF A类会议)论文大体内容:本文主要介绍了FTRL模型的工程实现,并且作者尝试的多个trick。这也是LR模型处理大规模数据下,所发展出来的优化模型FTRL...

2019-05-28 11:57:55 124

原创 #Paper Reading# Deep Neural Networks for YouTube Recommendations

论文题目: Deep Neural Networks for YouTube Recommendations论文地址: https://dl.acm.org/citation.cfm?id=2959190论文发表于: RecSys 2016(CORE B类会议)论文大体内容:本文主要介绍了DNN在YouTube视频推荐上的应用,这也是DNN模型在视频推荐领域的开山之作,后续所有的视频推荐...

2019-05-27 11:52:20 156

原创 #Paper Reading# The YouTube Video Recommendation System

论文题目: The YouTube Video Recommendation System论文地址: https://dl.acm.org/citation.cfm?id=1864770论文发表于: RecSys 2010(CORE B类会议)论文大体内容:本文简单介绍了YouTube的推荐系统,从整个推荐架构层面来描述。1. YouTube主要在首页以及播放页场景做推荐。2. ...

2019-05-24 17:34:00 75

原创 #Paper Reading# ESMM: An Effective Approach for Estimating Post-Click Conversion Rate

论文题目: Entire Space Multi-Task Model: An Effective Approach for Estimating Post-Click Conversion Rate论文地址: https://dl.acm.org/citation.cfm?id=3210104论文发表于: SIGIR 2018(CCF A类会议)论文大体内容:本文主要使用Multi-t...

2019-04-10 12:50:21 121

原创 #Paper Reading# A Unified Search Federation System Based on Online User Feedback

论文题目: A Unified Search Federation System Based on Online User Feedback论文地址: https://dl.acm.org/citation.cfm?id=2488198论文发表于: KDD 2013(CCF A类会议)论文大体内容:对于搜索/推荐系统来说,面临一个挑战,就是怎么样才能更合理的对多路召回的结果进行merge...

2019-01-16 12:07:08 184

原创 #Paper Reading# A Survey and Empirical Analysis of Beyond-Accuracy Objectives in Recommender Systems

论文题目: Diversity, Serendipity, Novelty, and Coverage: A Survey and Empirical Analysis of Beyond-Accuracy Objectives in Recommender Systems论文地址: https://dl.acm.org/citation.cfm?id=2926720论文发表于: TIIS 2...

2019-01-14 16:43:26 104

原创 #Paper Reading# Modeling Impression Discounting in Large-scale Recommender Systems

论文题目: Modeling Impression Discounting in Large-scale Recommender Systems论文地址: https://dl.acm.org/citation.cfm?id=2623356论文发表于: KDD 2014(CCF A类会议)论文大体内容:本文主要解决了怎么去衡量、量化推荐系统中的隐式负反馈问题,并实验了4种衰减方式:①线性...

2018-12-26 18:41:51 204

原创 #Paper Reading# Learning to Selectively Transfer:Reinforced Transfer Learning for Deep Text Matching

论文题目:Learning to Selectively Transfer: Reinforced Transfer Learning for Deep Text Matching论文地址:https://doi.org/10.1145/3289600.3290978论文发表于:WSDM 2019(CCF B类会议)论文大体内容:本文主要提出了一个RTL(Reinforced Trans...

2018-12-12 23:59:43 324

原创 #Paper Reading# Towards Enabling Binary Decomposition for Partial Label Learning

论文题目:Towards Enabling Binary Decomposition for Partial Label Learning论文地址:http://cse.seu.edu.cn/people/zhangml/files/IJCAI'18.pdf论文发表于:IJCAI 2018(CCF A类会议)论文大体内容:本文主要采用两层分类器来解决偏标记学习问题,通过one-vs-on...

2018-12-08 20:13:26 433

原创 #Paper Reading# Joint Matrix Factorization and Manifold-Ranking for Topic-Focused Multi-Document Sum

论文大体内容:本文将矩阵分解与流形排序(Manifold-ranking)组合起来,得到JMFMR的新模型,用于多文档摘要(extractive式),通过实验发现取得了不错的效果。

2017-07-01 18:48:43 547

原创 #Paper Reading# Manifold-Ranking Based Topic-Focused Multi-Document Summarization

论文大体内容:本文将流形排序(Manifold-ranking)应用到多文档摘要(extractive式)中,通过实验发现取得了不错的效果。

2017-07-01 17:58:49 923 1

原创 #Paper Reading# Abstractive Sentence Summarization with Attentive Recurrent Neural Networks

论文大体内容:本文使用一种conditional RNN来生成摘要,条件是卷积注意力模型(convolutional attention-based encoder),用来确保每一步生成词的时候都可以聚焦到合适的输入上。模型仅仅依赖于学习到的features,并且很容易在大规模数据上进行end2end式地训练,并且在Gigaword语料上和DUC-2004任务中取得了更好的效果。

2017-06-11 09:19:03 1370

提示
确定要删除当前文章?
取消 删除