• Machine Learning – Basic points


    Machine learning is the translation of English machine learning. It mainly studies how to make computers simulate or realize human behavior. Just like a student, it can acquire new knowledge or skills through learning, improve its existing knowledge structure, and continuously improve its own performance. It is the core of artificial intelligence, and its applications […]

  • Paper reading: multimodal graph networks for compositional generalization in visual question answering


    Title: multimodal graph neural network for combinatorial generalization in visual question answeringSource: neurlps 2020https://proceedings.neurips.cc/paper/2020/hash/1fd6c4e41e2c6a6b092eb13ee72bce95-Abstract.htmlcode:https://github.com/raeidsaqur/mgn 1、 Questions raised a key:Combinatorial generalization problem Example: taking natural language as an example, for example, people can learn the meaning of new words and then apply them to other language environments. If a person learns the meaning of a new […]

  • Introduction and implementation of multi task learning model based on esmm


    Introduction: This article introduces the paper “entire space multi task model: an e” effective approach for estimating post click conversion rate “published by Alibaba team in SIGIR ‘2018. Based on the idea of multi task learning (MTL), this paper proposes a CVR prediction model called esmm, which effectively solves the two key problems of data […]

  • Excel skills: how to achieve vlookup matching of separated data conditions?


    During the recent training in a German enterprise, a small partner asked me a very typical matching problem during the Q & a process, but this matching problem is quite thorny. Niu Shanshan found that such problems often occur in the workplace, so she quickly shared them with you. I simplified his data model. After […]

  • Unity shader for creeping effect


    Unity shader for creeping effect I. Preface Although shaderforge has the creeping function, as an excellent shader program ape, we still need to have a certain understanding of the principle. Only in this way, when we write the shader, we can be handy, and make certain optimization and changes to the related functions. This article […]

  • Enterprise intelligent transformation meetup review | open source Bi & AI helps enterprises transform in three stages!


    On May 22, 2022, xingce community, together with the technical experts from Weizhong bank, the fourth paradigm, ZTE and other co construction units, jointly held the first “enterprise intelligent transformation meetup” in the community. This activity introduced how to use open source Bi & AI technology to help enterprises complete the three stages of informatization […]

  • Fastapi learning path (21) request body – update data


    Series:   Fastapi learning path (I) fastapi — high performance web development framework   Fastapi learning path (II)   Fastapi learning path (III)   Fastapi learning path (IV)   Fastapi learning path (V)       Fastapi learning path (6) query parameters and string verification   Fastapi learning path (VII) string verification     Fastapi learning path (VIII) verification of path parameters and values […]

  • Fastapi learning path (34) database multi table operation


      Previously, we shared the operation based on a single database table. We also designed a cross table when designing the database. We can take a look at the database design.          class User(Base): __tablename__ = “users” id = Column(Integer, primary_key=True, index=True) email = Column(String, unique=True, index=True) hashed_password = Column(String) is_active = […]

  • “Neural factorization machines for sparse predictive analytics” – paper abstract


    1. Foreword FM can effectively find the second-order combination features, but the problem is that the second-order combination features captured by FM are linear combination (its expression is linear combination), and can not capture the nonlinear combination features. Now, deep neural networks can find nonlinear combination features, such as Google’s wide&deep and Microsoft’s deepcross, but […]

  • MindSpore! I love this new open source deep learning framework!


    Reprint address:https://bbs.huaweicloud.com/f… Author: red stone MindSpore! I love this new open source deep learning framework! 1.jpg I still remember that at this year’s Huawei Developer Conference HDC 2020, mindspire, a deep learning framework that has always attracted attention, was finally open source. I’ve been following mindspire, and I’m looking forward to it. Mindspire is a […]

  • R language back test transaction: create stock return curve based on historical signals / transactions


     Original link: http://tecdat.cn/?p=23808 This article describes how to create stock curves based on historical signals / transactions. Let’s take the historical signals of mark timing and decision moose as examples to create a stock curve for this strategy. #***************************************************************** #Load signal #***************************************************************** #Extract transaction history temp = extract.table.from.webpage(txt, ‘Transaction History’, has.header = F) temp = trim(temp\[-1,2:5\])   colnames(temp) = spl(‘id,date,name,equity’)    tickers = toupper(trim(gsub(‘\\\)’,”, sapply(temp\[,’name’\], spl, ‘\\\(‘))))\[2,\] load(file=filename)     #plota(make.xts(info$equity, info$date), type=’l’) […]

  • “Wide & deep learning for recommender systems” – paper abstract


    Introduction One of the difficulties of the current recommendation system is to implement both memory and generalization at the same time. This difficulty is similar to the search ranking problem. Memorization: The previous processing of large-scale sparse input is: linear model + feature crossover. It can bring good effect and strong interpretability through feature crossing. […]