Tag：vector
-
System recall too slow? Upper Milvus × Padderec double sword combination method!
Introduction to the authorLi Yunmei, zilliz Data Engineer, graduated from the computer department of Huazhong University of science and technology. Since joining zilliz, he has been committed to exploring solutions for the open source vector database Milvus and helping users create scenario applications. I like to search for books and recommend natural language processing system. […]
-
Omnipotent embedding4 – skip thought & TF seq2seq source code analysis
In the previous chapter, doc2vec mentioned that in fact, doc2vec only joined doc_ ID captures the subject information of the text without really considering word order and context semantics. N-gram can only solve this problem locally. Is there any other solution? It is still a general text vector. Skip thought tries to use encoder decoder […]
-
Omnipotent embedding5 – brothers of skip thought [trim / CNN LSTM / quick thought]
In this chapter, we will talk about the three brothers of skip thought, who have made different attempts to solve the remaining problems of skip thought [REF1 ~ 4]. The following paper may not give the optimal solution (there is no optimal solution for different NLP tasks, but only the most appropriate), but they provide […]
-
40000 word comprehensive and detailed explanation of the attention mechanism in deep learning (conclusion)
This section introduces various variants of the attention mechanism. Including but not limited to: “Attention mechanism based on reinforcement learning”: a part of the optional append input “Global & local attention mechanism”Among them, the local attention mechanism can selectively attach some parts of the input “Multidimensional attention mechanism”Attention: capture different features in the attention space. […]
-
Machine learning algorithm series (0) – basic knowledge
The background knowledge required for reading this article: first, lose mathematical knowledge 1、 Introduction artificial intelligence (AI) plays a more and more important role in modern life. The support of AI is inseparable from the functions of various voice assistants, color restoration of old movies, intelligent recommendation of e-commerce websites such as Taobao Jingdong, […]
-
Machine learning algorithm series (III) – standard linear regression algorithm
Background knowledge required for reading this article: matrix derivation and yidui programming knowledge 1、 Introduction two binary classification algorithms, perceptron algorithm and pocket algorithm, are introduced earlier. These algorithms solve the problem of classification, but in reality, they are more problems such as predicting the house price in a certain area, how much credit […]
-
Sign language can also be translated by machine! Machine learning gesture recognition function
Huawei machine learning (ML Kit) provides hand key point recognition service, which can be used for sign language recognition. The hand key point recognition service can recognize 21 key points of the hand and find the sign language alphabet by comparing the direction of each finger with the sign language rules. Application scenario Sign language […]
-
Intrigue with rust · point
Previous:preface The origin of the three-dimensional world is the point whose projection is 0 on the three number axes. Assuming that the three number axes are x, y and Z, the projection of the origin on them can be expressed as x = 0, y = 0, z = 0, which can be expressed as […]
-
Flare animation developed by flutter
Introduction to flare animation Flutter’s animation can be roughly divided into three categories: basic animation controlled by animationcontroller and animation, transition animation controlled by hero and custom animation controlled by custompainter. In addition, fluent also supports vector animation, which is similar to Lottie animation in Android development. Flare is a website that can quickly make […]
-
Natural language processing — Text vectorization (2)
1、 Abstract This sharing is based on the continuation of the previous text vector method. In the last content, we mainly shared two methods of text vectorization: the word bag model representation method and the word vector method based on deep learning. Although the word bag model can simply express words as vectors, it will […]
-
Machine learning algorithm series (IV) – ridge regression algorithm
Background knowledge required for reading this article: standard linear regression algorithm and yidui programming knowledge 1、 Introduction In the previous section, we learned the machine learning algorithm series (III) – standard linear regression algorithm, and finally obtained the analytical solution w of the cost function of standard linear regression as follows: $$ w=\left(X^{T} X\right)^{-1} X^{T} […]
-
Calculation algorithm of the shortest distance and line segment of two lines / line segments in three-dimensional space
There are two space segments $L_ S $, whose starting and ending coordinates are $s_ 0、s_ 1 $, direction vector $\ VEC u = s_ 1-s_ 0 $ $L_ T $, whose starting and ending coordinates are $t_ 0、t_ 1 $, direction vector $\ VEC v = t_ 1-t_ 0 $Note that the line corresponding […]