Developer Community Tech Week is back with us, so let’s take a look at some of the most important news of the past week for us developers.

  • The C++20 standard is officially released

  • The scale of China’s AI cloud service market grows much faster than expected

  • Nature: AI recognizes 16 facial expressions common to all human beings, transcends regions and cultures

  • DeepMind’s latest AI product, MuZero, can master a game without knowing the rules

  • The highest availability guarantee reaches 99.995%. Jd Zhilian cloud upgrade SLA service standard

  • Mimik works with IBM to make edge computing resources more accessible

  • “Microservice Fault Location Based on Trajectory Analysis”, Doctoral dissertation of Fudan University

  • AAAI 2021 | from the perspective of semantic relation modeling matching sentence semantic relations

News of the technology

Industry   News

1, the C++20 standard was officially released

Recently, the ISO C++ committee officially released the C++20 standard, named ISO/IEC 14882:2020. This major update introduces a number of new features:

  • Modules (Modules)

  • Coroutines (Coroutines)

  • Range (Ranges)

  • Constraints and Concepts

  • Designated Initializers

  • Operator <=>! == =, constEXPr supports: new/ delete, dynamic_cast, try/ catch, virtual

  • Constexpr vector and string

  • Timing: calendar, time zone support

  • std::format 

  • std::span 

  • std::jthread

Concepts/Ranges/Modules/Coroutines these new features will be a big impact on c + + developers and ecological, also add a lot to the “old age” programming language charm.

The scale of China’s AI cloud service market is growing much faster than expected

According to the latest REPORT “China’S AI Cloud Service Market (First Half of 2020) Tracking” released by INTERNATIONAL Data Corporation (IDC), the scale of China’s AI cloud service market reached 1.09 billion yuan in the first half of 2020, an increase of more than 100% compared with the first half of 2019. Cloud service vendors can provide rich and rapidly iterative single-point AI capabilities, as well as the ability to create application scenarios and overall industry solutions for industry users. The AI public cloud service market is still in its early stages, and in the future, AI cloud service vendors will gain a higher and higher market share in the overall AI software and application market.

3. Nature: Beyond regions and cultures, AI recognizes 16 facial expressions common to all human beings

Researchers from the University of California, Berkeley, and Google used machine learning and YouTube videos to compare common facial expressions with the situations that trigger them. It turns out that humans use the same facial expressions — smiles, frowns, grimaces — in similar social situations, whether it’s a birthday party in Brazil, a funeral in Kenya, or square dancing in a Park in China. In other words, the human variety of facial expressions can transcend regions and cultures. The study, published December 16 in the journal Nature, confirms the universality of human emotional expression across geographical and cultural boundaries.

DeepMind’s latest AI product, MuZero, can master a game without knowing the rules

MuZero, the latest ai product from DeepMind, Alphabet’s ai focused subsidiary, can master games without knowing the rules, according to foreign media reports. DeepMind’s mission is to show that AI can not only truly master a game, but can do so without being told the rules. On Wednesday, local time, the company detailed MuZero, an ARTIFICIAL intelligence that can master a game without having to learn its rules beforehand. MuZero can achieve this not only with visually simple games with complex strategies such as Go, chess, and Jai, but also with visually complex Atari games. Instead of using rules to find the best case, it learns to consider every aspect of the game environment and see for itself if it matters.

5, the highest availability guarantee of 99.995% jd Zhilian cloud upgrade SLA service standard

Recently, JD Zhilian cloud released a new upgraded cloud computing service level agreement, the highest availability guarantee of 99.995%. This means that jd Zhilian cloud as a cloud computing service provider, service availability level has been among the world’s top cloud computing manufacturers. “99.995%” behind this value, jingdong Zhilian cloud high level of cutting-edge technology advantages and rich business experience accumulation. Jingdong Zhilian cloud’s technical ability, after jingdong retail, Jingdong logistics, Jingdong data department and other core business of jingdong Group, and through the test of JINGdong 618 and JINGdong 11.11 years, has long-term technical accumulation and practice in ensuring service availability.

6. Mimik works with IBM to make edge computing resources more accessible

Edge computing startup Mimik is partnering with IBM to create edge-based computing resources that it says will help accelerate automation and digital transformation in healthcare, manufacturing, retail, and other industries. The growth of trends such as the “Internet of Things” and 5G has led to the emergence of billions of connected devices and applications at the edge of the network, Mimik said on December 24. But these devices and applications rely on computing resources located in centralized data centers, and they are becoming a bandwidth burden and therefore difficult to manage.

Mimik’s partnership with IBM could help solve this problem. The company has created a hybrid EdgeCloud platform that works by turning devices such as smartphones and PCS into “cloud servers” that can run on the edge of the network. The edgeCloud platform can take advantage of these devices, enabling them to communicate with each other in small local clusters regardless of the operating system and network they are running on. This makes it possible to use their combined computing resources to power applications at the edge of the network.

The academic frontier

Academic News

1. Fudan University doctoral thesis “Micro-service Fault Location Based on Trajectory Analysis”

As a new form of cloud-based software system, microservice system lacks a corresponding summary of its software fault characteristics, and also lacks open microservice system (such as open source project) for research. Therefore, in this paper, Trace Analysis Based Fault Localization for Microservice Systems conducted an industrial investigation in response to the development practice of Microservice Systems, and systematically understood the characteristics of Microservice Systems in the industry. A group of typical microservice fault cases and corresponding fault analysis and debugging practices are collected, and the practical problems and challenges are analyzed. From there, they developed a mid-scale microservices benchmark system that recreated 22 industrial failure cases and released it through the open source community. Based on the open source microservice benchmark system and the corresponding fault cases, a series of microservice fault assisted location methods and technologies are studied.

Links to papers: arxiv.org/abs/2006.06…

2, AAAI 2021 | from the perspective of semantic relation modeling matching sentence semantic relations

Sentence semantic relation matching is a basic but important task in natural language processing, and has a wide range of applications, such as retrieval, dialogue, question and answer and other practical scenarios. Existing methods can be roughly divided into two categories: one is to directly model the sentence itself, using BERT, LSTM and other methods, or generate semantic representation vector for matching analysis, or directly analyze the matching between words and phrases from a fine-grained perspective; The other category also considers the information of labels, because labels also contain certain semantic information. Rich information is supplemented by the method of label embedding, so as to obtain the semantic matching relationship of sentences.

This paper argues that the semantic meaning of tags indicates the semantic relationship between the two sentences, which has been neglected by existing researches. Through the mining of semantic relationship, semantic information of sentences can be further mined, features related to semantic relationship extracted, and better semantic relationship modeling of sentences can be realized. Therefore, relation of Relation Modeling is proposed in this paper to conduct modeling analysis of sentence semantic relation matching from the perspective of semantic relation modeling.

Links to papers: arxiv.org/abs/2012.08…

The above information comes from the network, edited by the “Jd Zhilian cloud developer” public number, does not represent the position of JD Zhilian cloud

Recommended reading:

  • K8s will deprecate Docker; TypeScript overtook C++ as the fourth largest language

  • IDC released the top 10 cloud computing forecasts in China in 2021; Docker Desktop launches tech preview for M1

  • A Chinese Internet of Things security testing technology has become an international standard; Windows 10 will support Android apps

Welcome to [JD Zhilian Cloud] to learn about the developer community

More wonderful technical practice and exclusive dry goods analysis

Welcome to [JD Zhilian cloud developer] public account