Africa is building a completely different artificial intelligence and Silicon Valley industry

Alt

Alt

Indaba-- growth in Africa's artificial intelligence community

In late August, in the shade Nairobi, Kenya (Nairobi) an arched pepper tree, hundreds of artificial intelligence researchers chatted about their algorithms. Some people stood in front of a poster, posters wrapped around the tree roots messy, depicting a machine learning system, commitment can predict everything, nutrients from the soil to a small-scale farmer whether it will repay the loan, then how autonomous vehicles in Cairo driving on a crowded street.

Over the past three years, academics and industry researchers from all over the African continent at a meeting called Deep Learning Indaba on, began to sketch the future of their own artificial intelligence industry. The conference brings together hundreds of researchers from more than 40 African countries to present their research results and to deal with artificial intelligence, natural language of ethics on all issues discussed.

Indaba was established in 2017, is a direct response to the Western Conference, the Western Conference often allows researchers from around the world to participate difficult. To Neural Information Processing Systems Conference (NeurIPS), for example, which is the most famous artificial neural network meeting. It has been held in the distant and expensive resorts before NeurIPS.

For the researchers were able to afford the cost, it is a vacation. In 2006 and 2007, which are located in The Westin Resort & Spa Whistler, British Columbia and Hilton Resort & Spa, a form of "informal discussions, skiing and other winter sports," the. For researchers from Africa, NeurIPS usually out of reach. 2016, meeting no papers of any African country was accepted. 2018, more than 100 researchers have been denied a visa to enter Canada, enter NeurIPS.

We need to find a way to build our image of machine learning methods in Africa.

Therefore, in 2017, South Africa University of the Witwatersrand (University of Witwatersrand) of former students and colleagues created some close together Indaba, they name it in a Zulu word meaning " important meeting or gathering. "

"How many papers have at least one of the research institutions from Africa?" The answer is: 0. "In the contemporary field of machine learning, two whole continents are gone."

Organizers estimated that about 50 people attended the first session of the General Assembly Indaba, but there are nearly 750 people apply, 300 people invited to participate. The next year, Indaba invited 400 people, and extended to 13 IndabaX meeting. Scale conference this year almost doubled, with 700 participants and 27 IndabaX activities.

Alt

Depth study of artificial intelligence community Indaba has become the connective tissue of Africa - not just community meeting space, it is also part of the community itself. The meeting with a clear agenda to strengthen the relationship between the African continent researchers: the establishment of a vibrant pan-African technology community, rather than re-invent the art, but through tailored to the challenges facing the region solutions to solve traffic congestion, insurance claims paid, and drought patterns.

Google, Microsoft, Amazon and other technology companies to Indaba provided about $ 0.3 million in funding, but organizers still wants to create a new, unique research - an area not controlled by Silicon Valley.

As Indaba organizer, University of Pretoria, South Africa (University of Pretoria) data science department Oecussi Marie Witt (Vukosi Marivate) said to me, "We need to find a way, in accordance with our vision to build machine learning in Africa. "

Derived from the technology giant's dilemma

This year's "Deep Learning Indaba" held at the Kenyatta University (Kenyatta University), for a period of six days. Kenyatta University in Nairobi sits Xi Kalu (Thika Road) next to it is a bustling 8-lane highway, motorcycles, taxis and shuttle buses to the city center.

In the Indaba participants, students accounted for a large proportion, which is why this conference is so concerned about a major cause of education. The first day of the conference was devoted to the popularity of AI knowledge and introductory courses, such as the basics of statistics and build neural networks. In this week's lesson, course adds more advanced topics. The participants took part on natural language processing, computer vision, depth specialized courses of intensive study and AI moral ethics. Some people attended the hackathon, where they established the spread of African wildlife can be automatically identified in order to better study and protect endangered species of artificial intelligence, while others use health data to predict and control malaria.

现在,在非洲建立业务意味着从用户的数字生活一开始就与他们建立有价值的关系

这些天通常是以一个主题来开始全天的议程:IBM 研究员 Aisha Walcottt - Bryant 谈到了与肯尼亚政府合作进行数据收集的问题;普林斯顿大学的 Ruha Benjamin 开了一门关于从编码到算法系统中的不等式的课程;Salesforce首席科学家 Richard Socher 谈到了他的团队在构建更通用的人工智能系统方面的研究。

作为非洲举行的卓越的人工智能大会,Indaba吸引了科技巨头的极大关注。Indaba的34个赞助商中有11个是美国公司,其中包括了 Google,Microsoft,Amazon,Apple和Netflix。

这是硅谷公司在非洲大陆进行重大投资的趋势之一。Google赞助了非洲数据科学和非洲数学科学研究所等组织,2018年还宣布了在加纳阿克拉建立首个非洲研究中心。与此同时,微软和比尔·盖茨基金会向尼日利亚数据科学捐赠了近100,000美元,希望在未来10年内可以培训100万名尼日利亚工程师。

Alt
非洲投资的吸引力是显而易见的。75%的非洲大陆仍然无法访问互联网。这对当地人来说是一个挑战,但对国际科技公司来说也是一个投资机会。现在,在非洲建立业务意味着从用户的数字生活一开始就与他们建立有价值的关系。最近的一份投资者报告显示,在发展中国家,Facebook每年每位用户的收入为2.13美元,高于2015年的0.90美元。

Indaba创始人在一次会议后写道:“ Indaba的大部分资金来自国际组织,我们的许多国际发言人来自国际技术公司。” “这可能给人留下这样的印象,即最好的工作发生在大型科技公司以及非洲大陆以外的国家,人们必须离开非洲大陆才能在该领域拥有有意义的职业生涯。”

这就是组织者试图在国际赞助商与本地赞助商之间寻求平衡的原因,并着重强调了非洲大学和科技公司的机会。

今年的Indaba会议向南非电信MTN首席转型官Bayo Adekanmbi以及数据科学尼日利亚(DSN)的创始人颁发了Maathai Impact奖,该组织已经培训了成千上万的尼日利亚人来支持本国的IT部门。

“ Indaba的资金很大一部分来自国际组织,我们的许多国际发言人也来自国际技术公司。”

阿德坎比(Adekanmbi)穿着一件塞在牛仔裤里白色短袖纽扣衬衫。他一只手拿着公文包,看上去随时准备传播数据科学的福音。而学生们也都很尊敬他。

阿德坎比(Adekanmbi)希望在未来10年内培训100万尼日利亚数据科学家,因此,他非常重视人才流失的威胁。“这是一个很大的问题。人才总是转移到最集中的领域。”他说。“如果我们不在这里建立另一个社区,人才将继续流失。”

远程工作是留住人才的一种方式。阿德坎比(Adekanmbi)启动了一个名为“数据科学家随需应变”(Data Scientists on Demand)的项目,计划帮助尼日利亚的工程师远程为世界各地的公司工作。

但这还不够。阿德坎比(Adekanmbi)说,希望从与非洲的合作中获益的公司应该在非洲大陆有实体存在,并通过投资当地的技术社区来显示决心。

他说:“如果你真的想让包容性,公平性和知识的分布与当地相关,那么企业应该愿意在世界各地建立卓越中心,并在全球范围内建立知识中心。”

Alt

突尼斯人工智能公司InstaDeep的联合创始人卡里姆·贝吉尔(Karim Beguir)结束了在伦敦的金融职业生涯回到非洲,建立了自己的公司。这就是典型的初创公司故事:放弃一份轻松的工作,梦想着用两个人和两台笔记本电脑创办一家新公司。他在Indaba教过一门入门数学课程,并举办了一个为期两届的研讨会,讨论如何建立一家初创企业:如何找到联合创始人,如何获得最佳的税收优惠以及吸引投资者。Beguir说他并不担心非洲的人才流失——他认为自己的发展轨迹是一个跨洲合作的模式。

对许多非洲人来说,与国际科技巨头的合作对他们开展工作至关重要。Indaba的与会者Tejumade Afonja是来自尼日利亚的工程师,他与全球AI Saturdays组织联合创办了一个为期16周的AI编码研讨会。她表示,英特尔的赞助使整个工作得以持续进行。英特尔要求组织者和讲师为学生提供使用英特尔软件和硬件的选择,但不需要求他们支付这笔费用

加纳数据保护委员会前执行董事、非洲数字权利中心创始人Teki Akuetteh Falconer表示,美国科技公司是唯一对扩展互联网基础设施和研究技术生态系统抱有浓厚兴趣的组织。“老实说,你逃不掉的。我经营着一家非政府组织,我必须为它提供资金,”Falconer说。“我的资源不够用。奇怪的是,只有这些(互联网)公司能理解我。”

未来一直在来

Indaba仍在寻求扩展。那些对构建非洲人工智能未来感兴趣的人反复听到的一个短语是“能力”(原文单词为“Capacity”)。“把更多的专业人士留在这片大陆,意味着我们的研究人员可以获得更多的专业技能和知识的传承,同时也可以获得更好的教育下一代的能力与资源。它也是关于Indaba作为全球深度学习社区在全球你范围内积累声誉的重要一部分。

这一进程的一部分是在各个国家/地区创建“Indaba分社”。在南非、塞内加尔和索马里等国家之间,现在有27个IndabaX活动,这些活动借鉴了TED大会上的“X”的术语概念。根据比赛大纲,他们的规模从几个人到几十人不等,而比赛的目标就是追求“努力和卓越”。

比赛获胜者和IndabaX的组织者随后将会被邀请参加主要的Indaba会议。在那里,将会有两名Deep Learning Indaba的优胜者可以获得全额赞助参加NeurIPS。其目标是最终让非洲的人工智能研究人员与西方的研究人员能在学术上处于平等的地位。

"Now we're convey this movement. Now you will see an equal delegation NeurIPS." They do not say, 'Oh, these Africans, we should bring them', but equally to the negotiating table before. "

"Currently, we are promoting the concept of the game. You will see the delegation Indaba in NeurIPS, we and other researchers and scholars are equal." Marivate said. "They will not say, 'Oh, these Africans, we should bring them', but equally we sit at a negotiating table with."

Original link: https://onezero.medium.com/africa-is-building-an-ai-industry-that-doesnt-look-like-silicon-valley-72198eba706d

If you want to learn more about cloud Jingdong product information, please click on the " Jingdong cloud " for more exciting content.

Alt
The above information comes from the Internet, compiled by "Jingdong cloud developer community" public editor numbers do not represent the position of Jingdong cloud.

Alt

Guess you like

Origin www.cnblogs.com/jdclouddeveloper/p/12118466.html