SXSW洞见:追一场全球最大文化盛典带来的“燥科技”潮

全力奔跑的 复星锐正资本

与CES(消费电子展)和MWC(世界移动通信大会)等知名活动相比,大多数亚洲人可能不太熟悉SXSW(“西南偏南大会”)。

 

2018年3月9日至18日,SXSW在德克萨斯州的奥斯汀市举行。这个为期一周的活动集音乐节、电影节以及互动/前沿技术会议于一体。与行业展览性质的CES和MWC不同,SXSW为创作人和创新者提供独特的机会,让业界专家齐聚一堂,深入探讨如何在不同的用例中将技术运用于电影、音乐、多媒体和医疗等行业。


本文是复星锐正资本美国团队同学参与这场科技与艺术狂欢盛宴的随手笔记,带大家一起感受这场“极客狂欢节”和它在2018年掀起的新浪潮



SXSW简介


一年一度的电影、音乐和互动媒体盛会SXSW于1987年在德州奥斯汀市创办,其范围和规模正在逐年扩大。早期主要是音乐和电影节,现在已成为音乐和电影发布的主要场所,经常有重量级制片人、导演、演员亮相。SXSW的总参与人数从1987年的700人左右增加到2017年的30多万人。

 

近年来,展示和探讨前沿技术的SXSW互动会议正成为SXSW的重头戏。回顾SXSW历年的科技行业演讲嘉宾/主旨演讲嘉宾,可以明显看出SXSW为何享有新鲜创意和创新技术温床的美誉:


SXSW的历年演讲嘉宾:

 

2000年– Sean Parker,Napster(纳普斯特)

2004年– Jonathan Abrams,社交网站Friendster

2006年– Craig Newmark,Craigslist

(克雷格列表网站)

2008年– Mark Zuckerberg,Facebook

2009年– Tony Hsieh,Zappos

2010年– Ev Williams,Twitter

2011年– Brian Chesky,Airbnb(爱彼迎)

2012年– Ben Silbermann,Pinterest

(品趣志)

2013年– Elon Musk,SpaceX(太空探索技术公司)


该活动还吸引了越来越多知名人士的参与,其中包括奥巴马总统(2016年),以及阿诺德·施瓦辛格、梅琳达·盖茨和斯蒂芬·斯皮尔伯格(2018年)。



会议亮点


SXSW会议可分为以下6个特别振奋人心的议题:

 

1. 民主化以及人工智能将开启的新体验

2. 电影和医疗行业中的AR/VR

3. 未来交通

4. 区块链及其应用

5. 公司如何拥抱开放式创新、多元化,并建立品牌形象

6. 专题会议:棒球、商业与救赎


我们将此六个议题分成两期,并用中英文交织的方式,带您了解原汁原味的SXSW。



 民主化以及人工智能将开启的新体验







1

设计下一波自然语言和人工智能


在这个激动人心的小组讨论中,来自谷歌、微软和梅赛德斯 - 奔驰的代表齐聚一堂,讨论公司开发人工智能功能时要考虑的关键因素,以及这些科技巨头也要面临的挑战。

 

谈到开发人工智能嵌入式解决方案时公司应考虑的关键因素,以及影响人工智能的采用和开发的因素时,谷歌智能家居(Google Home)的产品负责人、微软人工智能研究的首席项目经理和梅赛德斯 - 奔驰的机器学习主管与观众分享了以下内容:


-  人工智能功能开发过程成本高昂,投入资源之前要考虑应用人工智能的确切场景。


-  在开发谷歌智能家居时,谷歌注意到,在推出带有人工智能的新体验时,应逐渐向人们的生活方式引入变化,并且体验需是无缝的。


-  对于梅赛德斯来说,集成人工智能的关键目标是更有效地利用驾驶者的时间,因此人与自主车辆的信任/关系非常关键。


-  关于人工智能输出的信任问题,用例确实决定了输出数据中的错误容限。例如:在使用人工智能进行图像分析时,识别网上的时尚商品,而不是识别用于诊断的医疗图像中的特征,对误报的容限会有很大差异。


关于科技巨头们自身所面临的挑战及其所看到的趋势,分享的一些要点包括:


-  即使是这些大公司,在开发人工智能时需解决的挑战仍然围绕着数据质量,数据的取舍。小组成员强调,为了提供来自人工智能的正确见解,训练质量及由此的输出数据是最重要的。


-  科技巨头和学术界也在进行研究,以检验如何用更小的数据集来训练人工智能。


-  在自然语言处理领域,虽然已经有充足的使用语音的应用程序,但仍有机会在未来将人工智能应用于手势、面部表情等,以帮助用户从人工智能获得更完整的语境。



The democratization, and new experiences to be unlocked with AI





1

Designing the Next Wave of Natural Language and AI


At this exciting panel, representatives from Google, Microsoft, and Mercedes-Benz were assembled to discuss key factors to consider when a company develops its AI capability, as well as the challenges that even these tech giants face.

 

Speaking about the key factors companies should think about when developing AI-embedded solutions, and what affects the adoption and development of AI, the Product Lead of Google Home, Principal Program Manager of Microsoft AI Research, and Director of Machine Learning at Mercedes-Benz shared with the audience:


- Developing AI capability is a costly process, think about the exact scenarios to apply AI to, before investing the resources.


-  When developing Google Home, Google was mindful that when introducing new experience with AI, changes should be introduced gradually to people’s way of life, and experience needs to be seamless

-  For Mercedes, a key goal for integrating AI is to enable more efficient use of driver’s time, hence the trust/relationship one has with their autonomous vehicle is critical.


-  Further on the trust issue of AI output, the use case really determines the tolerance of errors in output data. For example, when using AI for image analysis, to identify fashion merchandises online, versus to identify signatures in medical images for diagnostics, the tolerance for false positives would be very different.


And regarding the challenges faced by the tech giants themselves, and trends they see going forward, some of the points shared include:


- Challenges that even these large companies need to tackle when developing AI is still around data quality, the inclusion and discard of data. The panelists reiterated that in order to provide the right insights from AI, quality of training and hence output data is upmost essential.


-   Researches are also underway between the tech giants and academics to examine how to train AI with smaller data sets.


- On the area of natural language processing, while there’s already ample applications using voice, there’s opportunities in the future for AI to be applied to gestures, facial expressions, etc to help users get a fuller context from AI.



2

为个人和组织将人工智能民主化


这个环节的演讲嘉宾有:当今人工智能领域中最具影响力的人物之一、谷歌云计算业务首席科学家、斯坦福大学计算机科学教授兼人工智能实验室主任李飞飞博士;以及曾担任奥巴马总统的美国首席技术官,前谷歌高管,同时也是支持TechHire、全民计算机科学和Image of STEM活动等科技教育民主化倡议的关键人物Megan Smith。后者也是许多促进多元化倡议(如马拉拉基金会)的倡导者。

 

讨论围绕人工智能的民主化进行。由于人工智能在改善生活方式方面具有巨大潜力,但要发挥其全部潜力,必须移除制度性进入壁垒,从而吸引更广泛和多样化的参与者。该技术必须提供给最广泛的开发者、企业家和用户。为了实现这一目标,李博士和Megan主张每个人都有一席之地,而不仅仅是硅谷的科技巨头,此外,初创公司、教育工作者和政府等都需要参与进来。

 

对于政府来说,这有助于影响学校和其他地方的课程开发,并使具有适当技能的人员进入劳动力队伍。例如,Megan在奥巴马总统的领导下帮助创建了TechHire,该倡议旨在通过在美国建立高科技人才管道来推动当地科技行业发展,目前已覆盖70多个社区。

 

为了实现人工智能的民主化,李博士正在与梅琳达·盖茨合作一个名为AI4All的非营利组织。该组织支持旨在让不太受重视的高中学生接触到人工智能的教育项目。

所有这些努力都是为了使大众更容易接触到人工智能教育和研究,并吸引更多不同的人群。

非营利组织AI4All▲

在奥巴马政府期间创建的TechHire倡议▲



2

Democratizing AI for Individuals & Organizations


In this session, the audience got to hear from one of today’s most influential minds in AI – Dr. Fei Fei Li, the Chief Scientist of Google Cloud, Stanford CS professor and the Director of the university’s AI Lab, and Megan Smith, who had served as the U.S. CTO under President Barack Obama, previously an Executive at Google, and was a key figure behind initiatives to democratize science and technology education such as TechHire, Computer Science for All, and the Image of STEM campaign. She’s also an advocate behind many initiatives to promote diversity, such as the Malala Fund.

 

The discussion was around democratizing AI. As AI has enormous potential for improving the way of life, but to realize its full potential, the systematic barriers of entry must be fixed to engage a much broader, and diverse set of participants. The technology must be made available to the largest possible community of developers, entrepreneurs, users. To achieve this cause, Fei Fei and Megan advocated everyone has a seat at the table, and not just Silicon Valley tech giants, in addition, startups, educators, and Governments alike, need to all participate.

 

For Government, it can help influence curriculum development at schools and elsewhere, and enable people with appropriate skillsets to enter the workforce. For instance, the TechHire initiative Megan helped created under President Obama aims to expand local tech sectors by building pipelines of tech talents across the US, which now covers more than 70 communities.

 

In an effort to democratize especially AI, Dr. Fei Fei Li is working with Melinda Gates on a non-profit called AI4All. The organization supports educational programs designed to expose underrepresented high school students to AI.

All these efforts are aiming to make AI education and research more accessible, and appealing to a more diverse population.

The AI4All non-profit▲

TechHire Initiative established

 during the Obama Administration▲




3

为什么机器学习是艺术的下一个前沿


该会议首先提出“机器能否确定艺术的价值?”这个问题。带出的一个例子是2017年达芬奇的“救世主”拍卖会,这幅画以4.5亿美元的天价成为有史以来市场上出售的最昂贵艺术品。但是,是什么决定了这一价值?是否有一种方式让机器做出这样的预测?


这次会议是人们对SXSW可以有所期待的一个典型例子,一个技术与媒体/艺术交汇的独特机会,也是我在SXSW参加的最有趣讨论之一。

 

在探索人工智能的功能边界时,除了路况、购买模式和健康诊断等实实在在的内容外,还有一个问题值得思索,即人工智能能否对审美、味道等非常主观的东西做出判断?

 

演讲嘉宾Hugo Liu博士是Artsy的首席科学家。作为麻省理工学院媒体实验室的校友,他是人工智能和消费者偏好预测前沿应用的领军人物。

 

目前,艺术品收藏是一个价值500亿美元的市场,Hugo认为现在每年参观美国博物馆的人数为8.5亿人次,如果能使5%的参观者成为艺术品收藏者,平均每人花费1000美元,就可以为艺术品收藏市场再增加425亿美元(8.5亿* 5%* 1000美元)。但问题是,尽管机器可能拥有一切洞察力,但它能否有助于新晋艺术家客观地鉴定艺术品价值?


Hugo一直致力于研究如何创建一个透明的、数据驱动的市场,就像是艺术界的Zillow(译注:美国房地产信息查询网站)。在这个过程中,他的团队使用机器来分析数以万计的艺术品和交易历史,并试图根据艺术家的展览历史、参展画廊和博物馆以及画廊/博物馆的历史来对艺术品进行分类,从而预测获胜者等。此外,在分析了一些传世名作后,Hugo的团队试图用人工智能来研究这些名作的主题、构图、风格和色彩主题等,并将它们与机器辅助艺术品分类文化影响联系起来,从而与艺术品的交易价值相关联。这个项目还在进行中,不断产生着引人关注的见解。

输入一组数据帮助训练机器预测艺术品的价值▲




3

Why Machine Learning is the Next Frontier for Art


The session began by asking the question ‘can a machine determine the value of art?’ An example brought up was the auction of Leonardo da Vinci’s “Salvator Mundi” in 2017, for a whopping $450M, the most expensive artwork ever sold. But what determines that value? And if there’s a way for machines to make such a prediction? 


This session was a great example of what one can expect at SXSW, a unique opportunity to be at the intersection of technology and media/art, and one of the most intriguing talks I attended at SXSW. 

 

When exploring the boundary of AI’s capability, beyond something more specific such as road conditions, purchasing patterns, and health diagnostics,..etc, a question arises regarding can AI make judgements around beauty, taste, which is something so subjective?


The speaker Dr. Hugo Liu is the Chief Scientist of Artsy. An MIT Media Lab alumni, he is a pioneer in cutting-edge application of AI and consumer taste prediction.

 

Art collection is currently a $50B market, and Hugo argues that with now 850M visits a year to US museums, if it’s possible to just activate 5% of these visitors to become art collectors, with an average of $1,000 a person spending, it can add another $42.5B to art collection market (850M * 5% * $1,000). The question is with all the insights it may have, will machine be able to help objectively assign artworks a value, even to first time artist?


Hugo has been examining the possibility of creating a transparent, data-driven marketplace, or a ‘Zillow for art’. In this process, his team uses machines to analyze tens of thousands of artworks and previous transactions, and attempted to categorize the artworks based on an artist’s exhibition history, participating galleries and museums, and the history of the gallery/museum to predict winners, etc. In Using machine to aid the categorization of artworks addition, upon analyzing some of the historic masterpieces, Hugo’s team attempted to use AI to look at the topic, composition, style, and color theme,..etc of these masterpieces, and correlate them with the cultural influence, and hence transaction value of the artworks. It’s still a project in the works, but continues to yield interesting insights. 

A set of inputs to help train machines to predict the value of artworks▲



电影和医疗行业中的AR/VR







1

电影和多媒体中的AR/VR 


这是SXSW最热门的话题之一,我参加了几场讨论和演示。在过去一年左右,人们怀疑AR/VR是否达到了炒作周期的低谷,是否在市场上占有一席之地。在SXSW期间,很明显,它不仅非常活跃,而且VR内容开发中的利益水平和投入的努力也不断增长。

 

认识到AR/VR在电影行业的潜力后,除了电影本身以外,影视制作公司日益为观众创造AR/VR体验。

 

例如:对于最新的《蜘蛛侠》电影,漫威开发了一款应用,让用户与角色自拍,试穿蜘蛛侠的服装,与角色互动等。《星球大战》、美剧《行尸走肉》等人气系列影视作品也有类似的应用。


对于《蜘蛛侠》、《星球大战》等经典卖座电影,制作公司意识到粉丝群在不断变化,因此希望利用AR/VR等新技术继续留住现有观众,并吸引新的观众。为此,VR在传统媒体和沉浸式媒体之间实现完美连接。

 

动画片方面,皮克斯公司最新大热影片《寻梦环游记》与Oculus合作,制作了电影的VR版本。


纵观AR/VR的未来,内容创作者仍然有很大空间去开发更具吸引力的内容。在SXSW,大量电影都展示了VR版本,让观众沉浸式体验电影。

SXSW上各种电影的VR体验▲



ARVR in the filming and healthcare industries







1

ARVR in Filming and Multimedia


This is one of the most popular topics at SXSW, where I attended several talks and demos. In the past year or so there’s doubt on whether ARVR reached the trough of the hype cycle, and whether it has its place in the market. During SXSW, it’s evident that not only it’s very much alive, the interests level and works gone into developing VR content continues to grow.


Recognizing the potential of ARVR in the film industry, film and TV production companies increasingly create ARVR experiences, in addition to the films themselves for audiences. 


For instance, for the new Spiderman movie, Marvel created an app where people can take selfie with a character, explore Spiderman’s suit, and have interactions with the characters, etc. Similar apps are created for popular franchises such as Star Wars, and TV series Walking Dead. 


For these long-time hits such as Spider-Man, Star Wars, as the fan base changing, production companies are conscious of it and hence hope to leverage new mediums such as ARVR to continue to engage existing audiences, and to attract new ones. For the purpose VR provides a good connection between traditional and immersive media.

 

For new animations such as Pixar’s recent hit Coco, the studio worked with Oculus to create a VR version of the movie.


Looking at the future of ARVR, there’s still much room for content creators to develop more engaging contents. At SXSW, numerous films were showcasing their VR version, to give audience immersive experience of the films.

A variety of VR experiences for films at SXSW▲



2

医疗前沿的沉浸式现实


会上讨论了AR/VR如何应用于医疗行业。美国两家领先医院——西德斯西奈医院和波士顿儿童医院分享了如何使用AR/VR来增强患者体验。

 

西德斯西奈医院使用由AppliedVR公司开发的VR内容,来测试如何帮助减轻患者(有时是得绝症的病人)的疼痛。SXSW上展示的VR内容包括:冰岛旅行体验,舒缓的瀑布或雨林景观,等等。参加试验的患者表示,VR体验确实有助于他们放松,并在一段时间里忘记了痛苦。

 

西德斯西奈医院正在与三星、旅行者保险公司(Travelers Insurance)等合作开发另一个类似项目——结合VR打造“数字疼痛减轻套件”,旨在帮助患者减少目前在美国盛行的阿片类药物依赖性。

 

AR/VR也可以成为患者接收和理解医疗数据的有效新方法。

例如:波士顿儿童医院也在与AppliedVR合作建立“虚拟现实指南”,帮助儿童了解MRI程序。此外,该医院还在与加拿大初创公司Klick Health合作,开发一款名为“Health Voyager”的应用,该应用从年轻患者的内窥镜检查记录中获取数据,然后转化为虚拟体验,帮助患者了解自己的身体状况。

 

由于MRI和内窥镜检查都可能是创伤体验,特别是对于幼儿,因此医院发现,有了VR作为辅助,不仅有助于减轻这些患者的焦虑,还能让他们更好地了解自己的检查结果。

 

AppliedVR的解决方案目前已应用于180多家医疗机构的5000多名患者。实践证明,该方案可以教育患者,帮助他们培养应对能力和放松技巧,以缓解疼痛。

 

AR/VR技术真正开拓了医疗行业的新边界。除了上述例子之外,AR/VR还在孤独症、中风和弱视等治疗中进行测试。在被问及应用方面的挑战时,小组成员表示,解决方案一旦部署到医院后,对其使用情况的监控,以及如何向付款方和医院管理方面证明其经济意义,这些都还存在困难。



2

Immersive Realities at the Frontier of Healthcare


This session discussed how ARVR is applied in a very different sector – healthcare. Two of the leading hospitals in the US - Cedars Sinai and Boston Children’s Hospital discussed how they are using ARVR to enhance patient experience.

 

Cedars Sinai using VR content developed by AppliedVR, to test how it helps alleviate pain from patients, sometimes terminally ill. The VR content, which were shown during SXSW, includes anything from an experience traveling to Iceland, to just soothing sceneries of waterfalls or rainforests. The patients participated in the trial reported that the VR experience really helped them relax and forget about the pain for the period of time. 

 

Another similar program that Cedars Sinai is developing in collaboration with Samsung, Travelers insurance, etc, is to incorporate VR as a ‘digital pain reduction kit’, aiming to help patients reduce Opioid dependencies, which is now an epidemic in the US.

 

ARVR can also be a new and effective way for patients to receive, and understand medical data.

One example is Boston Children’s Hospital also working with AppliedVR to build a ‘virtual reality guide’ to help young children understand MRI procedure. In addition, the hospital is working with a startup - Klick Health from Canada to develop an app called Health Voyager, which takes in data from young patient’s endoscopy records, and transforms it into a virtual experience to help the patients understand their own body.

 

As both MRI and endoscopy can be traumatic experience, particularly for young children, the hospital found that with VR as an aide, it not only helped reduce anxiety for these patients, but also enabled them to better understand their own results.

 

AppliedVR’s solution has now been used by over 5000 patients at more than 180 healthcare sites. They’ve been proven useful in educating patients, and helping them developing coping skills and relaxation skills to relieve pain.

 

ARVR technologies truly open up new frontier in the healthcare sector. Besides the examples above, ARVR has also been tested on therapy for autism, stroke, and treatment of lazy eye, among others. Though when asked about challenges in adoption, the panelists mentioned is the difficulty to monitor how the solutions are being used once they are deployed into hospitals, as well as how to prove the economics to payers, and hospital administration.



往期阅读:


复星锐正加注婚礼纪C轮融资 结婚细分行业领导者诞生

【盘点】那些就读湖畔大学的复星锐正系企业家们 创业之路“永不毕业”

开思汽配B轮累计融资4亿元,复星锐正超额跟投

国际资本潮涌东南亚,中国资本如何捕获“东南亚独角兽”?


    + 关注

    + 订阅

    阅读:378

    5

    精选留言

    扫描二维码推荐公众号

    微信公众号