铂链CEO 宋欣在斯坦福区块链峰会关于Bottos 人工智能新生态的演讲

旧金山时间3月1日16:00—21:30,北京时间3月2日8:00—13:30,铂链受邀参加在斯坦福大学举办的加密货币与交易所峰会,同时铂链在国内社区内进行实时图文转播。随着峰会的圆满成功,宋欣先生关于Bottos 人工智能新生态的演讲收到广泛关注,社区内图文直播也收到超过800名新铂粉的热烈反响。

此次会议由铂链与加密学之父Diffie院士和CPC联合在斯坦福大学举办,多位加密实验室教授、全球顶级交易所、顶级VC基金汇聚一堂。嘉宾包括:加密学之父、图灵奖获得者Whitfield Diffie,全球顶级交易所 Bittrex CEO Bill Shihara,币安联合创始人何一,全球首个获得公民资格AI机器人Sofia的开发公司Hanson Robotic CFO David Chen等。大咖云集、机会难得,又异常火爆、一票难求。

在会议上,铂链CEO宋欣提出了”人工智能新生态”的想法,这一新颖的想法已经提出就引发热烈讨论,宋欣先生说:”我们提出的人工智能新生态分为三层——底层为铂链公链,我们在这里设置共识、P2P网络和安全加密;第二层是服务部分,可以在那里找到标识、分布式事务、存储、智能合约、数字资产管理和信誉模型。在顶部是dAPP的一层,属于分布式人工智能经济,数据和模型市场以及存储/算力共享网络。”

此次会议是全美乃至全世界优质数字资产、交易所及学术的激烈碰撞,对于业内顶级合作伙伴之间交换关于加密技术的想法有着非常重要的意义,同时也为新的合作创造更多新可能。

在此,铂链附上宋欣先生演讲全文,欢迎广大铂粉及区块链、人工智能、加密数字货币爱好者共同讨论人工智能新生态!


宋欣先生演讲全文如下:

Good afternoon, everyone. Thanks for coming. Thanks for the speakers. I just learned a lot from your speeches. Hope you can learn something from my presentation.

大家下午好,非常感谢大家的到来。感谢所有的演讲嘉宾。我从大家的演讲中收益颇丰,同时也希望我的演讲为大家带来一些收获。

My name is Xin Song. I’m the CEO of Bottos project. Today I’m introducing you a new AI ecosystem based on blockchain technology, the Bottos platform. I’ll talk about three parts. First, why we do Bottos. Second, our understanding how blockchain can help AI. Third, the Bottos details and highlights.

我是铂链的CEO宋欣。今天,我将向在座各位介绍一种新的基于区块链技术的人工智能生态系统。我的演讲将分为三个部分:首先,我们为什么要做Bottos;其次,我们如何理解区块链对人工智能的帮助;第三,我将详细介绍铂链和它的亮点。

OK. Let’s talk about the first part. Why did we do Bottos. The story started 3 years ago, we started a project called Scream, an exoskeleton AI product, a wearable AI medical device helps you walk and recover from accident injury. The idea was great. The computing power was not an issue. The model was also excellent. The valuation VC gave was 30 million USD. But the product is still a kind of a toy until now. Cannot be commercialized. The Key reason for this was lack of training data. Not enough data provided to train the model. Because the data is too expensive and not enough people have the incentives to provide the data. Only the industry big guy has the data. And this is a typical pain point in the AI world. Then we met Blockchain last year and we believed Blockchain can great help with this issue. So we get into the blockchain area with the attention to solve the data sharing issue but we found blockchain can certainly do more to help AI. And that’s what Bottos will do in future. To form a new AI ecosystem based on block chain.

那我们现在开始进入正题,进入第一个部分,我们为什么成立铂链?故事要从三年前说起。三年前,我们成立了一个叫”尖叫”的项目,产出机器人外骨骼人工智能产品,它是一个可穿戴的人工智能医疗设备,可以帮助从事故伤害中造成的伤残情况恢复。这本是一个好想法,计算能力不是问题,我们也成功建立了很好的模型,当时的VC的估值为3000万美元。但是,产品至今仍是一种玩具,无法商业化。造成这种情况的主要原因就是,缺乏培训数据,没有足够的数据来训练模型。因为数据太贵,而且没有激励人们提供数据的方式,造成了行业巨头对数据的垄断。这是人工智能领域一大痛点。在去年,我们接触了区块链。我们相信区块链可以有助于解决这个问题。所以,我们进入区块链行业,并致力于用区块链解决数据共享的问题,区块链技术也将有效帮助人工智能。这就是铂链在未来即将做的事情——基于块链形成新的人工智能生态系统。

Then we take a look at how Block chain can help AI. So we all know the future is an AI world. Everything is somehow related to AI sometime in future. However, there is still a long way to go. We see more and more AI applications in the real world but the implementation of AI is much lower than expectation. We think there are three key factors in the AI development. The computing power has been improved over the decades thanks to the upgrade of hardware capability, the cloud computing and network technologies. Generally speaking it’s not a significant issue in terms of capability, but still have much room to reduce the cost. Second is the algorithm, the development of genetic algorithm such as deep learning and multi-level learning network has been improved and open sourced. The issue of this part is no secure open market for the model distribution and communication. The last part is data, we just talked about. We think data is the most important part of the AI development. You can only build a good AI product based on rich data, no other way. But the data is expensive and difficult to access. We need to have a way to encourage people to share data with incentives and protect their right and privacy at the same time. Overall, the issues in AI development can be concluded as how to reduce the cost, how to encourage people to share resources and how to protect people’s right and privacy. How can we solve those issues? We found block chain can help.

然后我们看看区块链如何帮助人工智能。我们都知道,未来将是一个人工智能世界。在将来,一切都与人工智关联。然而,目前来看仍有很长的路要走。在现实世界中,我们看到越来越多的人工智能应用,但人工智能的落地情况比预期的要低得多。我们认为人工智能发展有三个关键因素。由于硬件能力、云计算和网络技术的升级,计力在过去几十年中得到了巨大改进。一般来说,这不是急需解决的问题,但仍然有很大的空间来降低成本。二是算法,深度学习和多层次学习网络等遗传算法的发展推进了开源代码的发展。这一部分的问题是模式分发和开放市场的安全问题。最后一部分是数据,这是我们刚刚谈到的。我们认为数据是人工智能发展最重要的部分。基于充足的数据,只能构建一个好的人工智能产品,没有其他扩展。但数据昂贵且难以获取。我们需要有一种方法,可以鼓励人们分享数据,同时保护他们的权利和隐私。总的来说,人工智能开发中的问题可以归结为如何降低成本、如何鼓励人们共享资源以及如何保护人们的权利和隐私。我们如何解决这些问题?我们发现块链可以有所帮助。

What is blockchain. Many people are talking about it. Let me summarize it quickly. It’s decentralized, consensus based, immutable ledger and smart contract based with token incentives. What does that mean to AI? First, blockchain helps to reduce cost. Because it’s decentralized, creating trust for the people who do the transactions. Reducing the cost by doing transactions without a centralized organization, which is always costly. Second, provide enough incentives through tokens for people to share resources, could be data, could be algorithm, could be computing power, could be storage space, whatever is required for AI development. Thrid, it’s secure and help to protect the people’s right and privacy. Each model, all the data, all the transactions could be recorded can almost cannot be changed. The identification and ownership is clear. So Blockchain provides a whole solution for AI issues. Reducing cost, encouraging resource sharing and providing ownership and privacy protection.

区块链是什么?大家众说纷纭。让我快速总结一下:它是分布式的、基于共识的、不可篡改的分布式记账和基于数字资产激励的智能合约。这对人工智能的意义是什么呢?首先,区块链可以帮助降低人工智能获取数据的成本。因为它是分布式的,交易双方互相信任。通过去中心化的交易来降低成本,而这一环节在中心化环境中是非昂贵的。其次,通过数字资产提供足够的激励,让人们共享资源,可以是数据,可以是算法,也可以是算力,也可以是存储空间,是人工智能开发所的一切需要。再次,它是安全的,有利于保护人民的权利和隐私。每个模型的数据,所有的信息都可以记录几乎无法篡改。所有权和版权确权。所以,区块链为人工智能急需解决的问题提供完整的解决方案:降低成本,鼓励资源共享,提供所有权和隐私保护。

Take the exoskeleton project as example. You issue a token to the people who provide the algorithm training data. People have the incentives to share data because their ownership and privacy is protected. In addition, the token value could appreciate in future. People would like to share. When the model is trained well we could publish the model in the blockchain based model marketplace so people could pay the token to use the product. The project could get cheaper computing power and data storage space in the blockchain, too. So the Blockchain reduces the cost for data/computing power and storage space acquisitions. The training data and model is clearly registered in the chain so the ownership is clear. The project can also benefit from the token value appreciation because the application of the model is mature. Amazing, right?

以机器人外骨骼项目为例,你向提供算法训练数据的人员提供数字资产,人们就有了分享数据的动力,因为不仅所有权和隐私受到保护,数字货币价值在未来也会升值。人们就会愿意分享。当模型进行训练,我们很可能会在区块链模型市场发布模型,使人们能够通过支付数字货币使用产品。人工智能项目能通过区块链得到更便宜的算力和数据存储空间的。所以,区块链降低了数据的算力和存储空间的成本。训练数据和模型在链上也能清晰地标注,因此所有权也是清晰的。基于模型的应用已经成熟,该项目也能受益并增值。

Then let’s take a look at what Bottos is. A public chain specifically designed for AI development. AI could use data market place to get training data, use model market to distribute the model and train the model, use storage and computing power sharing network to store data and retrieve computing power. Moreover, each AI could issue their own token based on Bottos to realize the value exchange among the data provider, model user and storage/computing power sharing people.

接下来我们来看看铂链是什么?它是一条为人工智能的发展设计的公链。人工智能可以在数据市场上获取训练数据,运用存储和算力共享网络存储数据并检验算力。此外,每个人工智能可以基于铂链发行自己的数字资产实现去实现数据提供方、模型使用者以及存储/算力分享的价值交换。

This is the structure of Bottos. I like this fansion graph. It helps you to remember Bottos. In short, three layers, the bottom one is the Bottos public chain, we have the DPOS consensus method, P2P network and security cryptography there. Second layer is the service part, you find the identity, distributed transaction, storage, smart contract, token management and trust & reputation model there. At the very top is the dapps layer, the decentralized AI economy, the data and model market place and the storage/computing power sharing network. Basically AI can find every technical product in this layer.

这是铂链的结构。我很喜欢这张新颖的图片,它可以帮助你很好的了解铂链。简单来说,分为三层:底层为铂链公链,我们在这里设置共识、P2P网络和安全加密;第二层是服务部分,可以在那里找到标识、分布式事务、存储、智能合约、数字资产管理和信誉模型。在顶部是dAPP的一层,属于分布式人工智能经济,数据和模型市场以及存储/算力共享网络。人工智能可以在这一层找到自己所需的科技产品。

What can Bottos do for AIs. We summarize it in four parts. In addition to the platform product we just described, we also provide technical support for you to build your AI product based on blockchain. We provide marketing support for promotion in Bottos community. Globally there are over 130,000 people in the Bottos community, extremely strong in China. Last, we also invest in the AIs collaborating with Bottos, especially with the ones issue token based on Bottos. That’s how we build the full ecosystem. You must have a token to exchange the value.

铂链可以为人工智能做什么?我们总结了四个部分。除了我们刚才所描述的平台产品,我们也会为基于区块链技术的人工智能产品提供技支持。我们在社区内提供市场支持。全球有超过13万人在铂链的社区里,尤其在中国。最后,我们还会投资于与AI公司的合作,尤其是与基于铂链的AI项目。这就是我们如何建立完整的生态系统的过程。前提是你需要数字资产来交换价值。

Well, this is the brief introduction of Bottos as well as our understanding of the connection between AI & Blockchain. Bottos will have product version 2.0 released by the end the March and mainnet released in May. Follow us on Telgegram and Twitter. You will get all the updates about Bottos.

以上就是我对Bottos的简要介绍以及我们对于区块链与人工智能之间的关系的理解。铂链将会在三月底发布2.0版本产品并在五月发布主链,大家可以在tg及推特上关注我们并获得实时更新。

Thank you very much!

非常谢谢大家!