cool hit counter Cortex: Taking AI from virtual "Westworld" to real "community experiments"_Intefrankly

Cortex: Taking AI from virtual "Westworld" to real "community experiments"


Set in a futuristic world, a huge high-tech adult theme park where highly anthropomorphic AI "receptionists" allow visitors to indulge their lust, violence, and killing desires on robots. To make the human experience more realistic, programmers continue to make AI robots have richer emotional experiences by writing new code until some of them experience an awakening of self-awareness...

Westworld is about the "moral dilemma" that humanistic ethics shuns, but the "technological realization" and "community experimentation" that geeks spend their lives pursuing.

Are you sitting still? It's time to drive again.

As one of the projects in the bear fever retrograde, the official Cortex telegraph group has close to 100,000 people in both English and Chinese combined. The technical threshold is extremely high (referred to as "unintelligible") white paper, the core technical team graduated from the world's top schools (about Tsinghua, Shanghai Jiaotong University, Columbia University level), the advisory team led by Turing Award winner and "father of modern cryptography" Whitfield Diffie, covering investors from 20 mainstream cryptocurrency investment institutions such as Bitmain, FBG, Nirvana Capital, True Capital, etc. The legend is that it does not accept investment from domestic individuals and institutions but has prepared high-value airdrops in the community... Cortex has come out of nowhere with a geeky, mysterious, emotional and pragmatic approach, "plotting" a cool "AI community experiment".

Before looking at the project, a small egg is released.

The highlights are self-searching and are revealed at the end of the article.

Smart contracts + AI = on-chain inferred consensus

Cortex's vision is to provide state-of-the-art machine learning models and build a machine learning community platform on the blockchain, and by modifying and extending the instruction set, Cortex adds AI algorithm support for smart contracts so that anyone can add AI to their smart contracts. At the same time, Cortex proposes incentives for collective collaboration, allowing anyone to submit and optimize models in Cortex, and offering rewards for model contributors. Due to its open and shared nature, Cortex will create many models that go beyond human capabilities. Also, as a (geeky) "social experiment", the project also expects to give birth to "Artificial General Intelligence" (AGI) on Cortex.

Unlike other AI on-chain projects, Cortex has two innovations: first, AI is put into a smart contract to form an inferred consensus, and all nodes agree based on the inferred results, thoroughly ensuring blockchain autonomy; second, the trained data model is put into the public chain storage layer, and members in the community can pay tokens to use it, benefiting more community development.

Rant: Here I need to add a little background knowledge, artificial intelligence/machine learning is to refine big data, using arithmetic iteration to continuously find reasonable parameters to get the data fit optimization model, with this model can be used to generalize the data inference, so the developed model needs to go through very many data fit optimization "learning" before "mature" for analysis. "Learning" data is confidential and requires a lot of iterations, while blockchain's properties are an important helper in solving machine learning. Most of the many projects combining machine learning and AI with blockchain are doing on-chain training or data sharing sessions, the Cortex project is positioned to solve the problem of inferring consensus for machine learning models on the blockchain, and although based on the machine learning community to build, Cortex itself is focused on smart contract scalability and model storage. Adding AI inferred consensus mechanisms to smart contracts brings autonomy to the in-chain ecology, a project that, if realized, is believed to bring a considerable technological impact to the existing Ethernet architecture.

Tsinghua Academic Group + Father of Modern Cryptography + Twenty Mainstream Blockchain Investment Institutions

Founder Tzuki Chen studied at Tsinghua University, Carnegie Mellon University, and the University of California, Santa Cruz. During his early learning experience, he studied machine learning theory and various algorithmic applications, including Go algorithms, under David P. Helmbold. Proficient in consensus algorithms and public chain ecology, he leads a team of companies with first-line experience in developing mining pools, multi-digital asset wallets and mining software.

CTO Weiyang Wang studied at Tsinghua University and the University of Chicago, and learned statistical-based machine learning under Peter McCullagh, the founder of Koppers Prize Logistic Regression in 1989. OCR/image classification/NLP expert using deep learning, Top 1% in the world in Kaggle competition and won silver medal, leading research and code development for multiple cutting-edge papers in deep learning, leading 4 projects on OCR/migration learning/face/distributed computing, entered Awesome MXNet boutique project. Proficient in number theory, cryptography and other related fields.

Yan Liao and Yang Yang are the chief engineer of deep learning and the chief engineer of blockchain respectively, both with high achievements and development capabilities in their respective fields.

The biographies of these backgrounds are well documented. It is one of the few project teams that have frontline development experience in both AI and blockchain and have achieved excellent results that can be documented.

Consultant.

Investors.

Rant: Although the white-bearded "father of modern cryptography" grandfather has recently appeared frequently in the white paper of popular projects, so many people joked that "the Turing Award winner also want to walk to earn money", but look at the core members and the square team of the same investment institutions, but still for this "top of the fighter" like the team of heartfelt respect, and the confidence of the project on the ground.

Smart Contracts with AI + CVM

The overall technical plan consists of two parts: AI-integrated AI contracts, which allow users to write machine learning programs on the blockchain and enable interaction with other contracts; and the Cortex virtual machine, owned by the project itself and fully compatible with the Ethernet virtual machine instructions.

The specific solutions are as follows.

Cortex inferred consensus basis: when a user initiates a transaction to a contract, the full node needs to execute the code of the smart contract. The difference between Cortex and normal smart contracts is that Cortex's smart contracts involve inferring instructions, and then all nodes need to agree on the outcome of this inference.

Model submission architecture: Cortex also provides an off-chain submission interface, including a model instruction translation virtual machine. This will create a bridge for transactions and collaboration between providers of computing power and providers of algorithms.

Cortex Intelligent Inference Framework: Machine learning researchers around the world can upload "trained" corresponding data models to the storage layer. Other users who need such AI models can make inferences and pay providers by using the models.


Recommended>>
1、Fantasy westward travel reproduces wei shao level tycoon multiple service warfare number spend more than ten million unmatched
2、Blockchain smartphones can be financially empowering for users
3、Stantec selected as the first batch of Jiangsu Province industrial Internet service resource pool units
4、The heavy new cars you cant miss in 2018 Import Dirtbag Edition
5、Color TV industry to fundamentally transform highend

    已推荐到看一看 和朋友分享想法
    最多200字,当前共 发送

    已发送

    朋友将在看一看看到

    确定
    分享你的想法...
    取消

    分享想法到看一看

    确定
    最多200字,当前共

    发送中

    网络异常,请稍后重试

    微信扫一扫
    关注该公众号