Decentralized Machine Learning (DML) - Unleash untapped private data, idle processing power and crowdsourced algorithms
Name: Decentralized Machine Learning
Project Type: Platform
Project Description: Decentralized platform where users (Data providers) allow DML to securely access their untapped data with the purpose to provide new algorithms.
Token Symbol: DML
Token Type: ERC20
Early Whitelist Program: Feb 19th, 2018 – Feb 24th, 2018 (Ended)
Whitelist: Feb 25th, 2018
ICO Date: Mar 15th, 2018
Company: Kyokan Labs Team
Industry: Artificial Intelligence & Machine Learning.
The Machine Learning Industry has been increasing substantially; the International Data Corporations have forecasted revenues up to USD$200 billion by 2020. Nonetheless, the industry seems to have its limitations due to the way companies are currently performing the machine learning development. It is at this point where DML offers its innovative decentralized platform. Decentralized Machine Learning is willing to open a new way of “teaching” by eliminating the barriers the Machine Learning Industry is facing. Their main value proposal aims to provide access to untapped private data with the user’s consent.
DML will achieve this through an app that will securely collect the data and will function with the processing power of the device that is not being used by any other app (Idle).
People, who are willing to provide an access to their private data in exchange of money, are one of the key factors for DML’s success. However, DML does not have a base or projection of those potentials users. What will define if the project is a success or a failure is the number of adoptions the idea would get.
In the same way, DML has not made any partnerships yet, which could boost the idea for future spread of their products. A strong strategic plan is needed to reach the potential data providers’ market through diverse marketing campaigns and to make it feasible.
The biggest threats are the top techs companies. Nowadays, there is a kind of oligopoly in the Machine Learning Industry due the high investment that top techs can provide to the research and development of this industry. The fact that this is the only way to afford this development (but with a decentralization as a lower alternative for the industry) opens the door to new companies. Likewise, it is a great alternative to the top techs companies as well.
On the other hand, in the cryptocurrency world there are also many projects (such as AF Security & SynapseAI) that just like DML are developing alternatives aiming the boost of the Artificial Intelligence Industry through their rewards plans for data providers and transfer value for developers.
Nowadays, in the digitalized world that we’re living, every taken action within the Internet world leaves a footprint, which is massively stored as data. Businesses are taking advantage of this, collecting this data to offer their customers more suitable products or services. Due to the Big Data revolution, the Artificial Intelligence development has increased considerably. The reason of this is that the only ways developers can “teach” the machines is through following algorithms patterns and make a given decision; this method is called: Machine Learning.
Even though we have seen a great advantage due to the Big Data, this is only the top of the iceberg. The Big Data collects all the public information that each user leaves unaware, like the cookies that our Web Browser saves from each website we visit, but it does not collect all the information any user could provide.
Consequently, the biggest advantage in the Machine Learning development would be to attain when developers access to untapped data; this being the private data each Internet user has in their device. This could provide a valuable insight; however, accessing to this data will provoke an overwhelming amount of information that regarding of the limited processing power that centralized computers clusters have, they will not be able to handle.
Even though we might see the centralized hard-wares as inefficient, compared to the potential data, it requires a huge investment to build a computer cluster that only some large tech companies can afford. As consequence, top tech companies are building an oligopoly around the industry, being the only ones who can run the business.
DML’s mission is to develop a Decentralized Machine Learning Protocol that solves the current and forthcoming limitations the industry could face through its app. This app can be installed by thousands of users and will provide their private data to be collected directly from their personal devices; and at the same time, preventing any possible leak during the extraction of the information. Moreover, this app functions thanks to the processing power of each device, specifically using the power that is not used for any other app (idle). It also eliminates the limited processing power that the centralized systems have in their computer cluster. This is the easiest way to untappe the desired private information with the solely purpose of recollecting it to create new algorithms that allow the Artificial Intelligence Industry to give a big step forward.
The resulting and collected algorithms would be stored in a Federated Node where all connected devices would transfer them encrypted to an agglomerated storage, leaving the “raw data” in the owner’s device. Furthermore, these results are shared in the DML marketplace, where stakeholders who want to run any analytical predictions can access and acquire the algorithm that is best suitable for their purpose. As any business, DML needs to have economical exchange. For this purpose, DML has its own token, which allows the data providers to get paid for contributing to their private data and the use of idle processing power of their device. Also, stakeholders could use the DML tokens to acquire algorithms for their personal manners. Last, another utility this token has is to reward those developers who contribute with their effort and knowledge the DML platform.
Victor Cheung - Blockchain Developer
Michael Kwok - Project Lead Director
Founder of 2 tech companies with startup winning awards Seasoned growth lead in early stage startups, veteran digital marketer and community manager Specialty in business development. Online branding and SEO expert
Jacky Chan - Blockchain and Software Developer Founding engineer at blockchain consultancy firm Kyokan Labs, contributor to Metamask & DFINITY Former software developer at Uber Early engineer at Symphony Communication acquired by Goldman Sachs, serving Wall Street firms
Pascal Lejolif - Machine Learning Engineer Former CTO at Alkia IT Services, specialized in cloud computing, AI and cyber security services Over 8 years as Alcatel-Lucent Enterprise technical project manager Machine Learning Certificate at Stanford University
Wilson Lau - Machine Learning Engineer Technical manager of cryptocurrency mining consulting firm Involved in machine learning, agriculture robotic and system design Proficient in Python, C, Mathematica and Arduino
Patrick Sum - System Security Engineer Over 20 years of IT experience in banking, education sector and semi-government authority Software development, system administrator, project management. Wide-area network design expert MBA, AGSM BSc (Computer Studies), HKU Kyokan Labs Team
Matthew Slipper - Machine Learning Engineer Co-founder and CTO at Kyokan Labs. Contributor to Machinomy Former co-founder and CTO at Spectrum Labs, an investor-backed machine learning startup Early engineer at Symphony Communication, serving Wall Street firms
Alex Tseung - Software Developer Over 4 years at Cisco Systems as hardware validation engineer, and 2 years at Avi Networks as frontend engineer First engineer at Kyokan Labs, contributor to Metamask MSEE, Santa Clara University BSEE, UCLA
David Yoo - Software Developer Software engineer at Uber Led efforts in building user facing features for mobile and web applications Former full-stack engineer at an e-commerce startup for 3 years, brought progressive design and frontend expertise
Guillaume Huet - Big Data/Machine Learning Advisor 12 years in banking, consulting and business development in Europe, Africa and Asia Frequent speaker on data science and machine learning MIB, EDHEC Business School Specialization Certificate in Machine Learning, Stanford University
Google Federated Learning:
Collaborative Machine Learning without Centralized Training Data. Google, with its robust and secure cloud infrastructure is willing to interact with mobile devices, with the purpose of collecting learning data for their prediction model while keeping all the raw data in the device.
Kaggle: The Home of Data Science & Machine Learning Kaggle is a platform that encourages developers to complete Machine Learning Tasks through competition. Datasets are given to the developers and each one of them has to build a Machine Learning Model and post it in their profile, then others give them their reviews.
Top coder: Deliver faster for your business through crowdsourcing Top coder is a developer’s community with more than 1,000,000 experts. It provides on-demand solutions with the proposal of velocity and great capability of development.
Synapse AI: Decentralized AI Network Like DML, Synapse AI rewards their users while they transfer value to the developers through the access to their data.
Taking into consideration the limitations that the Machine Learning Industry is facing nowadays, Decentralized Machine Learning’s ICO arises as a strong potential alternative. With its innovative proposal, DML’s ICO looks forward to lower the barrier and let an entrance to the industry to savvies developers who are willing to contribute to its development but don’t have the capital to invest in those expensive hardware’s. Also, DML looks to create a milestone in the Machine Learning Development by getting access to private information, which is the body of the iceberg. This would provide thousands of new algorithms that will help with the accuracy and boost of the models used for predictions. .
Nevertheless, the Machine Learning Industry is being held by the large technological companies and nothing seems to indicate that this could change. It is true that the decentralization method is the alternative to continue industry’s development, but it is also true that the large companies will migrate to this method because of its lower costs. As we mentioned before, Google is developing its Federated learning system, and so on the biggest tech’s company will follow their path.
To conclude, in a potential profitable industry, there is a chance for everyone to take a piece of the cake, whether it is a big one or a small one.
In this case, with DML’s ICO, we consider that everyone could just take a small one. The main reason we consider this is because in an industry where top techs are running, they will always have the bigger scope, regarding their brand awareness, customer’s attachments to the company, infrastructure or wealth. Therefore, depending of the risk that each investor is willing to have, they could or could not invest in this project. A risky investor could invest willing to resale its share in any high fluctuating cryptocurrency market, or even just before any drop. On the other hand, a conservative investor would rather not invest against the big techs.
You can buy DML on the above exchanges in the DML/BTC, DML/ETH, DML/BNB (Binance) pairs.
Easily Calculate or Convert DML to US Dollars(USD), Euro(EURO) or Great Britain(GBP) Pounds and other World currencies below.
Our Crypto Converter makes this as easy as a breeze.
This Wiki is provided courtesy of Crypto cue Magazine
Price Prediction: Price Prediction