Ethereum machine learning

AI, Machine Learning and Ethereum Blockchain Come Together to Offer Maximum Profit

4 stars based on 34 reviews

At my company we are starting an experimental project to extend EVM with basic deep learning capabilities. This is not to train a neural network, but to use a ethereum machine learning neural network inside a smart contract.

Computation-wise using a pre-trained neural network is actually not so much more expensive than doing, say, RSA. I understand that this may be a bit too heavy for the official Ethereum blockchain, so in our case we will run the EVM on a separate permissioned cluster with BFT-like consensus.

A pre-trained network is saved on the blockchain. We can use some of existing neural network serialization ethereum machine learning such as the ones used in Keras framework. In the simplest case there we add a single predict instruction similar to predict from Keras framework.

This instruction will take a fully qualified name of the neural network and an input data array, run the neural network and produce output data. As an example input data could be an English-language string, and output will be a German translation of this string.

One problem that we will need to solve in the process is introducing deterministic floating point numbers such as IEEE into the EVM ethereum machine learning some way.

If there are other people interested to run AI on EVM, we would be willing to cooperate on this to establish a standard that everyone uses …. Can I ask what the goal of this ethereum machine learning Running deep learning inside the EVM is probably not the solution. You might want to take a look at this approach: Basically the idea is to use a rational activation function and encoding the trained neural network inside a zkSNARK circuit.

You can build it so that the weights are part of the private input and then basically have a smart contract confirm the result of a neural network applied to a certain input while not revealing the actual network weights. An application ethereum machine learning this would ethereum machine learning verifying results to a competition that are ranked using a neural network.

The reason why you might not want the weights to ethereum machine learning public is that it ethereum machine learning usually easy to fool a neural network once you know the weights.

The only problem with this approach is that the person generating the proof has to know the internal weights. Well - I think many people have expressed different ideas about running a neural network as a trusted application so that all parties agree to the outcome.

As a toy example, you can consider an example of a smart contract that is an decentralized Uber which pays to drivers based on their behavior. The smart contract needs to differentiate bad drivers from good drivers by running ethereum machine learning neural network on driver historical behavior.

Good drivers are get paid and bad drivers do not get paid. Another example is when a smart contract buys apples from a supplier and it needs to find out whether an apple is a good apple or a bad apple ethereum machine learning on chemical analysis data.

Essentially you feed into the neural network data points of chemical analysis and the network tells you whether the apple is good or bad. In general, I think to answer your question, anytime a smart contract runs on data, there is no need to use neural networks. On the other hand if a smart contract runs on Big Data, then, arguably, you need a neural network to extract important info from the data, as the data itself is too complex.

In the examples above the neural network does not need to be confidential or ethereum machine learning in anyway, it ethereum machine learning be a pre-trained network which is trusted by all participants. Another possibility is to run a totally different thing and then feed the results into Ethereum somehow …. I mean that the trained network will have a very low probability in being perfect from a passengers point of view.

Then I guess either training would have to be done on the blockchain somehow, or users should regularly be allowed to vote for a new trained network or training scheme to be used. This could also apply to the apple sale scenario I guess, but on a much larger time scale. Like the apple salesman tampering with the genes through generations to maximize the profit he would get form the network on the blockchain but minimizing his costs in production. Agree ethereum machine learning actually Uber drivers are getting really bad - have you noticed that if they want you to cancel sometimes they keep on moving their car slightly to fool the algorithm?

I think if they just do not move the car for 5 min after they accept they are automatically penalized. This example shows though, that simple algorithms are easy to fool, a machine learning algorithm is harder to fool.

Actually the best way to find how to fool a neural network is to run another neural network called adversarial network. I personally think this makes sense. The weights of the network can then be updated through a smart contract call itself. However, even though the weights might be completely in the open, the fact that you as the owner of the contract is the only one allowed to train the network and upload the weights, means that the people interacting with this contract will need to trust your good intentions.

In the case of the decentralized Uber ethereum machine learning the driver and the passenger will need to trust you as a company to train a balanced model. Another point is that neural nets can have millions of parameters. Updating such a network often would be very costly. Do you think it could be fixed by a procedure like that: How to form this body then? Should there be some kind of a token for that? Imho governance is a huge question - I was born in a country USSR which tried a communistic experiment - it failed profoundly!

In the USSR taxi drivers were ethereum machine learning advantage ethereum machine learning passengers in all possible ways, in addition to that they were stealing gas all the time.

A simplistic approach, where you pass all the power to the drivers will probably fail, you need to take into account both the drivers and the passengers somehow, the question is, is there a ethereum machine learning governance model to do it?

Usually if one retrains deep neural nets, all parameters change. I guess techniques such as quantization ethereum machine learning become very important in this scenario.

Interesting - never used it before. My understanding of the idea is that you cant train in low precision, but once trained, you can quantize to ethereum machine learning precision and it will still work …. I mean this is only an issue in usecases like the dUber case, and to be honest the more I think about this, the further I go down the rabbithole, and the less it has to do about the specific issue of doing a forward pass in a neural net.

If you could somehow verify that the network is trained on a set of transactions…. Why not solve it in a simpler way: To work on a job, you will need to stake some amount of ether When creating a job, you ethereum machine learning a swarm hash for the set of data to be predicted upon.

Along with this hash, the sender also submits a salted verfication set A number of nodes then perform the prediction and salt their result hash, which is merkle tree encoded it needs to be a quantized result to avoid differences in ethereum machine learning point implementations When enough nodes have submitted their hash, they reveal their hash.

Any node that tried to cheat is robbed off their staked eth. The rest split the reward for the computation. To me, it seems like an off-chain solutions like Truebit could be a better way of tackling this problem. If you create multiple nodes and submit the same answer multiple times, you will get a bigger piece of the pie. A more sophisticated version of your proposal by kladkogex was discussed in this thread and seems to encourage centralization.

The problem is that there is currently not a single person ethereum machine learning the world that can mathematically prove TrueBit security. Ethereum machine learning wonder if it would be feasible to take something like keras. Keras is a Deep Learning framework that uses TensorFlow behind the scenes. I wonder if the same approach could work for Ethereum. So you take a trained Keras model which essentially is just a huge bag of numbers and then run a couple of matrix multiplications and geometric functions over them to get the prediction.

This is basically what keras. Just the fact that the more complex the model is the more gas would it take to run the predictions. Chris - thank you, I was aware of Keras but not aware of Keras. Keras is definitely the leading library now for abstracting out deep learning implementation details. The question is how whats the MVP implementation to add a Ethereum machine learning.

Experiments run on our cloud infrastructure though. But even without that it should be possible to just come up with an MVP state solidity library that does the basic math on top of todays Ethereum machine learning. So the way Ethereum machine learning understand it, one could build a keras-solidity library today even though it may be inefficient unless further opcodes are added to improve the speed of some underlying mathematical operations.

The idea to integrate support for neural networks into Ethereum recently started to rotate in my head too, glad to have found this thread - kladkogex please keep us updated!

But the point is to use models ethereum machine learning off-chain and only do inference on-chain which is much cheaper.

Ethereum machine learning way it should be possible to run these computations more efficiently e. It may be a bit of a challenge to guarantee deterministic results, ethereum machine learning quantization may be a good idea not only for performance.

The EVM would only need a mechanism for making a call to the neural network runtime and handle its result. Putting an AI in charge of distributing economic rewards this could be tokens created out of thin air sounds like a pretty powerful possibility to me…. The current plan is that: We can use some of existing neural network serialization standards such as the ones used in Keras framework The EVM will need to pull the neural network from the blockchain. Re-training on a decentralized network would require GPUs - could also ethereum machine learning done though ….

I do think doing this just for the purpose of doing it has merit though. In the USSR taxi drivers were taking advantage of passengers in all possible ways, in addition to that they were stealing gas all the time A simplistic approach, where you pass all the power to the drivers will probably fail, you need to take into account both the drivers and the passengers somehow, the question is, is there a sustainable governance model to do it?

Keras is definitely the leading library now for abstracting out deep learning implementation details I agree. After all, you can do matrix multiplications by hand:

Buy bitcoin coinbase vs

C rotate right bitstamp
Buy ledger blue

Icon pack robot android like ios

Bitcoin projection chart

Robot chicken indiana jones raiders of the lost ark online free
Profittrailerder bitcoin trading botswana
Bitcoin trading tutorials

Bitcoin ethereum bitcoin cash ripple stellar litecoin cardano neo eos price analysis march 8

49 comments

Dogecoin solo mining calculator event

Machine Learning algorithms are being developed and improved at an incredible rate, but are not necessarily getting more accessible to the broader community. DanKu enables anyone to get access to high quality, objectively measured machine learning models. At Algorithmia, we believe that widespread access to algorithms and deployment solutions is going to be a fundamental building block of a balanced future for AI, and DanKu is a step towards that vision.

The DanKu protocol utilizes blockchain technology via smart contracts. The contract allows anyone to post a data set, an evaluation function, and a monetary reward for anyone who can provide the best trained machine learning model for the data.

Participants train deep neural networks to model the data, and submit their trained networks to the blockchain. The blockchain executes these neural network models to evaluate submissions, and ensure that payment goes to the best model. The contract allows for the creation of a decentralized and trustless marketplace for exchanging ML models. This gives ML practitioners an opportunity to monetize their skills directly.

It also allows any participant or organization to solicit machine learning models from all over the world. This will incentivize the creation of better machine learning models, and make AI more accessible to companies and software agents. Anyone with a dataset, including software agents can create DanKu contracts.

For more info, please refer to the last section regarding the competition below. As of , Artificial Intelligence and Blockchain continue to dominate tech news everywhere. Earlier in , we asked ourselves if we could brings these two things together and solve a problem in Machine Learning.

We immediately noticed a diversity of ideas, where solutions were proposed for all kinds of problems. One good example is OpenMined , which allows you to train ML models on data that you never have access to. Since putting Blockchain and AI in the same sentence sounded very click-baity, we decided early on to show and not only tell. This led us to focus our efforts on a narrow problem definition. And the idea for DanKu was born: Trustless machine learning contracts for exchanging machine learning models on the Ethereum Blockchain.

Bob creates a new DanKu contract. He submits a dataset, an evaluation criteria, and a reward amount to the contract. ML practitioner Alice downloads the dataset submitted by Bob, and works independently to train a ML model. After successfully training a model, Alice submits her solution to the contract aka. Other participants like Alice can also submit their solutions. Bob reveals the testing dataset after the submission period ends. The testing dataset will be used for evaluating the submissions.

At some point in the future, the blockchain will evaluate the submitted models, and payout to the winning submission. If no submission fulfills the criteria, the reward is refunded back to Bob. Bob and other participants just exchanged ML models in a completely trustless manner. The contract also ran a fully functional ML model on the blockchain too! Some additional steps are necessary to ensure the trust and fairness of the competition.

Refer to the white paper for more details. For demo purposes, we decided to write a DanKu contract where a neural network was implemented. Solidity, the programming language for Ethereum contracts, was not designed with Machine Learning in mind. It did not have a math library, or even floating point numbers. Ethereum also had issues running code that was too computationally expensive.

It was a software engineering nightmare. Initially, we started with something simple: After gaining a bit of confidence, we tackled the problem of implementing a neural network from scratch. We initially started with a neural network with no hidden layers. We developed everything locally, until we were ready to test it on a live blockchain. We started off by testing it on the Ethereum Ropsten test net.

Everything worked without much of an issue. Afterwards we wanted to try it on the Ethereum blockchain to make sure it worked properly on the real thing. What looked like machine code to these everyday miners, was actually a fully functioning neural network. Feb 15th was a good day, like a Friday. It was exciting to see the first DanKu contract in the wild.

Like with most new tech, DanKu also disrupted a few things:. Since the DanKu protocol does not require trust, it removes the need for a middleman for exchanging ML models. This game changing feature will further democratize access to Machine Learning for the masses.

Hopefully this will create an uptick in open-source ML models available for everyone. The protocol also makes it possible to collectively raise money for things like cancer research. Universities and research groups can create contracts for open-problems like protein folding, etc. Any person could directly donate money to these contracts. Donating money for medical research will never be the same ever again. These pools will probably be managed by Data Scientists, who will try to solve these ML problems.

The contract reward could later be divided among the data scientist and GPU suppliers. Another interesting application would be in finance.

If submitted models result in a tangible financial result, it would be a lot easier to create DanKu contracts and finance them. DanKu contracts also create the opportunity for AI systems to self-improve. AI systems could contract out work to improve itself in an automated and seamless manner when it encounters new data. The use of cryptocurrencies also makes it an attractive method of self-improvement, since crypto payments and transactions are accessible by anyone or anything.

Improvements in Ethereum, further optimizations in the contract, and improved protocol design can all make DanKu contracts better. With these improvements, a more diverse set of ML models could be further supported. S Presidential county election data as our dataset.

Every county is represented with 3 data points: For example, a data point can look something like this: The first two values refers to the latitude and longitude of Seattle.

The third value, 0 refers to the Democratic candidate, whereas 1 refers to Republican candidate in this data format. Since the training and testing datasets are randomly selected by the contract, the fairness of the competition is assured. Participants of the contract are required to train a simple forward pass neural network where they can define the neural network structure in terms of layers, neurons and biases.

After training, they are required to submit their network definition, weights and biases to the contract. The DanKu contract was initialized on block Feb 19th, The reward for the winning submission is 5 Ether ETH. You can find the guide here to help you participate in the competition. Comments, questions or criticisms are welcomed at algorithmia.

Algorithm Engineer at Algorithmia, helps make complicated things simpler. Believes that Machine Intelligence will have a huge impact on our lives in the days to come, and hopes to have a defining role in shaping this new future. Algorithmia Blog - Deploying AI at scale. Some background As of , Artificial Intelligence and Blockchain continue to dominate tech news everywhere. How does it actually work? You can describe the DanKu protocol at a high-level in 4 steps: Running the first neural network on the blockchain For demo purposes, we decided to write a DanKu contract where a neural network was implemented.

Some disruptive features It was exciting to see the first DanKu contract in the wild. Like with most new tech, DanKu also disrupted a few things: The third value, 0 refers to the Democratic candidate, whereas 1 refers to Republican candidate in this data format random data points are selected as the dataset for this competition.

Besir Kurtulmus Algorithm Engineer at Algorithmia, helps make complicated things simpler. More Posts Follow Me:

desukedo.biz