Blockchain and Artificial Intelligence

Blockchain and Artificial Intelligence

Guest blog by Vinod Sharma.

Abstract – Blockchain is a mystery story or provides the foundation for cryptocurrencies like Bitcoin. What’s different about blockchains compared to traditional big-data distributed databases like MongoDB. Its like featuring a product that contains small blocks of brain in form of dust but consider that the innovation efforts of several publicly traded asset managers and banks are also on this brain block dust quest.  Computers start simulating the brain’s sensation, action, interaction, perception and cognition abilities. Blockchain is a new approach to manage/monitor financial and other transactions, Guarding an innovation department or powerhouse lab is a smart setup without inbuilt component of artificial intelligence is like an effort of joining blocks without reference of previous block. In this article the idea is to draw a rough sketch of   exaggerated scenario of how these two technologies may interact with us in the future and what warrants the, perhaps perplexing, 2 super powers. AI’s Control systems are widely used. They govern how a simple thermostat adapts to a target temperature. Before we discuss this further, let’s first review. Views here are from many of my friends, colleagues and reading through web. All credits if remains on the original  contributor only.

Introduction – How Future Technologies affect our life’s, How many jobs will be taken by robots and how much will left for humans; difficult to predict today. At the same time software development/ coding jobs will be less and less thats for sure with machine learning concept and its not very far may be just another decade and all software development jobs are gone. The future computing will have extraordinary capabilities with expected factoring of a 3,000 digit number (40 power to 10) faster than today. Personal health devices will send data to databases and glucose levels will be analyzed from tears by contact lenses. AI will form its own team on the fly with combination of hardware, software, infrastructure and machine learning trained programs organized to facilitate planning, control, coordination, and decision making in an organization and this process will happen any time or many times for various tasks. Blockchain to transform financial services. Disruptive technologies transform traditional industries.

Main Story – Artificial intelligence and blockchains are the 2 closed components of digital business. While Blockchains can help us verify, execute and record. AI helps in decision making, assessment, understanding and recognizing. While the machine learning methods that are a part of AI help us find opportunity and improve decision making, smart contracts and blockchains can automate verification of the transactional parts of the process. Connected industries require new innovative solutions as in new technology scenarios human errors replaced by technical failure.  Google search looks like a fair deal to the consumer because it is an amazing free product but wait is it really free may be not.  Google uses its algorithms to index web pages in a way that provides a near-seamless service to anyone who wants to locate content on the web.  In return it ask for the right to collect and record information on your browsing history, which it uses to; better understand the content that users want to see on the web, feeding this information back into Google search and improving its search engine service in the process and also help its customers sell you better targeted advertising, based on the customer profile it’s built up on you with the data it collects.

Blockchains might get use to work with digital human mind-files. Formulating human mind thought process and on-going thinking process to machine readable digital files and uploading and saving them on physical machine’s storage devices. Blockchain will also 

help their current prototypes, digital identities like Linkedin, Twitter or even facebook etc. Through their asset management, property registration, and access control features it will become usable, sellable and marketable. Blockchain thinking is as an input-processing-output computational system. Blockchains are much more than the already extraordinary scope of potential activity that has been envisioned for their deployment in reinventing currency, finance, economics, government, legal services, science, and health – blockchains are a basic substrate for computing itself. I guess calling bitcoin as bubble out of  blockchain would not be wrong. Blockchain thinking can also be compared with human brain power of storing our memories on a decentralized storage disk.

Concept of blockchains, a new form of information technology that could have several important future applications. Blockchains  is like a Blue Ocean Database at the expense of near-term earnings. But is the alternative, sprinkling a dash of innovation brain blocks dust approach may have been sufficient to check the planning-for-the-future box. Emerging markets are early-adopters of new technologies i.e. Drones and 3D modeling instead of on-site risks visits. Refining existing and developing new risk services is key for both technologies and businesses to prepare jointly for the next industrial revolution. Embrace digitalization and customer centricity will give super sonic experience. As a concept thinking blockchain as a thinking process which can help in formulating human thinking as a blockchain process to make machine learning process stronger. This can also help to firm up deep learning process after potential integration in turns this could have benefits for both artificial intelligence and human enhancements.

Looking into past through crystal ball on how people use to see future from their time, i.e todays time they must have seen Money in todays time as less and less Physical but more and more of  Electronic Tokens or Commodities that can have the properties like unit of account a defined value, Medium of exchange a acceptability, Store of Value a non-perishable safe from physical threats. In a simple term today e-money is an item which can be defined as sending and receiving numbers to and from email/internet addresses. Blended or overlapping memories could be stored as separate discrete units. Now adding Blockchain concept here will simply give freedom and welcome you to fork the chain and implement your own rules if you don’t like one of the changes, you are more than. Is given to the people, not to the bank or companies. The best example is Bitcoin which in simple term is the money which (supply) cannot be manipulated and is fixed at 21 million coins, each divisible up to 8 decimal.

Blockchains could be employed as a secure large-scale data management mechanism to coordinate the information of millions and billions of individuals. I remember my visit to Hong Kong in 2009, just 12 year after the biggest financial crises known as 1997 ghost  when the Thai baht lost its peg with the dollar. I heard stories  of some of the victims, including a high-flying stockbroker who was reduced to selling sandwiches, and a businesswoman whose boss told her to “take care of the work for him” before hanging himself. Now  where necessary, they try to neutralise heavy capital inflows with offsetting flows the other way, including central-bank purchases of foreign-exchange reserves. The AI sub-field called “Artificial General Intelligence” (AGI) is most relevant as AGI can be modeled as a feedback control system. If we try to simulate the entire episode with AI Blockchain today then solution and speed of solutions would have been beyond excellent mark. AGI is about autonomous agents interacting in an environment. Asia did not see the 1997 crisis coming precisely because it thought it had learned the lessons from earlier crises. Asian countries had high national saving rates, limited public debt and budget surpluses which brings us to AI.

Blockchain  matter to foundations of data management and it provides foundation to bitcoin which is mind game and a result of well thought process and cooked with ingredients like artificial intelligence, artificial general intelligence, Machine learning (blended with deep learning) and their components around. Perhaps the first and most straightforward element needed for thinking is memory. For blockchain computational purposes, a position can be articulated that each memory is a discrete unit and that these discrete units are encoded and stored somewhere. AGI’s Control systems have many great qualities. First, they have strong mathematical foundations going back to the couple of decades. Aspirations, linked to opportunity, can breed dynamism and inclusive, sustainable economic growth. That’s something we all want to see. Meanwhile, tech giants like Apple, Google and Intel are investing billions of dollars in using machine learning AI’s to build out their next big projects.

As on date all of us are involved in projects that involve blockchain technology implementation or related activities knowingly or unknowingly. Blockchain for Artificial Intelligence is like a global or planetary database in which AI and neuroscience have been moving towards a modular managed approach with memory to  unlocks opportunities by Data sharing and better data model. Creating audit trail on data and model for more trustworthy predictions and shared as global registry. AI DAO-  that can accumulate wealth that you cant turn off. A Decentralized Autonomous Organization (DAO) is a process that manifests these characteristics. It’s code that can own stuff.  Self-driving car is an excellent example for this.  What if you use blockchain to store the state of machine. The key move for blockchain-enabled thinking is that instead of having just one instance of a memory, there could be arbitrarily many copies of a memory, just as there can be many copies of any digital file.

The biometric authentication feature associated with mobile wallets is a great example with promising feature. With AI power to enable security features of mobile payments mean the technology could gain traction in other areas of B2B payments and escalate blockchain to generalize, any previous application of AI, but now the AI “owns itself”. We might have a future where humans own nothing, we’re just renting services from AI DAOs. Chatbots harness software that uses artificial intelligence to process language from interaction with humans in chat programs and virtual assistants. The concept of chat bots are coming up very fast but its limited to smart devices holders. They capture the interaction with the world (actuating and sensing), and adapting (updating state based on internal model and external sensors). Further that the number and location of any stored items, in this case memories, could be optimized dynamically for system operations. Blockchain will prevail as log source, event immutability, Key management credential for info-security and Quantum computing (will be a reality anytime soon).

Conclusion –Todays everyone in the said industry are looking for answers for How Artificial Intelligence Will Revolutionize Banking as we all know Artificial Intelligence is about probabilistic, constantly changing and its a algorithm to guess at reality and blockchain is deterministic, sort of permanent and its algorithm & cryptography to record reality. AI has taken some steps into banking, but it also is poised to revolutionize how banks learn from and interact with customers. Looking into future through the crystal ball, I tried describing how blockchain technology can help AI, by drawing on my personal experiences in both. As development trend for the operation support system, convergent billing, money storage as digital numbers i.e bits and bytes has broad scope that is not limited by a single standard. Artificial Intelligence (Cognitive Computing) will play a bigger role. Its development may follow different directions. Through different security threats. However, which is the best direction? The answer lies in the analysis of future technologies development within the 3GPP framework (For Telecom), FinTech, AI and AGI, Machine learning & Deep Learning, Threat Intelligence will play a bigger role coupled with an evaluation of the driving factors and key capabilities required by convergent systems and requirements. There’s no single answer to this without end-to-end architectural analysis. AI and blockchain  combination is explosive! Blockchain technologies . It  can help realize some long-standing dreams of AI and data analysis work, and open up several opportunities. Let us open a discussion on this and build some intelligence around it. AI controls cancel noise in your expensive headphones. They’re at the heart of thousands of other devices from ovens to the brakes in your car.

About the Author

Read about Author at  About Me | FacebookPage | Twitter | ContactMe | LinkedinPage,     

Blockchain and Artificial Intelligence

New Approaches to Unsupervised Domain Adaptation

New Approaches to Unsupervised Domain Adaptation

This article was contributed by Nikita Johnson. 

The cost of large scale data collection and annotation often makes the application of machine learning algorithms to new tasks or datasets prohibitively expensive. One approach circumventing this cost is training models on synthetic data where annotations are provided automatically. 

However, despite their appeal, such models often fail to distinguish synthetic images from real images, necessitating domain adaptation algorithms to manipulate these models before they can be successfully applied. Dilip Krishnan, Research Scientist at Google, is working on two approaches to the problem of unsupervised visual domain adaptation (both of which outperform current state-of-the-art methods.)

What you can find in the full article: 

  • Tell us more about your work, and give us a short teaser to your session?
  • What started your work in deep learning?
  • What are the key factors that have enabled recent advancements in deep learning?
  • Which industries do you think deep learning will benefit the most and why?
  • What advancements in deep learning would you hope to see in the next 3 years?

To read the original article, click here.

Top DSC Resources

Follow us on Twitter: @DataScienceCtrl | @AnalyticBridge

New Approaches to Unsupervised Domain Adaptation

Recommendation System Algorithms

Recommendation System Algorithms

Today, many companies use big data to make super relevant recommendations and growth revenue. Among a variety of recommendation algorithms, data scientists need to choose the best one according a business’s limitations and requirements.

To simplify this task, my team has prepared an overview of the main existing recommendation system algorithms.

Collaborative filtering

Collaborative filtering (CF) and its modifications is one of the most commonly used recommendation algorithms. Even data scientist beginners can use it to build their personal movie recommender system, for example, for a resume project.

When we want to recommend something to a user, the most logical thing to do is to find people with similar interests, analyze their behavior, and recommend our user the same items. Or we can look at the items similar to ones which the user bought earlier, and recommend products which are like them.

These are two basic approaches in CF: user-based collaborative filtering and item-based collaborative filtering, respectively.

In both cases this recommendation engine has two steps:

  1. Find out how many users/items in the database are similar to the given user/item.
  2. Assess other users/items to predict what grade you would give the user of this product, given the total weight of the users/items that are more similar to this one.

What does “most similar” mean in this algorithm?

All we have is a vector of preferences for each user (row of the matrix R) and the vector of user ratings for each product (columns of the matrix R).


First of all, let’s leave only the elements for which we know the values in both vectors.

For example, if we want to compare Bill and Jane, we can mention that Bill hasn’t watched Titanic and Jane hasn’t watched Batman until this moment, so we can measure their similarity only by Star Wars. How could anyone not watch Star Wars, right? 🙂

The most popular techniques to measure similarity are cosine similarity or correlations between vectors of users/items. The final step is to take the weighted arithmetic mean according to the degree of similarity to fill empty cells in the table.

Matrix decomposition for recommendations

The next interesting approach uses matrix decompositions. It’s a very elegant recommendation algorithm because usually, when it comes to matrix decomposition, we don’t give much thought to what items are going to stay in the columns and rows of the resulting matrices. But using this recommender engine, we see clearly that u is a vector of interests of i-th user, and v is a vector of parameters for j-th film.

So we can approximate x (grade from i-th user to j-th film) with dot product of u and v. We build these vectors by the known scores and use them to predict unknown grades.

For example, after matrix decomposition we have vector (1.4; .9) for Ted and vector (1.4; .8) for film A, now we can restore the grade for film A−Ted just by calculating the dot product of (1.4; .9) and (1.4; .8). As a result, we get 2.68 grade.



The previous recommendation algorithms are rather simple and are appropriate for small systems. Until this moment, we considered a recommendation problem as a supervised machine learning task. It’s time to apply unsupervised methods to solve the problem.

Imagine, we’re building a big recommendation system where collaborative filtering and matrix decompositions should work longer. The first idea would be clustering.

At the start of a business, there is a lack of previous users’ grades, and clustering would be the best approach.

But separately, clustering is a bit weak, because what we do in fact is we identify user groups and recommend each user in this group the same items. When we have enough data it’s better to use clustering as the first step for shrinking the selection of relevant neighbors in collaborative filtering algorithms. It can also improve the performance of complex recommendation systems.

Each cluster would be assigned to typical preferences, based on preferences of customers who belong to the cluster. Customers within each cluster would receive recommendations computed at the cluster level.

Deep learning approach for recommendations

In the last 10 years, neural networks have made a huge leap in growth. Today they are applied in a wide range of applications and are gradually replacing traditional ML methods. I’d like to show you how the deep learning approach is used by YouTube.

Undoubtedly, it’s a very challenging task to make recommendations for such a service because of the big scale, dynamic corpus, and a variety of unobservable external factors.

According to the study “Deep Neural Networks for YouTube Recommendations”, the YouTube recommendation system algorithm consists of two neural networks: one for candidate generation and one for ranking. In case you don’t have enough time, I’ll leave a quick summary of this research here.


Taking events from a user’s history as input, the candidate generation network significantly decreases the amount of videos and makes a group of the most relevant ones from a large corpus. The generated candidates are the most relevant to the user, whose grades we are predicting. The goal of this network is only to provide a broad personalization via collaborative filtering.


At this step, we have a smaller amount of candidates that are similar to the user. Our goal now is to analyze all of them more carefully so that we can make the best decision. This task is accomplished by the ranking network, which can assign a score to each video according to a desired objective function that uses data describing the video and information about users’ behavior. Videos with the highest scores are presented to the user, ranked by their score.


Using a two-stage approach, we can make video recommendations from a very large corpus of videos while still being certain that the small number of them are personalized and engaging for the user. This design also enables us to blend candidates together that were generated by other sources.


The recommendation task is posed as an extreme multiclass classification problem where the prediction problem becomes accurately classifying a specific video watch (wt) at a given time t among millions of video classes (i) from a corpus (V) based on user (U) and context (C).

Important points before building your own recommendation system:

  • If you have a large database and you make recommendations from it online, the best way would be to divide this problem into 2 subproblems: 1) choosing top-N candidates and 2) ranking them.

  • How do you measure the quality of your model? Along with the standard quality metrics, there are some metrics specially for recommendation problems: Recall@k and Precision@k, Average Recall@k, and Average Precision@k. 

  • If you are solving recommendation problems with classification algorithms, you should think about generating negative samples. If a user bought a recommended item, you should not add it as a positive sample, and others as negative samples.
  • Think about the online-score and offline-score of your algorithm quality. A training model only on historical data can lead to primitive recommendations because the algorithm won’t know about new trends and preferences.

    This post originally appeared on here

Types of Machine Learning Algorithms in One Picture

Types of Machine Learning Algorithms in One Picture

Here is an interesting visualization of machine learning algorithms:

Originally posted here.  Also check out the following great visual summaries:

DSC Resources

Popular Articles

Types of Machine Learning Algorithms in One Picture