14 June 2016

Game Theory

Game Theory in many respects is the bedrock of advertising and algorithmic trading markets. Combine this with Complex Networks and formal Machine Learning and one has a decisive strategy model. The mathematical models are also applied in Multiagent Systems for studying Argumentation Theory and communication between agents. A Beautiful Mind was a movie that perhaps made Game Theory and the concept of Nash Equilibrium a more mainstream concept. There are endless applications to the field and a few reference sources of further study are provided below.

Complex Networks

The rising scale of data and the need for information gain has provided a greater need towards understanding patterns to form knowledgeable insights. In many cases, such patterns can be derived through machine learning and data mining. But, also through studying complex networks that form within contextual data. The below links provide useful sources of study in the science of complex networks.

10 June 2016

Brexit Pipeline

Studying Sentiment Analysis in context of Brexit (EU Referendum) is currently an intensive area as the polling stations will very soon be active for voters. Input sources from social media and news feeds can be a focal point for storytelling about the various events. Social media and news feeds can be utilized in form of stream processing which can then be used for machine learning analysis and then indexed for summary into Elasticsearch. A sample workflow example is provided below.  Reader will take notice that the sample workflow is also supported as an example for learning Apache Flink. The workflow can be modified as required for example, one could use a Redis cache layer between the machine learning process and Elasticsearch. Also, could extend with an NLP pipeline (Gate/UIMA) or simply OpenNLP/CoreNLP for extracting information. One could even replace Apache Flink with Spark or GraphLab. Alternatively, one could even replace Kafka with Kinesis and simply apply the AWS data pipeline. Also, the data sources can be stored using S3. Furthermore, one could even use DL4J with Spark on ElasticMapReduce to apply Deep Learning approach in form of convolutional neural network model. Although, Python developers may be more inclined to use Theano, TensorFlow and possibly RabbitMQ. For a graph representation one could use Titan, GraphX, Elasticsearch Graph, Cayley, PowerGraph, Gelly, among others. As one can see there are several ways of implementing a solution on a case-by-case basis to translate the requirements of stories. However, prototype in small is always the best way to go before scaling out incrementally i.e fail fast

Input->Kafka->ApacheFlink->Elasticsearch->Output

Steps:
  1. Collect
  2. Log 
  3. Analyze
  4. Serve & Store
List of Input Sources:

As a side note, GNIP and DataSift provide an entire data source pipeline for building out a firehose of streaming inputs. Live Polling data can also be used to gather voting trends as they happen. However, as the referendum is now past, one can probably get a hold of the dataset or API.

Deep Learning Software

Below are a few links to current deep learning libraries available for the developer community.

software links
deep learning libraries by language
popular deep learning tools
15 Libraries for Deep Learning
Comparison of Deep Learning Software
deep learning bibliography

Intelligent Health Insurance

Private Health Insurance needs to be transparent with claims handling. They also need to profile their clients ethically and cover for legitimate claims. But, many private health insurance companies even lack basic customer service. Some providers are so irresponsible that they even make it difficult for clients to contact them directly. Increasingly, it seems many providers avoid to cover for legitimate claims. In what respects are they being regulated for such practices. CrowedFunded Insurance providers should be the next step forward. This way everyone can be a customer as well as a health provider. Collectively, everyone is reaching a win-win situation rather than a single insurance provider that is driving business out of the medical misery of others. This may also increase medical accessibility for all concerned as well as improve research. Furthermore, deep analytics can be attained on health profiles on the basis of handling claims. Such practices can further be shared in form of peer-to-peer networks between countries so people can take advantage of quality and cost-effective healthcare in a timely fashion without having to wait in a queue or being limited on the choice of a medical professional. However, stringent regulatory and compliance requirements would have to be met. One might wonder that the NHS is partly doing this already. However, the NHS is badly managed, deeply seated in bureaucracy, and some times uncaring medical staff. Also, doctors can directly converge on claims handling avoiding malpractice and increasing the scope for specialist consultation. DNA profiling can also be a step forward towards healthy living for the future as well as personalization in healthcare. The driving force for any healthcare should be that the health of the patient comes first, above all they receive the most appropriate treatment and diagnosis, in a timely fashion, as should be expected without an excessive financial burden.

5 June 2016

Reasoning in Artificial Intelligence

Reasoning is an important aspect of knowledge representation in artificial intelligence. It provides a formalism for deduction and induction over representation of a set of knowledge utilizing constraints for inference defined in form of logical rules. Hence, why knowledge representation and reasoning are so interrelated in theory. There are various forms of reasoning in knowledge-based systems. A few reasoning approaches are listed below. 

4 June 2016

Intelligent Agent for Retail

Retail stores can utilize intelligent agents to solve dynamics of shelving, customer service, and pricing. Tiny drones could be deployed around the retail supermarkets during inventories and during random peak hours to manage stock levels and monitor for pricing. This will also replace the option of having to require shelf pickers or even dual purpose cashiers for smaller stores. Going further, intelligent agents could be used for more mobile and flexible cashiers. In process, the whole pricing and supply chain management can be automated even going as far as to implement more intelligent ways to mine for personalized offers. Retail stores need to manage prices, customer behavior, and optimization of sales. In process, there are quite a few variables at play: price drifting, not all customers are the same, not all items in a category are the same (basket of goods), price execution, and having the right sales strategy at the right time. It is inevitable that retail stores will look towards more insightful and intelligent price optimization strategies to meet supply and demand while attaining maximization in sales revenue from satisfied customers. Sentiment analysis will also come to play a significant factor towards connecting with the customer's intentions and conversion. Eventually, a self-driving option through use of effective artificial intelligence such as with deep learning may spearhead for more cognitively and semantically aware retail stores.

2 June 2016

Coursera Specializations

Coursera have interesting specializations especially in data science. A list of some useful options that are regularly offered and are quite thorough in their coverage is included below. One interesting aspect of the specializations is that they can further be included as credit options towards a degree. For example, data mining and cloud computing specializations cover the MOOC aspect of the Online MCS-DataScience degree course at UIUC. It appears that it will essentially become the trend as more degrees are offered online for working professionals.

Machine Learning
Big Data
Data Science
Data Science at Scale
Functional Programming in Scala
Data Mining
Cloud Computing

Alternatives to Namecheap

Namecheap has significantly deteriorated in service and this has mostly happened since the time they swapped over to the new clunky and unreliable website. Not only are there significant practical features missing that were available in the older website but it is also significantly less intuitive. One can hazard a guess that their UI/UX design team needs to be fired for such a terrible job for not even having the basic sense of usability principles. However, domain providers have huge competition so a consumer has the privilege of shopping around. Cloud service providers like AWS and Google have also entered the domain market adding further competition. However, it is advisable to keep the hosting provider separate from the domain provider. But, chances are that many will be looking to AWS for domain registration and DNS services inclusive of all the additional scale out options, especially for production ready applications. Companies like namecheap will eventually be losing market value as they fade into the dust through their own failures of service. A few alternative options over namecheap for domains is provided below.