Search

Big Data & Analytics

Towards Augmented Intelligence

ARTIFICIAL INTELLIGENCE AND AUGMENTED INTELLIGENCE

Artificial Intelligence and Augmented Intelligence have entered the business realm, been the subject of academic and online discussions and captured the collective imagination with great enthusiasm.
The terms Artificial Intelligence and Augmented Intelligence are often used interchangeably. However, although they do share many methodological and technological similarities, these two approaches are based on two different assumptions, one placing machines and the other humans at the heart of the decision making process.

Artificial Intelligence is the collection of mathematics, computing and statistics that makes it possible to perform tasks once deemed exclusively human in nature, tackling problems typically resolved by humans through intelligence. Within this context, it is acknowledged that the algorithm has an element of independence: once the training has been completed, the system voluntarily initiates the action within its environment and pursues the objectives set without interacting with the human agent.
Conversely, Augmented Intelligence integrates and supports human thinking, analysis and planning, maintaining, however, the intent of a human player at the centre of the human-machine interaction.

The evolution

During the development of Augmented Intelligence, the methods by which computer systems learn have evolved, starting with traditional statistics and moving towards the modern Deep Learning and Augmented Intelligence approaches.

From humans

The machines learn from humans: the field of traditional statistics and the first steps in supervised learning. In this case the goal is known and Artificial Intelligence is used to improve system performance, with the algorithm helping to predict the link between inputs and outputs.

As humans

The machines learn as humans: the similarity level with human learning grows and artificial intelligence elements are transformed to accommodate the change. Deep Neural Networks mimic the behaviour of the human brain and attempt to induce solutions learning from heterogeneous and unstructured data.

With humans

The machines learn with humans: the learning process becomes collaborative and the Augmented Intelligence realm is introduced. Machine learning and the human expert come together: the algorithm produces results and modifies its behaviour based on the feedback received from the human user.

  • strip-0

    TARGET REPLY

    Target Reply operates on many aspects: it supports companies through an end-to-end consulting process that extends from identifying customer requirements to designing and implementing concrete solutions, working with some of the most advanced technologies in the Data Integration, Data Modelling, Predictive Analytics, Machine Learning and Data Intelligence realms. Target Reply provides customers with a complete and up-to-date skill set encompassing all Data Science technologies, both commercial and Open Source. Target Reply analyses and meets customer requirements from building the architecture of the data infrastructure, to the structuring of analysis dashboards, providing customers with clear insights, applicable to their specific business in a tangible way.

  • AUGMENTED INTELLIGENCE IN THE BUSINESS WORLD, SOME CONCRETE EXPERIENCES

    Every day, terabytes of data are produced and exchanged by companies, organisations and users, a large percentage of which is not structured data and it therefore cannot be attributed to fixed structures nor is it easily usable. Managing such a large amount of data is not possible for human beings without the support of computer systems. However, the unstructured nature of the data signifies a capacity for abstraction and analysis that is not readily manageable by traditional analytics.

    strip-1

Target Reply's experience in data analysis has enabled the company to successfully implement many projects, juxtaposing the teams with Artificial Intelligence algorithms and Machine Learning techniques designed to provide a significant amount of added value and to facilitate results that are otherwise unattainable.

Target Reply’s Data Scientists team has successfully implemented various projects, applying Augmented Intelligence logic and technologies in different business contexts and highlighting the flexibility and value of this approach. Some examples:

  • At a leading consumer credit company, the anti-fraud team received support from a Machine Learning engine based on Neural Networks that makes it possible to increase the efficiency and effectiveness of the human team in identifying fraudulent behaviour.
  • For businesses, the ability to be able to identify in advance customers that may leave the company for a competitor’s solution offers a major strategic advantage. Based on customer history data, algorithms and neural networks, a large company operating in the B2C market was provided with information on possible churners, leaving the teams with the task of deciding on the strategy.
  • Target Reply assisted with the Claim Management process for a major player operating in the Oil & Gas sector. The approach for this project was drastically different from previous ones: without human supervision, the algorithm constructs correlations, identifies connections and simplifies content searches, enabling people to easily navigate a large amount of seemingly disconnected information.