Tech Blog

Neural networks: building block of radical change

Guest post by Daria Petrova, Marketing Manager at NTR Lab

BPU Neural Network

Today, neural networks have a high profile at AI conferences. Becoming more popular within tech companies, neural networks are becoming increasingly popular for forecasting, decision-making, pattern recognition, classification, optimization, and data analysis. Here are some practical uses of neural networks:

Self-driving cars:

Neural networks are used for image analysis and object recognition. Some deep-learning models specialize in streets signs while others are trained to recognize pedestrians.

Automatic text generation:

Neural networks are used to learn punctuation, grammar and style of a specific text and/or script, then automatically create entirely new texts with the same style.

News aggregator based on sentiment:

Neural networks are used to filter content based on sentiment, so users can create news streams that only cover what he or she needs. BPU Holdings, an Artificial Emotional Intelligence (AEI) company, does just that with their product called Neil – a personal news curator using Artificial Intelligence (AI).

Image caption generation:

Neural networks are used for object detection in photographs and then turn the labels into coherent captions, with proper sentence structure for the corresponding image.

Automatic colorization:

Neural networks are trained in large organized image databases to be used for adding color to black and white photographs and videos.

Recommendation systems:

Neural networks are used for e-commerce websites to contextualize images for recommendation systems. This enables companies to better understand a consumer’s buying pattern; in turn, enabling the consumer to receive recommendations based on their preferences, swiftly and efficiently.

Translation:

Similar to automatic text generation, neural networks are helping enhance automatic translation of text by using stacked neural networks.

Painting:

Neural networks are being used to paint a picture that combines the content of one image with the style of another by extracting the techniques of two different paintings to create a single image.

Music:

Neural networks are trained to generate music and create tunes in a chosen style at the click of a button.

Sales:

Neural networks simultaneously consider multiple variables, such as market demand for a product, customer income, population, and product price.

Banking:

Neural networks are used to forecast future price and exchange rates, stock performance, and credit rating. (In 2000, Russian banks did not offer credit, because they had no system for evaluating risk. NTR created and built the first credit scoring system for a Russian bank and based it on a neural network.)

Healthcare:

Neural networks allow researchers to model different parts of the human body and recognize diseases from various scans. BPU Holdings has tapped into the Healthcare industry with their product, aiMei Framework. Currently researched with the Cloud and Autonomic Computing (CAC) Center, BPU hopes to further their research and assist cardiac patients with artificial emotional intelligence.

Marketing:

Neural networks predict consumer behavior, in order to create and understand a more sophisticated buyer segment; they also produce content and are giving marketers a myriad of new, more efficient and more dynamic tools.

Leading tech companies have been recruiting top computer scientists and global engineers for years, but it is only recently that they have figured out how to develop powerful neural networks that are true game changers for how we live, communicate, purchase, and interact with one another.

NTR’s recent projects used models based on deep neural networks to solve the problems of document classification, named-entity recognition, image segmentation, image classification, and object detection. One project we are currently working on for hospitals involves developing software that checks if required procedures are performed, e.g., if a specific patient received an injection on time.

We have been deeply involved in using neural networks for a variety of clients in many of the uses listed above.

In Artificial Emotional Intelligence (AEI), neural networks are used to make machines/technology more sensitive, heightening their ability to understand human emotions, such as people’s psyche of thinking and their feelings to specific topics, moods, and even reactions to social media content. BPU Holdings believes that with EI, everyone can improve themselves and their relationships. They seek to lead the market in AEI and expand the possibilities in multiple industries.

Zimgo