Deep learning networks are infamous for their ability to detect cats in images. Advances in computer vision and the application of Convolutional Neural Networks (CNN’s) have yielded exciting advances in image classification and computer vision applications. CNN’s are used to classify images and identify the objects that are in them. They essentially translate pixels values to information about what is in the image. There are often many layers between pixel values and outcomes. The layers in these networks can be used to determine the style of an image. Early layers tend to identify lines or colours, whereas later layers identify more complex objects and derivations.
Combining data that has been generated from 2 images that have passed through CNN’s allows a principal content image to be mixed with style from another image. Content and style are weighted, and the algorithm iterates through numerous passes of the images to align the images. The style of an image is derived from comparing convolutional channels filters and the correlation between them to produce gram matrices. Further details on the approach and specification can be found here.
We have been experimenting with various content images and style images over the 2017 holiday period. Although the commercial value of such image generation is difficult to quantify (as with traditional art), the neural style transfer approach allows AI to generate amazing new vivid images by combining a “content” image and a “style” image. We have posted various examples to the Algospark Neural Style Transfer Art gallery. These can be found here:
The implication is that an already large image library of content and style can be combined using AI to generate exciting new computer art derived art libraries.
How can you offer great service without an exploding product list? Meeting an increasing number of customer needs from a growing list of customers can lead to exponential growth in product offerings. Do you really want to be the one stop shop for everybody for everything?
Most organisations follow the 80:20 product rule. This means that 80% of customers buy 20% of the product offerings. Products that are not in the top 20% make up part of the “product long tail”. Whenever there is an efficiency drive, these products typically appear in the cost saving table of a PowerPoint presentation. But these products have been developed to meet customer requirements, and are nearly always part of a portfolio of products that customers buy. How can product investment or divestment opportunities be made for specific products without jeopardising customer relationships? How should the product tail be cut? Or more importantly, what new products should I recommend to customers? The answer is learn from supermarkets and shopping baskets.
Market basket analytics and product graph analytics are excellent ways to determine “hero products” and the dependencies with other products. These type of analytics measure products by their support (% of transactions in which the product appears), confidence (probability of buying product X if you also buy product Y) and lift (strength of product inter-relationships). Product portfolio dashboards are an excellent way to visualise these metrics. They allow fast understanding of key product relationships that make it easy to determine core product clusters and the most important product associations. This can then be linked to evaluation of product financials (ie sell products that make money) and development of recommender systems (suggest products that customers want).
So using a product portfolio analytics tool will help keep product development in line with demand patterns. It also helps guide customers to more consistent product portfolios without “exploding” the product list.
See an example of an Algospark product portfolio dashboard here: https://wilkinsondi.shinyapps.io/newproddev/
This forms part of a suite of optimisation tools for sales, product and process. Further details are here: https://algospark.com/#sales