Delve into some of the most inspirational, outlandish or exciting data and AI news, and explore the art of the possible, one story at a time. Today, we’re jumping from AI and machine learning to mapping some pretty large data. Enjoy.

The scent of a machine

Perfume is a provocative thing. Scent plays on our emotions to invoke feelings and rouse memories. Smell is a very human sense, it’s not something a machine could understand. Is it? Well, Symrise, a major fragrance company that works with the likes of Estée Lauder, Avon, and Donna Karan, is now working with IBM too, to apply machine learning to the world of fragrance. IBM has developed an algorithm that studies existing fragrance formulas and compares the ingredients to other data sets, like geography and customer age. This algorithm, which the company has named Philyra, can now develop new perfumes that will target very specific market segments. Symrise has 1.7 million fragrance formulas that it’s collected over the years – not just for expensive perfumes, but for toothpastes, pet foods, detergents, candles, foods, and drinks too – which have been shared, along with information on how they sold, customer data and demographic preferences. Philyra uses that information to create new formulas targeted to highly specific demographics.

Is AI art’s next medium – $432,500 suggests so

‘Portrait of Edmond Belamy’ depicts a portly gentleman. It’s not too dissimilar to art works you might see hanging in any gallery. This painting is different, though. Despite appearing to be unfinished – the facial features are somewhat indistinct and there are blank areas of canvas – it has just sold for $432,500 at auction. Take a closer look at the signature, however, and it might give you a clue as to why this portrait is so different. The portrait was created by an algorithm. The system was fed a dataset of 15,000 portraits painted between the 14th century to the 20th, and the AI makes a new image based on the data. A secondary part to the algorithm then tries to spot the difference between a human-made image and one created by AI. The aim is to fool it into thinking that the new images are real-life portraits.

Visualising the world’s population in 3D

Human Terrain - visualising the worlds population in 3DOur urban environments are growing at a rapid pace. As the populations migrate from rural locations to cities, new mega cities are growing, dwarfing our existing giants. Kinshasa is now bigger than Paris. Guangzhou, Hong Kong and Shenzhen are forming an epic, 40 million-person megalopolis… It’s hard to comprehend the scale of these cities. But data visualisation can help. The “Human Terrain” is a data view of the population of every major city, town and village around the world. The project draws on free and open data, namely the Global Human Settlement Layer, which brings together satellite images, census data and geographic information provided by volunteers to create global population density maps. The Human Terrain allows you to see the current populations around the world, compare the density to 1990 and see the change over a 15 year period. Go, get lost.

How AI and guitar hero could help you become the next Mozart

How AI and guitar hero could help you become the next MozartEver wanted to be able to play the piano like a pro – but really can’t be bothered learning? (Me too!) Well, now even the most tone-deaf of us probably can, thanks to an AI system called Piano Genie… The team at Google’s Magenta project trained a neural network on a dataset made up of about 1,400 pieces of piano performances. The system learned what notes typically follow one another, much like your phone’s predictive texting tries to guess what you want to write next. Inspiration for the project came from Guitar Hero, the video game that lets players bang on a plastic guitar’s buttons, following the notes displayed on the screen – reducing the 88 piano keys with just 8 plastic buttons.

Using AI to analyse movie trailers and find out what films audiences will like

Using AI to analyse movie trailers“Machine learning is, at heart, the art of finding patterns in data. That’s why businesses love it. Patterns help predict the future, and predicting the future is a great way to make money.” Film studio 20th Century Fox, is hoping to cash in by finding patterns and using AI to predict what films people will want to see. Researchers from the company have been analysing the content of movie trailers using machine learning. Machine vision systems examine trailer footage frame by frame, labelling objects and events, and then compare this to data generated for other trailers. The idea is that movies with similar sets of labels will attract similar sets of people, allowing the studio to micro segment their content.



Posted in AI, Big data, Media roundup On November 9, 2018 By