Artificial Intelligence, Augmented Reality and Deep Learning – the Hot Topics at the Events in the Fourth Quarter of 2017

In the fourth quarter of 2017, the IT-Personalberatung participated in various technology and innovation events to be able to deliver top candidates for our clients' critical positions within the scope of the digital transformation.

Amongst others, we attended the following events:

• Stuttgart Knowledge Forum - as Top 100 Entrepreneurs chosen by Speakers Excellence
• EHI Technology Days in Düsseldorf - with Manfred Wenzel and Viktoria Roenick as experts for the retail and logistics sectors
• Private banking event in Vienna - to acquire knowledge in this sector

Please find below a summary of the most exciting presentations:

We would like to single out the EHI Technology Days 2017, in which the topics of artificial intelligence, augmented reality and deep learning ran through all presentations. Top decision-makers from the trade industry came together in the Rheinterrasse Düsseldorf for the largest trading technology conference in Germany, Austria and Switzerland.

Retail companies today are faced with the major challenges of winning and retaining customers in an uneven competition between retail and internet shops and of impressing these customers with ingenuous omnichannel strategies.

We have selected three exciting presentations:

1. Presentation

Artificial Intelligence - How Automated Decisions will Revolutionize Business Processes in the Future. Prof. Dr. Michael Feindt, Professor of Physics at the Karlsruhe Institute of Technology

Dr. Michael Feindt is the founder and chief scientist of the Blue Yonder GmbH. The Blue Yonder GmbH is a leading provider of AI Solutions (Artificial Intelligence) for retailers.

What is intelligence? Max Tegmark defines intelligence as the ability to achieve complex goals. The ultimate goal of artificial intelligence research is to create an artificial general intelligence (AGI) equivalent to the human mind.

Artificial intelligence for specific applications (superhuman narrow AI) already exists, e.g.

• Image recognition (image classification (Google), face recognition)
• Audio translation: (text-to-speech, speech-to-text)
• Two-player games (chess (IBM), Jeopardy (Watson), Go (DeepMind)
• Atari video games (Space Invaders, Breakout (DeepMind))
• Trading decisions (supply chain, product pricing (Blue Yonder))
• Elementary particle physics research (event reconstruction (M.F. et al., KIT))

In artificial intelligence, we distinguish between:

Artificial Intelligence I

Machines learn to accomplish tasks that are easy for human beings, e.g. image recognition. It aims to facilitate and automate human-machine communication.

Artificial Intelligence II:

Develop smart algorithms to optimize highly complex sequences of actions. This involves deep neural networks - reinforcement learning. Google invested $1 billion in computing time to train such neural networks.

One example of applied artificial intelligence II is the computer program 'Alpha Go', which is exclusively playing the board game of Go (the most difficult game in the world). In October 2015, the program defeated multiple European champion Fan Hui. By playing against itself, the program became better than the human champion.

Artificial Intelligence (general facts):

Human intelligence is influenced by emotions (gut feelings), the frame of mind on a particular day, the overall situation, etc. Artificial intelligence is based solely on data - learning processes and algorithms. Any disturbances are eliminated to prevent wrong decisions.

Artificial intelligence is the key to better decisions in the age of digitalization, as it provides us with:

• objective data
• good/correct forecasts including uncertainty information
• good/correct cost-benefit assessment
• good/correct optimization
• optimized automated procedures based on many forecasts

The big question many entrepreneurs ask themselves is: Will artificial intelligence be required for successful business processes in the future?

The fact is that 99% of all operational business decisions will be fully automated in the future, made by algorithms and based on data. Thanks to data-driven processes, companies can increase their competitiveness, boost their sales and achieve higher margins. In goods management and pricing, processes are already optimized and controlled through artificial intelligence. But trade is not the only sector about to get "Amazonned". New technologies and the associated solutions, be it artificial intelligence, machine learning or data science, will have a big impact on many sectors.

2. Presentation

Artificial Intelligence for the Trade Industry: Karin Kaltenkirchen, Managing Director of the Modehaus Marx GmbH, Franziskos Kyriakopoulos, Managing Director of the 7lytix GmbH

The Modehaus Marx uses deep learning algorithms to "learn" about the style of its customers on the basis of transaction and product data. This permits the generation of real-time personalized recommendations that are perfectly tailored to the individual customer profiles.

On a daily basis, the trade industry faces the following issues that have a negative impact on sales:

OUT OF STOCK:

Effect: The customers abandon their shopping cart because they can't find the desired item. Solution: Data-based sales forecasts: e.g. future weather data are synchronized with product features (e.g. warming - rainproof) by means of machine learning and deep learning.

Effective product placement at the POINT OF SALE: Which offers should be grouped together? Which combination leads to cross- or up-selling? Solution: Association rule analysis: Analysis of itemsets, i.e. items purchased together (network analysis). Strong association rules are automatically identified at multiple levels (item, brand, section, etc.)

Why don't my customers shop more often?

Solution: Offering the right kind and amount of items to the customers thanks to reliable customer purchase predictions.

These predictions are acquired by means of collaborative filtering (analyzing the behavior patterns of user groups to make automatic predictions about the interests of a single individual), recommender systems (automated recommendation technologies), item-item & customer-customer similarities (identifying similarities to recommend similar products) as well as network theories.

Conclusion: By using the right AI methods, common issues of the trade industry could be resolved. By means of sales forecasts, out of stocks, surplus stocks and price reductions can be avoided. Thanks to customer purchase predictions, personalized offers and targeted advertising concepts can be created. And, lastly, association rule analyses can be used to optimize product placement at the point of sale as well as advertising campaigns.

3. Presentation

Augmented Reality Indoor Navigation with Google Tango, Peter Krämer, Managing Director of Cologne Intelligence

First, a quick detour to Wikipedia:

"Augmented reality (AR) is a live direct or indirect view of a physical, real-world environment whose elements are "augmented" by computer-generated perceptual information, ideally across multiple sensory modalities, including visual, auditory, haptic, somatosensory, and olfactory. [...] Augmented reality is used in order to enhance the experienced environments or situations and offer perceptually enriched experiences. [...] Augmented reality brings the components of the digital world into a person's perceived real world. " (Wikipedia).

The trade industry can use augmented reality to adapt to the new online customer behavior and can use its advantages to lure customers into stores.

BMW does the Tango - A virtual tour across instrument panels and ornamental seams:

BMW is the first car manufacturer to use augmented reality to sell cars. The Munich-based company developed the app in cooperation with tech giant Google. Consumers can use their smartphones to approach the car, open the door, take a close look at the interior and even turn on the radio. "We are the first manufacturer to use augmented reality in car sales", says Biermann, Head of Innovations Sales. BMW is convinced that the app has great marketing value. Procedures in car sales are changing. Twenty years ago, the average customer made four visits to the car dealership before buying a car. Today, he goes there once. "In most cases, the decision was made online", says Biermann.

One of the top trends at the Consumer Electronics Show in Las Vegas was the combination of real-life data with computer-generated visualizations. At the same fair, the car manufacturer introduced its BMWi Visualizer.

Lowe's (a U.S. home improvement retail company) offers an augmented reality app to help customers find items in their stores. Customers can use Tango-enabled smartphones to easily search for products, add them to a shopping list and locate the product within the store.

Another example of applications in the retail sector is IKEA Place, a free app that is based on Apple's ARKit. The application allows people to virtually place IKEA products in their homes.