More Intelligent Retail Clouds

More data, more interactions with parts of the business, new patterns to leverage in a changing world.

How is Artificial Intelligence Changing the Retail Landscape in Maize

Francis Oyewole heads up Business Development at Aura Vision, a startup focused on integrating AI with retailers’ existing security cameras. We talked with Francis about how Covid-19 has changed the retail landscape and the role AI plays as more stores need to monitor customers’ actions.

How has Covid-19 changed the current retail landscape? And how does the possibility of another pandemic anticipate the future of retail?

Covid-19 has really accelerated digital transformation. There were some retailers who were dragging their heels on becoming more digital within brick and mortar operations. Retailers are starting to inquire and explore the technologies that they possibly should have been using before Covid-19.

That's a big shift towards taking more of the off-line world into a digital space. Data is absolutely massive during this time. And there are a lot of retailers who simply don't have the historical data that they need to be able to predict how things could be going forward. So they're scrambling to get into that position. Brick and mortar stores are disappearing in some sectors, but it seems like using AI and using data could help brick and mortar stores survive. But at the same time this crisis has really highlighted the importance of brick and mortar, not just from an economic standpoint, but also from a social standpoint. People may be in the position where they miss the opportunity to go into a store and buy something. A.I. has already impacted the online shopping experience, it only makes sense for that mindset to be taken into brick and mortar.  ... "

Counting Crowds

Interesting application, note the technique of training in various contexts of measurement.   We don't do that enough with other problems.

A bird’s-eye view of a huge crowd of people. AI Could Help Count How Many People are in Large Crowds

New Scientist  by  Chris Stokel-Walker

German Aerospace Center researchers have developed an artificial intelligence (AI) system that can accurately count the number of people in large crowds. The researchers hand-counted nearly 250,000 people in 33 images of a large crowd taken from planes, drones, and helicopters, then used this data to train an algorithm called MRCNet. The algorithm divides each image into small squares and analyzes how many people are in each. The algorithm’s results were at least 15% more accurate than those of other AI-powered crowd estimation systems. The system is much faster than hand counting, taking 0.03 milliseconds to compute the number of people in each square. .... "

Emotions in Group Photos

More than just spotting but also getting the emotion of multiple people.     Could lead to the further understanding  of crowds in contect.

Spotting Faces in the Crowd
from University of Delaware   By Julie Stewart

University of Delaware (UD) researchers are using machine learning and deep learning with neural networks to identify the emotions of people in group photos, with the goal of automatically classifying images uploaded to websites. UD's Xin Guo said, "When people search, they would see the images they are looking for because the algorithm would run and label whether people are happy or not. It could be used to analyze the emotions of a group of people pictured at a protest, a party, a wedding, or a meeting, for example. This technology could also be developed to determine what kind of event a given image shows." Guo's team scored first place in the Group-level Emotion Recognition sub-challenge of the 6th Emotion Recognition in the Wild (EmotiW 2018) Challenge at the ACM International Conference on Multimodal Interaction 2018 in October, with an algorithm that accurately classified people in a set of images as happy, neutral, or negative.  .... '

Drones Detecting Physical Behavior

Another case of complex pattern recognition.  In live video.  Here the drone is classifying a set of human behaviors and them classify them as a 'brawl'.  In English a violent, multi person fight.  Could be with or without weapons.  Could then be integrated further into a model of a crowd.  And identify individuals.    No indication that the drone would do anything other than alert the authorities.    Will see how accurate this is, and how its integration into police decision making is envisioned.

AI Drone Learns to Detect Brawls      in IEEE Spectrum  by Jeremy Hsu

Researchers at the University of Cambridge in the U.K., working with colleagues at the Indian Institute of Science, Bangalore and India's National Institute of Technology, Warangal, have used deep learning to develop a drone surveillance system that automatically detects small groups of people fighting each other. The system uses computer vision software that runs in real time to detect violent individuals, says the University of Cambridge's Amarjot Singh. The researchers trained deep learning algorithms to recognize violent actions by identifying body and limb poses in staged video footage of interns mimicking violence. Singh replaced some of the neural network layers at the front-end with fixed parameters, and used supervised learning toward the back-end, exchanging some of the deep learning process with human engineering input. This allowed the resulting ScatterNet Hybrid Deep Learning (SHDL) network to learn more quickly with less data and less available computing power. The researchers are securing permission from Indian officials to test the system at two upcoming music festivals, and Singh is working to incorporate crowd modeling into the deep learning models.  ... "

Getting Better Wisdom from Crowds

Like the quantifying aspect of this.  We experimented with WOC choices and how they linked with actual commercial decisions.   Would like to see it tested with  consumer as well as expert groups.

Better wisdom from crowds
MIT scholars produce new method of harvesting correct answers from groups.
Peter Dizikes | MIT News Office 

The wisdom of crowds is not always perfect. But two scholars at MIT’s Sloan Neuroeconomics Lab, along with a colleague at Princeton University, have found a way to make it better.

Their method, explained in a newly published paper, uses a technique the researchers call the “surprisingly popular” algorithm to better extract correct answers from large groups of people. As such, it could refine wisdom-of-crowds surveys, which are used in political and economic forecasting, as well as many other collective activities, from pricing artworks to grading scientific research proposals.

The new method is simple. For a given question, people are asked two things: What they think the right answer is, and what they think popular opinion will be. The variation between the two aggregate responses indicates the correct answer.

“In situations where there is enough information in the crowd to determine the correct answer to a question, that answer will be the one [that] most outperforms expectations,” says paper co-author Drazen Prelec, a professor at the MIT Sloan School of Management as well as the Department of Economics and the Department of Brain and Cognitive Sciences. .... " 

Baidu Understanding Crowds

In the CACM:   An approach that has been used to design retail environments.

" ... Baidu's Big Data Lab has devised an algorithm that can predict crowd formation, which researchers say could be used to help warn authorities and individuals of unusually large crowds that threaten public safety.

The algorithm correlates data from Baidu Map route searches with the crowd density of the places people search for to anticipate crowd formations at a certain place and time.

"Our algorithm is able to use crowd data from Baidu maps to predict how many people will be [at a certain location] in the next two hours," reports Baidu researcher Wu Haishan. ... " 

Mobile Sensors and Crowd Estimates

I like the idea of using data from  sensors in new ways.  In the BBC:  An example:  " ... It may be possible to estimate the size of a large crowd based on geographical data from mobile phones and Twitter, according to a new study. ... " .  Applications in Retail and elsewhere?

Crowd Governance Using the Wikipedia

This upcoming talk is intriguing:

OBAIS Seminar SeriesOBAIS Seminar Series
Crowd Governance: The Monitoring Role of Wikipedia in the Financial Market.
Ms Weifang Wu. Phd Candidate at Hong Kong University of Science and Technology
Wednesday, Feb 4, 1:30-3:00 PM EST, 219 Carl H Lindner Hall
At the University of Cincinnati

Abstract:  In this study, we explore whether Wikipedia plays a governance role in the financial market by reducing the information disadvantage of individual investors. We hypothesize that the aggregation of information on Wikipedia enables individual investors to collectively monitor insiders and institutional investors.  Using the creation of a firm Wikipedia page as an information event, our empirical results support our hypothesis and further show that the governance effect is stronger for firms with higher institutional ownership concentration as well as those with more intensive insider trading activity.  Taken together, these findings support the view that Wikipedia helps mitigate the information asymmetry among individual investors, institutional investors and corporate insiders.

More info contact Uday Rao:  uday.rao@uc.edu

Modeling Movement of Crowds

In Nature:  We looked at this kind of research for retail behavior insight.

Mathematical time law governs crowd flow

Pedestrians avoid bumping into each other by anticipating when their paths would collide.

Walking in crowds means predicting the future. When navigating heavily trafficked areas, people adjust their paths after subconsciously calculating how long it would take to collide with another person.

Researchers have come to this conclusion by analysing videos of crowds. They say that it could lead to safer design of public spaces and help in the development of crowd-monitoring methods to prevent deadly stampedes. .... "

Engagement AI Goes Public in the UK

Not really AI, as I read it, not autonomous, but tele-operation, and the researching of how people would act and interact with local avatar robotics.    What would these avatars be acting as: Police, workers, security, information providers?  The simple reaction of people to robotics in public is not clear.   Can they project authority?  What are the legal implications?   A natural extension of large scale UK CCTV networks: Making them mobile and interactive.  A stunning view of the future.   Full article in LaboratoryTalk: 

At the right a picture clip of Robothespian, one of the avatars mentioned in the article. It is not implied that this is the form of the robots being considered.

" ... Researchers from leading UK institutions are working on a £2m project designed to look at how remotely operated robots could enable people to take part in public spaces - without them actually being there. Alongside researchers from the Universities of Bath, Oxford and Queen Mary University of London, experts from Exeter University and Bristol Robotics Laboratory (BRL) will look at how using remotely operated robots might enable people to participate in public spaces.

The £2 million three-year project, Being There: Humans and Robots in Public Spaces, funded by the Engineering & Physical Sciences Research Council (EPSRC) and led by Exeter University, will examine how robotics can help to bridge the gap between the way we communicate in person and online. The aim of our research is for the robot to be an avatar for a remote person

The project will seek to look at the social and technological aspects of being able to appear in public in proxy forms, via a range of advanced robotics platforms. The robots will be controlled remotely - a method called tele-operation and a tele-operator will be able to see through the robot’s eyes and speak through its mouth, while directing where it looks and how it moves. .... "

Investor Herd Dynamics

Crowd and herding dynamics were always an interest.  With the thought of using them for physical retail understanding.  Investing is different, but this post has perhaps some related insights.  Perhaps for crowdsourcing behavior?

Crowd Modeling and Simulation

Crowd Modeling and Simulation. A favorite topic as we examined in store behavior. This takes it much further, but the elements of simulation are useful.

Disney Crowd Control Center

I mentioned the science of queuing theory here recently.   In the late 70's I had a chance to meet with scientists at Disney theme parks to talk crowd control.  As you might expect they were very interested in the dynamics of crowds as augmented by lines.   I was convinced they were the most advanced in this realm, combining both mathematical and psychological principals.  They have advanced, as is indicated by this recent NYT article.