In this game two agents, a red and a blue dot, have to gather green-dot apples.

Image: Google DeepMind/YouTube

Scientists at Google-owned DeepMind have found its AIs behave almost the way humans do when faced with scarce resources.

In a new study, DeepMind scientists plugged its AI agents, trained with deep reinforcement learning, into two multi-agent 2D games to model how conflict or cooperation emerges between selfish participants in a theoretical economy.

As DeepMind explains, they trained their AI agents to behave the way some economists model human decision making. That is, selfish and always rational.

"The research may enable us to better understand and control the behaviour of complex multi-agent systems such as the economy, traffic, and environmental challenges," DeepMind's researchers explain in a blog.

In one game two agents, a red and a blue dot, are faced with the task of gathering apples represented by green dots. The agents can simply collect apples together, suggesting cooperation, or they can 'tag' the other to prevent them collecting apples.

After several thousand rounds, they found that when there's an abundance of apples the agents collect as many as possible and leave each other alone. However, when DeepMind restricted the supply, the agents became more aggressive, figuring out that it may be optimal to block their rival to boost their chances of taking what's available.

"The Gathering game predicts that conflict may emerge from competition for scarce resources, but is less likely to emerge when resources are plentiful," they write in a new paper.

"These results show that agents learn aggressive policies in environments that combine a scarcity of resources with the possibility of costly action. Less aggressive policies emerge from learning in relatively abundant environments with less possibility for costly action," they note.

DeepMind also found that smarter agents with a larger network, enabling them to devise more complex strategies, tried to block their fellow gatherer more frequently, regardless of how much scarcity was introduced.

However, a second game called Wolfpack produced different behaviors when they were equipped to devise more complex strategies.

In this game, two wolves represented by red dots work together to capture the blue dot prey and face the risk of losing the carcass to scavengers.

If the wolves cooperate, they can get a higher reward since two wolves are better at protecting the catch than one. In this case, DeepMind found that a greater capacity to implement complex strategies resulted in more cooperation.

DeepMind found that in Wolfpack, cooperation behavior is more complex and requires a larger network size because agents need to coordinate hunting to collect team rewards.

Image: Google DeepMind/YouTube

They also found the wolves developed two different strategies for killing the prey and protecting the carcass.

"On the one hand, the wolves could cooperate by first finding one another and then moving together to hunt the prey, while on the other hand, a wolf could first find the prey and then wait for the other wolf to arrive before capturing it," they note in the paper.

DeepMind offers this explanation for why network size made the agents more competitive in the gathering game, yet more cooperative in the hunting game.

"In Gathering, defection behavior is more complex and requires a larger network size to learn than cooperative behavior. This is the case because defection requires the difficult task of targeting the opposing agent with the beam whereas peacefully collecting apples is almost independent of the opposing agent's behavior," they write.

"In Wolfpack, cooperation behavior is more complex and requires a larger network size because the agents need to coordinate their hunting behaviors to collect the team reward, whereas the lone-wolf behavior does not require coordination with the other agent and hence requires less network capacity," they write.

Read the original post:
Google's DeepMind: What can these battling AIs tell us about human behavior? - ZDNet

Most of us read the wrong things. As Haruki Murakami put it, reading what everyone else reads means youre probably going to think what everyone else thinks. All those books from high-school? Everyone else has read them too. The best-sellers? Same.

Thats not to say these books arent valuable. They are. Theyre just not going to help you get unique insights, see problems in a different way than others, or even help you solve more problems. They will, however, make you sound like youre smart because you can talk about the things everyone else is talking about. That said, there is the old adage: When you do what everyone else is doing, you shouldnt be surprised to get the same results everyone else gets.

While thinking the thoughts that other people have is enough to get a seat at the table, its not enough to win the game.

To win you need to see things that other people cant see. You need to connect things that other people cant connect.

Reading can help you develop insights, connections, and understanding that baffles others. To do this, you cant, however, follow in the same footsteps as everyone else because that leads you down the same path.

With that in mind, here are five books that youve probably never heard of (and one you have) that will change your life and enable you to see the world in a new light.

La Rochefoucaulds critical and pithy analysis of human behavior wont soon be forgotten. A list of people influenced by his maxims include Nietzsche, Voltaire, Proust, de Gaulle, and Conan Doyle. The readers best policy, Rochefoucauld suggests, is to assume that none of these maxims is directed at him, and that he is the sole exception. . After that, I guarantee that he will be the first to subscribe to them.

Ive never read this book in a cover-to-cover sense but Ive read each of the laws. More than that, Ive broken each of the laws. Ill give you an example. The first law is Never outshine the master. Once, I worked directly for a CEO. I worked as hard as I ever have to show off my talents and skills and at every turn it backfired over and over again. The lessonmake your masters appear more brilliant than they are and you will attain the heights of power. I wish I read this book earlier in my career, it certainly would have been helpful.

This book sat on my shelf for a year before I picked it up recently. This is the biography of Cyrus the Great, also known as Cyrus the Elder, who made the oldest known declaration of human rights. The book is full of leadership lessons. Heres an example.

Brevity is the soul of command. Too much talking suggests desperation on the part of the leader. Speak shortly, decisively, and to the pointand couch your desires in such natural logic that no one can raise objections. Then move on.

This no-nonsense collection of 20 letters from a self-made man to his son are nothing short of brilliant as far as Im concerned. This is a great example of timeless wisdom. The broad theme is how to raise your children in a world where they have plentybut the lessons apply to parents and non-parents alike. Check out a sample.

An autobiography of Nobel laureate Herbert A. Simon, a remarkable polymath who more people should know about. In an age of increasing specializing, hes a rare generalistapplying what he learned as a scientist to other aspects of his life. Crossing disciplines, he was at the intersection of information sciences. He won the Nobel for his theory of bounded rationality, and is perhaps best known for his insightful quote, A wealth of information creates a poverty of attention.

And heres one more just for good luck, even if youve probably heard of it:

OK, this is a bonus pick as I figured many of you might have read this already. However, the translation matters. Get this one. The best way to sum up this book is: A simple and powerful guide to life. This book was never intended for publicationit was for himself. How many people write a book of epigrams to themselves during a war? Get it. Read it. Live it.

This post originally appeared on Medium.

Read the original here:
Five books about human behavior that will change the way you see ... - Quartz

As part of their series on mobility, Deloitte explored how human behavior can cause delays in the adoption of new technology in the article Framing the future of mobility: Using behavioral economics to accelerate consumer adoption. Deloitte has predicted a shift in the automotive industry from personally owned, driver-driven cars to shared and self-driving vehicles. Despite the number of advantages generated from such a transformation, it could be met with skepticism because of limitations in our own human cognition.

Deloitte argues that the speed with which this future vision arrives likely hinges...on how quickly consumer expectations and behavior shift. The same research that revealed these change-prohibitive biases shed light on ways to overcome them and encourage consumers to welcome the future of mobility.

If/when the automotive shift that Deloitte anticipates comes to fruition, its not just the auto industry that will be majorly affected, but insurance, financing, technology, and energy industries as well. This isnt simply a change in how people use transportation, but one affecting government regulations and producing major infrastructure changes.

As ridesharing and self-driving transportation options become more prevalent, consumers of all ages could potentially benefit. The previously immobile generations who can not yet drive or are now unable to would no longer find themselves stranded, and families wouldnt have to worry about transporting them. Other societal benefits could result, like a decrease in traffic congestion and an increase in vehicle efficiency; resulting in reduced emissions and improved air quality.

Most importantly, autonomous vehicles would likely eliminate the element of human error, helping reduce the 30,000 deaths that occur each year during traffic accidents. A safer and more productive commutean average of 46 minutes per daycould reduce stress and be more affordable; Deloitte analysis shows that the cost of traveling per mile might decrease as much as two-thirds.

The positive results of an autonomous driving world could likely be abundant, but Deloitte cautions that just because a new technology offers benefits on paper does not mean customers will ultimately embrace it.

Continue Reading Below

ADVERTISEMENT

Studies in behavioral economics and social psychology have demonstrated that we as humans have a set of biases that affect the choices we make. Figure 1 in the article shows a list of these biases and their impact on how the future of mobility would be adoptedor, more specifically, why these biases could likely hinder the adoption of autonomous vehicles.

A loss aversion bias causes humans to overrate what we would lose compared to what we would gain from something new. This goes along with the endowment effect, where we overvalue things we already possess, and a status quo bias: a reluctance to change because we overvalue the current state.

These three biases together could cause individuals to feel like they are giving up more with their personally owned car than they would gain from a new autonomous driving state. To justify a change, the gain must overwhelm what is being giving up, so these biases make it even harder to achieve when you factor in the emotional attachment to a car. Trading a tangible good for a service also doesnt feel like a fair trade, so substituting personally owned vehicles for car/ridesharing may take longer than Deloittes initial time projections.

Three other types of biases related to risk would also predispose humans to resisting the change to a future mobility with autonomous driving. There is a risk miscalculation bias at play, which shows that humans are generally poor at assessing risk and assume the worst when faced with something new or unknown. In the instance of this new technology, there are no known effects as to how driverless vehicles will work, so it is perceived as more risky than it actually is.

The chart in figure 2 shows that the types of risk categorized as new, unknown, uncontrollable, involuntaryall of which would be associated with self-driving carsare viewed as the most risky. Regardless of the testing done by regulators or carmakers, the underlying technology of a self-driving car will likely remain mysterious to the average consumer. [T]he very nature of an autonomous vehicle makes it fundamentally uncontrollable (by the passenger, at least), which means customers are likely to see riding in them as particularly risky.

Likewise, an experience that can be controlled is an old risk, or is a known and observable technology that would automatically be viewed as less risky by the human brain. This is reflected in the optimism bias, where drivers overestimate their own ability and underestimate the probability of a bad event happening to them. Most drivers think they are better than the average driver and safer than they really are, which could reduce the likelihood that consumers will adopt self-driving cars due to safety reasons; they surely believe they are safer than trusting an unknown technology.

Another cognitive bias working against a future mobility system is the tendency to overemphasize a familiar or signature event that sticks out as the norm even though it may be an outlier. If a specific airline has a crash, people may easily associate that airline with crashing planes even though it may be a statistical anomaly and extremely rare. This tendency, known as the availability heuristic, might make a commuter focus on the few occasions when he was inconvenienced by ridesharing (by a long wait for his vehicle, for example) or a story of someone being harassed by a driver rather than the majority of instances when shared mobility was fast, convenient, and inexpensive.

After stepping into the psyche to see why we are predisposed to thinking in a certain way, Deloitte offers steps leaders can take to facilitate an accelerate adoption of autonomous driving technology. By manipulating the way a choice is presented or framed, we can overcome the aforementioned cognitive barriers.

Negative framing Using the loss aversion bias, we know losses are seen with more importance than gains. This method would involve making a consumer feel like they are missing out on something instead of gaining something. So a choice framed as costing time/money/lives instead of saving them would be more effective.

Aggregating When presenting data, expanding timeframes and aggregating costs over the longer period has more impact. Showing the amount of time or money that can be saved in a year seems a lot larger than the minutes or pennies from each day, so by changing the timeframe and forcing the consumer to look at the bigger picture can have a greater influence.

Creating social proofs Deloitte points out that we often look to the behavior of others for clues as to the correct course of action. As juvenile as it may sound, the saying everyone is doing it really does come into play here. By making it seem like our peers are participating, we are more likely to as well; especially in the case of a product that a consumer doesnt feel strongly about one way or another.

Using default options Pre-selected options give the illusion that something is the norm, so by making an option the standard selection, a consumer will be influenced to use it. Making an autonomous vehicle the default option could encourage consumers to use that technology like Uber does with the UberPool feature.

Packaging as an add-on According to Deloitte, research suggests that any new innovation is more readily accepted by consumers when it is packaged as an add-on to an existing, familiar item, rather than as a change to the central form and function of a product. By creating a familiar vehicle that has autonomous driving capabilities as an additional feature, it would mitigate the new-ness of such a technology and make it seem more acceptable.

How quickly the future of mobility takes hold in our society depends on a large number of factors; chief among them is the way it is marketed. By understanding the cognitive biases behind how consumers will perceive autonomous vehicles, decision-makers can alter their approach to make it more appealing and reduce the fear and hesitance that typically comes along with change.

More:
The Skeptical Consumer - How Behavioral Economics Can Influence the Adoption of Self-Driving Cars - Fox Business