A new computational model performs at human levels when subjected to standard intelligence test, making artificial intelligence (AI) system at par with human understanding capabilities. Researchers from Northwestern University built the new computational model on CogSketch, an artificial intelligence platform, that has the ability to solve visual problems and understand sketches in order to give immediate and interactive feedback. "The model performs in the 75th percentile for American adults, making it better than average," said Ken Forbus of Northwestern University, adding "The problems that are hard for people are also hard for the model, providing additional evidence that its operation is capturing some important properties of human cognition." Researchers noted that developing artificial intelligence systems that have this ability not only provides new evidence for the importance of symbolic representations and analogy in visual reasoning, but it could potentially shrink the gap between computer and human cognition. "Most artificial intelligence research today concerning vision focuses on recognition or labelling what is in a scene rather than reasoning about it," Forbus noted.

During the contest, which played out over a matter of hours, one bot proved it could find and exploit a particularly subtle security hole similar to one that plagued the world's email systems a decade ago--the Crackaddr bug. Until yesterday, this seemed beyond the reach of anything other than a human. "That was astounding," said Mike Walker, the veteran white-hat hacker who oversaw the contest. "Anybody who does vulnerability research will find that surprising." In certain situations, the bots also showed remarkable speed, finding bugs far quicker than a human ever could.

"Experimentation is the least arrogant method of gaining knowledge. The experimenter humbly asks a question of nature." Artificial Intelligence (AI) is coming for your job. But is it coming for the job of your photographer or inventor? The driver-less cars, automated factories, and automated laboratories of today may give way to AI capable of thinking, writing, creating or even diagnosing disease.

As business, policy, and technology leaders gathered at the annual World Economic Forum in Davos, Switzerland this year, the rise of populism with Brexit in the UK and the election of Donald Trump in the US drove discussions about the pros and cons of globalization. While globalization has improved the living conditions of vast swaths of this planet's population, it has also led to shifting employment patterns, as jobs leave the US for China and other low-wage countries. However, the Davos cognoscenti believe wage inequality is only part of the problem. Hand-in-hand with globalization is the topic of automation – seen as more of a culprit for increasing inequality than the usual scapegoat, low-cost labor. "[Automation] particularly represents a challenge for people in our economy with low skills, particularly the older workers who don't feel able to embrace and learn new skills and new technologies," says Philip Hammond, the UK Chancellor of the Exchequer, according to an article in the Washington Post.

When addressing the question of what it means for an algorithm to learn, one can imagine many different models, and there are quite a few. This invariably raises the question of which models are "the same" and which are "different," along with a precise description of how we're comparing models. We've seen one learning model so far, called Probably Approximately Correct (PAC), which espouses the following answer to the learning question: An algorithm can "solve" a classification task using labeled examples drawn from some distribution if it can achieve accuracy that is arbitrarily close to perfect on the distribution, and it can meet this goal with arbitrarily high probability, where its runtime and the number of examples needed scales efficiently with all the parameters (accuracy, confidence, size of an example). Moreover, the algorithm needs to succeed no matter what distribution generates the examples. You can think of this as a game between the algorithm designer and an adversary. First, the learning problem is fixed and everyone involved knows what the task is. Then the algorithm designer has to pick an algorithm. Then the adversary, knowing the chosen algorithm, chooses a nasty distribution over examples that are fed to the learning algorithm. The algorithm designer "wins" if the algorithm produces a hypothesis with low error on when given samples from . And our goal is to prove that the algorithm designer can pick a single algorithm that is extremely likely to win no matter what the adversary picks.

Machine learning has been a constant on tech trend lists for years. This year, it's time to embrace what humans can learn by interacting with machine learning. As Google's head of Machine Intelligence, Blaise Aguera y Arcas noted in a recent Medium article: "Machine intelligence will expand our understanding of both external reality and our perceptual and cognitive processes." In the spring of 2016, Google's AlphaGo software, fueled by machine learning, beat the world's greatest human Go player, Lee Sedol. The victory was a major milestone for a specific type of AI, called deep neural networks, that is more closely modeled on the way humans think.

Eric is responsible for the development of strategic products, services and solutions. He joined Esker in 2002 as Director of Marketing Communications, and in 2005, extended his responsibilities to include product management. Many businesses have yet to understand the full potential of machine learning and cognitive computing for back office processes. The benefits range far beyond simply increased productivity and faster supply chains. Today's advanced algorithms have the computational power and speed to gather, analyze and manage vast repositories of data.

Google's AI research lab, Google Brain is reportedly developing AI (artificial intelligence) software that can build more machine learning software. The ultimate aim to design such software is to reduce the costs of hiring experts for making machine learning software and make it more accessible and efficient in the future by spreading the benefits of the technology far and wide. If using AIs techniques become more practical, they could increase the speed at which new AIs can be made and implemented across the economy. Currently, companies are paying a premium to hire machine-learning experts, who are in short in supply. In recent months, several other groups have also reported progress on getting learning software to make learning software.

In this week's Abundance Insider: Self-organizing drone swarms, synthetic stem cells, and an AI that can detect heart failure better than human doctors. I'm launching an online course with SUCCESS Magazine called Xponential Advantage. It's aimed to inspire, educate and guide a new breed of "Exponential Entrepreneurs," and is an expansion of my core content from Abundance and BOLD, the keynotes I give to Fortune 500 executive teams, and some of the material I teach executives who attend Singularity University. These are the areas I truly believe an exponential entrepreneur can leverage to have a billion-person impact. What it is: Scientists at the London Institute of Medical Services have created an AI capable of predicting with 80% accuracy which patients would die of pulmonary hypertension within a year, beating the average doctor's prediction accuracy by about 20%.

What do you think of when you think about chatbots? If you're like many others, it's possible that you think of them as nothing more than that annoying little window that pops up when you're visiting a website. You know what I'm talking about - the one that claims to be able to answer your questions. You may have had both positive and negatives experiences with chatbots. Some of you may even ignore them altogether.