As machine learning models become more general, we need to characterise them in richer, more meaningful ways. We describe a method to infer the cognitive profile of a system from diverse experimental data. To do so, we introduce measurement layouts that model how task-instance features interact with system capabilities to affect performance. These features must be triangulated in complex ways to be able to infer capabilities from non-populational data -- a challenge for traditional psychometric and inferential tools. Using the Bayesian probabilistic programming library PyMC, we infer different cognitive profiles for agents in two scenarios: 68 actual contestants in the AnimalAI Olympics and 30 synthetic agents for O-PIAAGETS, an object permanence battery. We showcase the potential for capability-oriented evaluation.

But getting the right data to make AI and machine learning algorithms — and understanding it — is where many organizations are slipping up, a recent …

Artificial intelligence just doesn't pop up when you install tools and software. It takes planning and, most of all, it takes data. But getting the right data to make AI and machine learning algorithms -- and understanding it -- is where many organizations are slipping up, a recent study finds. Organizations face difficulties with data silos, explainability, and transparency, a study of 150 data executives commissioned by Capital One and Forrester Consulting finds. They say internal, cross-organizational, and external data silos slowed machine learning deployments and outcomes.

IT is facing significant talent shortages, and new human resources AI talent recruiting systems are touted as being able to help. How do these systems work, and are they effective? The purpose of artificial intelligence hiring and talent scouting systems is to reduce the amount of work that HR or IT conducts in the activities of talent seeking, candidate evaluation and hiring. For example, if you're looking for a senior project manager, it's not uncommon to receive 300 or 400 resumes. All these candidates are applying because they believe they have the experience and the requisite skills to do the job you want to fill.

As progress in AI continues to advance at a rapid pace, it is crucial to know how advanced systems will make choices and in what ways they may fail. Machines can already outsmart humans in some domains, and understanding how to safely build systems which may have capabilities at or above the human level is of particular concern. One might suspect that superhumanly-intelligent systems should be modeled as as something which humans, by definition, can't outsmart. However, as a challenge to this assumption, this paper presents the Achilles Heel hypothesis which states that highly-effective goal-oriented systems -- even ones that are potentially superintelligent -- may nonetheless have stable decision theoretic delusions which cause them to make obviously irrational decisions in adversarial settings. In a survey of relevant dilemmas and paradoxes from the decision theory literature, a number of these potential Achilles Heels are discussed in context of this hypothesis. Several novel contributions are made involving the ways in which these weaknesses could be implanted into a system.

How Artificial Intelligence will revolutionize financial services is a hot topic among Wall Street technologists. But, just as a home renovation might be delayed by the discovery of extensive foundation problems that need to be repaired first, the financial services industry must improve the quality of its data before it can realize the benefits of AI. Early front-office AI applications, such as Deutsche Bank using AI to predict equity pricing, have won headlines. But there is also significant momentum around applying AI to post-trade processing, compliance and risk management: Instead of manually searching for a needle in a haystack to learn why a trade failed to identify a real issue, AI can pre-emptively flag problems and solve the exception. However, while many capital markets firms have explored AI initiatives, few of these pilot programs (less than 15% according to Forrester Research) have made it into production to yield real business value.

Emergence of work Robo advisors If you're not Rumpelstiltskin, you've no doubt noticed the prolific rise of the talking speaker. There seems little to technically hold this back and from a cost optimization and process automation perspective, it makes all the sense in the world. Augmented Reality meets GoToMeeting A little play off #1, but if you mash what's going on with AR with the prohibitive cost of dedicated telepresence rooms, you get a budget busting way to see a screen with you and your colleagues sitting at the same table and interacting. You'll probably also be able to attend meetings in virtual person through a physically present android of some semblance (R2D2 style or perhaps like that annoying holographic person in the airports). Self-driving electric car innovations convergence hit a feverish pitch Chunky four German auto armies (Audi, Mercedes, Porsche, and BMW) release competing vehicles resulting in Telsa stock being challenged over their stretched diversification strategy and the need to refocus on the immediate autonomous competitive onslaught.