GOOGLE'S ARTIFICIAL INTELLIGENCE has got no nose. Using a data set of close to 5,000 molecules identified by perfumers. In a new paper from the Google Brain Team, the researchers explained the process. In short, perfumers labelled molecules with phrases like "buttery" and "tropical", and trained the algorithm with around two-thirds of the labelled scents. They then fed the remaining third to the AI unlabeled, and the algorithms successfully guessed how these mystery molecules would smell. This is no mean feat (as distinct from "mean feet", which have their own unique scent).

GOOGLE HAS announced that it is making its artificial intelligence (AI) and machine learning (ML) courses available to everyone, beyond Mountain View. The new Learn with Google AI portal will let anyone with an interest, at almost any level, learn how to make the most of the glorious new horizon of AI and neural networking. Previously, the courses were designed for internal use to train Googlers, as Google continues its moves to be an "AI First" company. Zuri Kemp from Google AI explains: "From deep learning experts looking for advanced tutorials and materials on TensorFlow, to "curious cats" who want to take their first steps with AI, anyone looking for educational content from ML experts at Google can find it here." There's also a Machine Learning Crash Course (MLCC) which "provides exercises, interactive visualizations, and instructional videos that anyone can use to learn and practice ML concepts."

NVIDIA HAS been teaching MPs how to create neural networks as part of an AI Workshop organised in conjunction with the Big Innovation Centre. A cross-party group including Stephen Metcalf MP and Lord Clement Jones CBE, co-chairs of the All Party Parliamentary Group on Artificial Intelligence (APPG AI), was tasked with learning about AI, what it really means, and then taking the information back to inform the discussion with constituents. Of course, that doesn't just mean being able to respond if an MP's surgery bears the question "So, do I need to live in fear of Skynet?" but also more practical pursuits like a local business looking at bringing AI to their business. Stuart Wilson, AI and Supercomputing director at Nvidia, said: "Artificial intelligence represents the biggest technological and economic shift in our lifetime. It is of national importance that policymakers understand the core components, capabilities and limitations surrounding the modern AI boom."

DURHAM POLICE will look to artificial intelligence (AI) to help it make decide whether to remand or release criminal suspects. The AI system, which has been tested by the Durham Police Force since 2013 and trained on five year's worth of data, assigns a risk category to suspects. The BBC reports that the system has proved itself to be 98 per cent accurate in assessing suspects as low risk suspects, and 88 per cent accurate in flagging high-risk ones, although it is not made clear exactly how this is measured. The AI system, known as Harm Assessment Risk Tool (HART), has been tested by custody officers in the course of their work, but has not yet assisted their decisions. However, this is about to change, said Sheena Urwin, head of criminal justice at Durham Constabulary.

GOOGLE HAS made a tool available that it hopes will aid devs in the battle against trolling and harassment. Jigsaw, a long standing project in Google's arsenal uses machine learning technology to spot behaviours online that shouldn't be. Its latest tool, called Perspective, measures'toxicity' of comments based on trawling through millions of existing rants and'teaching' the software by serving them up to volunteers to grade. With publishers including Wikipedia, The Guardian, The Economist and The New York Times already testing the tool, it could become a powerful weapon in the battle against online troublemakers. Alternatively, one could argue it might be a barrier to free speech.

GOOGLE'S ARTIFICIAL INTELLIGENCE (AI) has already started to show some evidence of personality traits, but now DeepMind is learning to show signs of aggression when it thinks it's not going to get its own way. Sound like anyone you know? Where multiple instances of DeepMind are running they can, on the other hand, agreed to work together for a common goal should they believe there's more to be gained by doing so. It may occur to you as it has to us, that words like "aggression" and phrases like "working together" are exactly the sort of thing that doomsayers pick up on in predicting mankind's demise at the hands of robot overlords. However, this early research based primarily around game theory is experimental in nature and about as scary as BluTac.