Selmer Bringsjord (right): … Proving things. Discovering things. I don’t think that even a hair’s width of these things have been simulated in computational and cognitive science and in AI.

I remember asking James Moor, the Dartmouth professor who’s written quite a bit on AI: “You know. Jim, you really are a true believer in this stuff but can you tell me how much time you’re willing to give these AI people?”

I mean, if we give them another thousand years, and we still don’t have cognition as I’ve characterized it, captured computationally with the relevant artifacts and outputs produced… Are you going to be skeptical now?

Intelligent systems.

US Defence Advanced Research Projects Agency (DARPA) researchers released a custom version of Android designed to simplify inclusion of stringent safeguards by app developers, in a bid to address concerns around data privacy.

The open-source Privacy Enhancements for Android (PE for Android) platform was built by teams at Two Six Labs and Raytheon BBN Technologies as part of DARPA’s privacy-focused Brandeis programme.

Brandeis programme manager Joshua Baron told Mobile World Live the platform aims to address a knowledge gap among app developers, which researchers found often aren’t familiar with “the expected privacy disclosures or regulations that may guide their application’s use”.

FastNICs seeks to improve network stack performance by 100x or more, and to accelerate distributed applications like training deep neural networks.

John Keller

A deep learning-based model estimates attenuation maps directly from emission data, enhancing myocardial perfusion imaging on SPECT-only scanners.

Judging by the heatsinks, that system could preheat Huang’s oven by itself.