AI Closer to Mimicking Human Abilities (and Musk’s Fears Called Out) – This Week in Artificial Intelligence 12-26-15

Daniel Faggella

Daniel Faggella is Head of Research at Emerj. Called upon by the United Nations, World Bank, INTERPOL, and leading enterprises, Daniel is a globally sought-after expert on the competitive strategy implications of AI for business and government leaders.

AI Closer to Mimicking Human Abilities (and Musk's Fears Called Out) - This Week in Artificial Intelligence 12-26-15

1 – Elon Musk Nominated for ‘Luddite’ of the Year Prize over Artificial Intelligence Fears

The Information and Technology Innovation Foundation, a Washington-based think tank, has nominate Tesla’s Elon Musk for its annual Luddite Awards. This announcement ironically comes after several innovative achievements, including TeslaX’s announcement this past week of successfully launching and returning a rocket to Earth. The Luddite Award is handed out to the   person who the group considers most set on holding back the introduction of new technologies. The nomination is linked to Musk’s, Gates’ and Hawking’s warnings about the existential threats of artificial intelligence. Other nominees include “advocates seeking ban on killer robots” and “States limiting automatic license plate readers”.

(Read the full article at The Guardian)

2 – Google Plans New, Smarter Messaging App

Google is hard at work behind the scenes on a messaging and chatbot service, as it tries to catch up with Facebook and other rivals in the already competitive domain. Messaging apps and chat bots have historically been a weakness for Google, which has had lukewarm success with apps like Google Hangout and Google Messenger. While a Google spokeswoman declined to comment, a source close to the matter stated that Google’s Nick Fox is leading a development team. In October, Fox’s offer to buy 200 Labs, Inc. was declined by the small startup, which specializes in building advanced, responsive chatbots. Those close to the source stated that Google is looking to build something similar.

(Read the full article on The Wall Street Journal)

3 – Artificial intelligence startup Arya.ai raises $750k from YourNest, VentureNursery

Arya.ai, a Mumbai-based AI startup, has raised $750,000 in an initial round of funding from VentureNursery and YourNest Angel Fund. The company, which creates deep learning algorithms and other sophisticated tools for developing AI systems, will focus the funds on new products. Cofounder and CEO Vinay Kumar said,

“Using AI, developers can build applications that can solve most complex problems. However, the biggest pain point in building reliable AI applications is solving multiple challenges they encounter in the process,”

According to Kumar, Arya.ai works across several industries and is planning to launch into the automobile, military, and defense industries.

(Read the full article on VCCircle)

4 – Darwin: a Neuromorphic Hardware Co-Processor Based on Spiking Neural Networks

Researchers at Zhejiang University and Hangzhou Dianzi University announced a mutually developed Darwin Neural Processing Unit (NPU), a neuromorphic hardware co-processor that more closely mimics neural networks in primitive brains. The NPU, which uses complementary metal-oxide-semiconductor (CMOS) technology (used for creating integrated circuits), is compared to a simple animal brain that supports up to 2048 neurons, 4 million synapses, and 15 synaptic delays. The pioneering effort is based on the desire to build a more effective and energy-efficient. The team plans to use the Darwin NPU as a Processing Element in a Network-on-Chip (NoC) architecture, hoping to scale from thousands to potentially millions of neurons.

(Full article on Zhejiang University News is no longer available as to this article update on September 2017)

5 – Machines that Learn like People

A research team from MIT’s McGovern Institute for Brain Research has theoretically proven that their model of an object-recognition system can makes accurate determinations of visual objects based on only a few examples, a feat still dominated by humans. Their findings will be published in the journal Theoretical Computer Science, which follows a previous paper published in October that demonstrates how their model aligns with empirical evidence about how the brain works. The hypothesis has important implications for bringing today’s machine learning to meet the level of primate visual systems, a function that could help produce much more powerful AI.

(Read the full article on MIT News)

 

 

Subscribe