A New Artificial Intelligence Makes Mistakes—on Purpose

A chess program that learns from human error might be better at working with people or…

New Machine Learning Theory Raises Questions About the Very Nature of Science

A novel computer algorithm, or set of rules, that accurately predicts the orbits of planets in…

Who needs a supercomputer? Your desktop PC and a GPU might be enough to solve some of the largest problems | ZDNet

A new method significantly reduces the amount of memory needed for brain simulations, freeing some AI…

‘Weird new things are happening in software,’ says Stanford AI professor Chris Re | ZDNet

The time has come, said Re, for AI scientists to focus on higher-level tasks such as…

An AI saw a cropped photo of AOC. It autocompleted her wearing a bikini.

Image-generation algorithms are regurgitating the same sexist, racist ideas that exist on the internet. Read more…

How Transformers work in deep learning and NLP: an intuitive introduction

Transformers work in deep learning and nlp: an intuitive introduction. How did we go from attention…

Scientists look to harness the power of AI and the mind combined

European Commission gives project £3m to explore a new kind of technology. Is it worth it?…

Here’s what happened when AI and humans met in a strawberry-growing contest

Top strawberry growers were pitted against data scientists in a smart-agriculture competition organized by Chinese e-commerce…

Fruit Fly Brain Hacked For Language Processing

A simulated fruit fly brain has learnt to perform natural language processing tasks. Read more at…

Microsoft patent details tech that could turn dead people into AI chatbots

It’s Black Mirror come to life. Read more at CNET…

AI-powered text from this program could fool the government

Volunteers couldn’t tell AI-generated comments from those penned by humans. Read more at Ars Technica…

AI Weekly: Meet the people trying to replicate and open-source OpenAI’s GPT-3

A grassroots group of researchers is attempting to create an open source language model as powerful…

Google trained a trillion-parameter AI language model

Researchers at Google claim to have trained a natural language model containing over a trillion parameters.…

AI models from Microsoft and Google already surpass human performance on the SuperGLUE language benchmark

AI language models have surpassed human performance on a popular benchmark. But what does that really…

MuZero: Mastering Go, chess, shogi and Atari without rules

In 2016, we introduced AlphaGo, the first artificial intelligence (AI) program to defeat humans at the…

DeepMind Has Reconciled Existing Neural Network Limitations To Outperform Neuro-Symbolic Models | MarkTechPost

DeepMind Has Reconciled Existing Neural Network Limitations To Outperform Neuro-Symbolic Models Read more at MarkTechPost…

Spotting elephants from space: A satellite revolution

Using the highest resolution satellite imagery currently available—Worldview 3—from Maxar Technologies and deep learning, (TensorFlow API,…

Deepfake queen to deliver Channel 4 Christmas message

Channel 4 will use deepfake technology for its alternative to the traditional Christmas broadcast. Read more…

Artificial intelligence solves Schrödinger’s equation

A team of scientists at Freie Universität Berlin has developed an artificial intelligence (AI) method for…

At NeurIPS 2020, researchers proposed faster, more efficient alternatives to backpropagation

Backpropagation has led to impressive gains in AI and machine learning. But it has limitations researchers…