Large language models like GPT-3 are trained using vast bodies of text, much it taken from the internet, in which they encounter the best and worst of what people put down in words. That is a problem for today's chatbots and text-generation tools. The models soak up toxic language--from text that is racist and misogynistic or that contains more insidious, baked-in prejudices--as well as falsehoods. OpenAI has made IntructGPT the default model for users of its application programming interface (API)--a service that gives access to the company's language models for a fee. GPT-3 will still be available but OpenAI does not recommend using it.

Did you miss a session from the Future of Work Summit? Large language models (LLMs) such as OpenAI's GPT-3, which can "write" sentences that read nearly like they were written by a human, can be prompted to perform a range of writing tasks given only a few examples of the tasks. For example, LLMs have been used to create marketing materials and video game levels in addition to recipes, poetry, and movie scripts. But because LLMs learn to write from examples taken from sometimes toxic communities, they can fall victim to parroting misinformation, sexism, ageism, racism, and conspiracies. Efforts have been made to combat toxicity in LLMs -- with mixed results.

In December 2019, DeepMind's AI system, AlphaFold, solved a 50-year-old grand challenge in biology, known as the protein-folding problem. A headline in the journal Nature read, "It will change everything" and the President of the UK's Royal Society called it a "stunning advance [that arrived] decades before many in the field would have predicted". In this episode, Hannah uncovers the inside story of AlphaFold from the people who made it happen and finds out how it could help transform the future of healthcare and medicine. Thank you to everyone who made this season possible! Find Seasons 1 & 2 on YouTube: http://dpmd.ai/3geDPmL

When voice actor Heath Miller sits down in his boatshed-turned-home studio in Maine to record a new audiobook narration, he has already read the text through carefully at least once. To deliver his best performance, he takes notes on each character and any hints of how they should sound. Over the past two years, audiobook roles, like narrating popular fantasy series He Who Fights With Monsters, have become Miller's main source of work. But in December he briefly turned online detective after he saw a tweet from UK sci-fi author Jon Richter disclosing that his latest audiobook had no need for the kind of artistry Miller offers: It was narrated by a synthetic voice. Richter's book listing on Amazon's Audible credited that voice as "Nicholas Smith" without disclosing that it wasn't human. To Miller's surprise, he found that "Smith" voiced a total of around half a dozen on the site from multiple publishers--breaching Audible rules that say audiobooks "must be narrated by a human."

With his 2,141 SoundCloud followers, veteran U.K. psychedelic-music producer Darren Sangita didn't exactly wound Spotify when he pulled his music from the streaming service last month. "I'm a zed-list music celebrity," he says. But he couldn't support a company whose founder, Daniel Ek, invested more than $114 million in Helsing, a European security startup that manufactures artificial-intelligence software to "keep liberal democracies from harm." "The circumstantial evidence points to a massive investment in military/AI tech," says Sangita, who runs indie label Sangita Sounds. "I was just enormously disappointed. Are there not any other alternatives that Mr. Ek could have possibly thought about investing in to make the world a more beautiful and perhaps a safer place to live?"

The sector's growth is expected to be fueled by AI technologies, including machine learning, deep learning, natural language processing (NLP) and …

Kedro, A Python Machine Learning Pipeline Framework Developed By McKinsey, Has Been Donated To The Linux Foundation.

Meta said it has built an AI supercomputer that will be the world's fastest by mid-2022. Called AI Research SuperCluster, it will be used to build AI models that can ingest vast amounts of information and process it like a human brain would, complete with contextual understanding, according to Meta's VP of AI. Potential use cases include the ability to "seamlessly analyze text, images and video together," augmented reality features and near-instantaneous flagging of harmful content, all of which could help build experiences in the metaverse.

The most common way to listen to recorded music nowadays is via streaming platforms which provide access to tens of millions of tracks. To assist users in effectively browsing these large catalogs, the integration of Music Recommender Systems (MRSs) has become essential. Current real-world MRSs are often quite complex and optimized for recommendation accuracy. They combine several building blocks based on collaborative filtering and content-based recommendation. This complexity can hinder the ability to explain recommendations to end users, which is particularly important for recommendations perceived as unexpected or inappropriate.

Although scaling up language model size has reliably improved performance on a range of NLP tasks, even the largest models currently struggle with certain reasoning tasks such as math word problems, symbolic manipulation, and commonsense reasoning. This paper explores the ability of language models to generate a coherent chain of thought -- a series of short sentences that mimic the reasoning process a person might have when responding to a question. Experiments show that inducing a chain of thought via prompting can enable sufficiently large language models to better perform reasoning tasks that otherwise have flat scaling curves.