Last Week in AI #192: Meta predicts shape of 600 million proteins, a teacher allows AI tools in exams, Twitter fires AI ethics team, and more!
Last Week in AI #192: Meta predicts shape of 600 million proteins, a teacher allows AI tools in exams, Twitter fires AI ethics team, and more!
Meta AI releases new database of previously unknown protein structures, what happened when students were allowed to use AI to help write essays, Twitter's entire AI watchdog team disbanded
Last Week in AI #192: Meta predicts shape of 600 million proteins, a teacher allows AI tools in exams, Twitter fires AI ethics team, and more!
Top News
AlphaFold’s new rival? Meta AI predicts shape of 600 million proteins
Last year, DeepMind published AlphaFold 2, a method for accurately predicting protein structures, and allowed the company to build the AlphaFold Database with 200 million protein structures. Now, Meta AI has developed a method that is 60 times faster at predicting protein structures and generates about 617 million predictions from just a few examples. Meta AI credits its method to a combination of large language models and sequence-to-structure relationships, and its researchers are already using their method to predict protein structures for unknown microorganisms from environmental DNA. It's clear that the protein prediction landscape is far from being settled, and new methods will continue to emerge.
Our take: This is certainly a major advancement in the field of protein prediction. While the specific AlphaFold-based approach used by Meta is similar to DeepMind's AlphaFold, the use of a language model instead of text data gives it an intuitive understanding of protein sequences which makes it more powerful than AlphaFold. While Meta's predictions are limited to only short sequences, the speed and affordability of these predictions will enable more exciting research down the line.
A teacher allows AI tools in exams – here’s what he learned
Eighth-grade students at a German high school were allowed to take an exam where they wrote written arguments using AI tools. This was after the school has already been experimenting with the use of digital devices and AI tools over the last two pandemic years. They learned about the advantages and disadvantages of different AI systems in class in the weeks leading up to the exam. It was found that the students are more reflective when dealing with AI texts and see AI tools more as support than as homework automation. On the other hand, AI texts overwhelmed students who did not know how to construct an argument.
Our Take: With more AI methods surfacing in our daily lives, perhaps the future of education will be AI-dependent. While the conservative thought is to ban digital tools and AI, this school has wholeheartedly accepted the reality and is trying to use it best. It is important to note that while students used AI tools to generate texts, they had to do an equal amount of research on the Internet to verify all the information. So, it can be argued that they ended up learning more in the process of using AI tools than they would have in conventional exams.
Welp, There Goes Twitter's Ethical AI Team, Among Others as Employees Post Final Messages
The Chief Twit has kicked more birds out of the nest. In the wake of the heartbreaking departures of Rahul Ligma and Daniel Johnson, Twitter's Ethical AI team has also evaporated. Multiple members of the META team--formed last year to audit Twitter's algorithms to investigate harms and biases--posted on Twitter saying they were no longer at the company, and one former worker suggested the entire watchdog team was being disbanded.
Our take: To many, running a tight ship means cutting off parts of an organization that are not deemed crucial to the business. But Twitter is, for better or for worse, more than just a business. Its algorithms and technology--whatever one may think of them--affect far more than the pockets of its shareholders. I think it's fine for organizations to be run in this way to an extent, but somebody has taken social responsibility. The transparency and accountability that the META team represented are vital.
Other News
Research
Google wants robots to generate their own code - "There are countless big problems left to solve in the world automation, and robotic learning sits somewhere near the top. While it’s true that humans have gotten pretty good at programming systems for specific tasks, there’s a big, open-ended question of: and then what?"
Using AI to compress audio files for quick and easy sharing - "Compression is an important part of the internet today, because it enables people to easily share high-quality photos, listen to audio messages, stream their favorite shows, and so much more."
The method may help clinicians determine which patients would likely benefit from aggressive treatment even at early stages of disease - "The method may help clinicians determine which patients would likely benefit from aggressive treatment even at early stages of disease"
Artificial intelligence and molecule machine join forces to generalize automated chemistry - "Artificial intelligence, "building-block" chemistry and a molecule-making machine teamed up to find the best general reaction conditions for synthesizing chemicals important to biomedical and materials research -- a finding that could speed innovation and drug discovery as well as make complex chemistry automated and accessible."
New AI Research Could Drive Progress in Medicine and Clean Energy - "Our new database of protein structures could lead to discoveries that help cure diseases and clean the environment."
Machine learning facilitates “turbulence tracking” in fusion reactors - "Fusion, which promises practically unlimited, carbon-free energy using the same processes that power the sun, is at the heart of a worldwide research effort that could help mitigate climate change."
Tracking trust in human-robot work interactions - "Researchers use functional near-infrared spectroscopy to monitor participant responses"
Machine Learning Helps to Solve Problems in Heliophysics - "A new special collection invites papers pertaining to the use of machine learning techniques in all sub-fields of heliophysics."
Machine-Learning Tool Solves Metamaterial Jigsaw - "A new tool can determine whether a collection of building blocks will assemble into a mechanically sound structure."
In the latest human vs. machine match, artificial intelligence wins by a hair - "A protein scientist, who competed against a computer program, says machine learning will advance biotechnology"
Nvidia’s eDiffi is an impressive alternative to DALL-E 2 or Stable Diffusion - "Nvidia’s eDiffi is a generative AI model for text-to-image and beats alternatives like DALL-E 2 or Stable Diffusion, according to the company. "
Applications
Lockheed Martin Successfully Completes First Autonomous Black Hawk Helicopter Flight - "While DARPA’s Liberty Lifter seaplane can be used to transport large payloads, Lockheed Martin’s autonomous Black Hawk helicopter can be piloted without actual pilots."
Exhibit aims to present AI images as real art - "The artists and curators of "Artificial Imagination" say it shows AI art is legitimate"
AI-Powered 'Iron Man' Boots Could Someday Help You Walk, Run Faster - "Talk about a new step forward: Stanford engineers have developed robotic boots that help you walk and run faster with less effort. Equipped with a motor, the boots use artificial intelligence to provide a personalized boost that’s just right for whoever is wearing them. "
Italy Turns to AI to Find Taxes in Cash-First, Evasive Culture - "Italy’s aggressive use of artificial intelligence to close Europe’s biggest tax gap is starting to make inroads, but still must overcome a national legacy of tax avoidance and strict EU data privacy laws."
Google’s new prototype AI tool does the writing for you - "Remember that time Google showed off its artificial intelligence prowess by demoing conversations with Pluto and a paper airplane? That was powered by LaMDA, one of Google’s latest-generation conversational AI models."
Google’s wildfire detection is available in US, Mexico, Canada and parts of Australia - "At an AI-focused press event today in New York, Google announced that it’s bringing its AI-powered wildfire detection system to the U.S. Canada, Mexico and parts of Australia."
Business
Digital assets marketplace Creative Fabrica launches generative AI tool - "Creative Fabrica, a marketplace for digital files like print-on-demand assets, fonts and graphics, announced today it will launch its own generative AI tool."
Google Pays $100M for AI Avatar Startup Alter - "Google has reportedly acquired artificial intelligence (AI) avatar startup Alter — which was formerly known as Facemoji — for $100 million. "
Ford, VW seeking buyer for Argo AI’s lidar unit - "Ford and Volkswagen are trying to squeeze any remaining value out of Argo AI, the autonomous vehicle startup the two automakers invested billons in before abruptly shutting it down last week."
Replit’s Ghostwriter AI can explain programs to you—or help write them - "Today, Replit announced Ghostwriter, an AI-powered programming assistant that can make suggestions to make coding easier."
OpenAI will give roughly 10 AI startups $1M each and early access to its systems - "OpenAI, the San Francisco-based lab behind AI systems like GPT-3 and DALL-E 2, today launched a new program to provide early-stage AI startups with capital and access to OpenAI tech and resources. Called Converge, the cohort will be financed by the OpenAI Startup Fund, OpenAI says."
Researchers encourage retailers to embrace AI to better service customers - "Three QUT researchers are part of an international research team that have identified new ways for retailers to use Artificial Intelligence in concert with in-store cameras to better service consumer behaviour and tailor store layouts to maximise sales."
Labrador Systems deploys its first assistive elder-care robots - "We’ve been keeping tabs on Labrador Systems since we caught a very early demo of its elder care-focused technology in a hotel suite several CESes ago."
Concerns
Invasive Diffusion: How one unwilling illustrator found herself turned into an AI model - "Last weekend, Hollie Mengert woke up to an email pointing her to a Reddit thread, the first of several messages from friends and fans, informing the Los Angeles-based illustrator and character designer that she was now an AI model."
Scientists Increasingly Can't Explain How AI Works - "AI researchers are warning developers to focus more on how and why a system produces certain results than the fact that the system can accurately and rapidly produce them."
Study urges caution when comparing neural networks to the brain - "Neural networks, a type of computing system loosely modeled on the organization of the human brain, form the basis of many artificial intelligence systems for applications such speech recognition, computer vision, and medical image analysis."
Machine-learning systems are problematic. That’s why tech bosses call them ‘AI’ - "Pretending that opaque, error-prone ML is part of the grand, romantic quest to find artificial intelligence is an attempt to distract us from the truth"
Explainers
History Of AI In 33 Breakthroughs: The First Expert System - "In the early 1960s, computer scientist Ed Feigenbaum became interested in “creating models of the thinking processes of scientists, especially the processes of empirical induction by which hypotheses and theories were inferred from data."
History Of AI In 33 Breakthroughs: The First ‘Thinking Machine’ - "In 1308, Catalan poet and theologian Ramon Llull completed Ars generalis ultima (The Ultimate General Art), further perfecting his method of using paper-based mechanical means to create new knowledge from combinations of concepts."
Fun
Machine Learning researchers are using wonderful terrible puns to title papers - "Injecting fun and play into life at any opportunity is a crucial part of the creative spirit that defines humanity. As a species, we are capable of truly terrible and wonderful things."
The Infinite Conversation - "an AI generated, never-ending discussion between Werner Herzog and Slavoj Žižek. Everything you hear is fully generated by a machine. The opinions and beliefs expressed do not represent anyone. They are the hallucinations of a slab of silicon."
Copyright © 2022 Skynet Today, All rights reserved.