Last Week in AI #167: Enzyme developed with AI to decompose plastic, Hugging Face reaches $2 billion valuation, new ambitious EU AI Act, and more!

AI helped researchers modify enzymes for plastic recycling, Hugging Face closes Series C funding run, EU proposes sweeping AI legislation to curb harmful uses

Top News

Scientists Discover Method to Break Down Plastic in Days, Not Centuries

Years ago, scientists sifting through plastic bottle debris discovered a bacterium that can degrade polyethylene terephthalate (PET), a compound used in plastic bottles that often ends up in landfills. The environmental impacts are clear, and PET is currently recycled by melting or reshaping--but its properties deteriorate each time it is recycled. But enzymatic depolymerization could present another promising way to mitigate waste, and researchers at UT Austin used a neural network to help them decide how to modify the protein scaffold of an enzyme to make it stable at higher temperatures and boost its catalytic activity. Their PETase can degrade 51 different PET products, including whole plastic containers and bottles. The enzyme can depolymerize PET at least twice as fast and at lower temperatures than the next best-engineered enzyme.

Our Take: Areas like enzyme engineering present a very exciting application area for machine learning. ML systems can make predictions about and aid design choices for biological systems like enzymes, which is an incredibly powerful ability. I think we are far from utilizing this capability to its full extent and hope we'll see plenty more research along these lines in the near future.

Hugging Face reaches $2 billion valuation to build the GitHub of machine learning

Hugging Face announced this week that it raised a $100 million round of venture financing, which makes Hugging face worth $2 billion. The company also made its debut appearance on Forbes’ AI 50 list. The company has become the main repository for all things related to machine learning; one can find pre-trained models, and datasets for various tasks, and also participate in the developer community with their own model. They intend to add more resources with a mission to make machine learning accessible and productive for beginners and experts alike. Other than stepping up its efforts in the realm of open-source, Hugging Face is also exploring the domain of ethical AI, where they are looking into removing bias from data used for training these machine learning algorithms.

Our Take: In many ways, Hugging Face is leading the ‘open and collaborative’ evolution of ML. Moreover, the fact that as of today more than 10,000 companies use their platform and that one of their libraries on Transformers on Github got over 62,000 stars and 14,000 likes are in itself a testament to the success and utility of the company. I think efforts like this can also be a push for the research community, especially the deep learning and AI community to ensure their work is reproducible, considering the vast amount of research work in these domains generated every month.

A quick guide to the most important AI law you’ve never heard of

EU recently proposed the new ambitious AI Act, which if passed, would be the most comprehensive AI legislation yet. The bill asks for extra checks for AI applications that have a high risk of harm, like systems that grade exams, recruit employees, or make decisions in the justice system. Among other things that may be included in the final bill are requiring companies to notify users when deepfakes are present, as well as restricting law enforcement usage of predictive policing and facial recognition. Some requirements in the current draft are technically impossible to comply with, like requiring humans to “fully understand” how AI systems work, and additional debates will take place in the months to come. EU lawmakers expect at least another year before this Act is finalized, and a couple more beyond that before when companies have to comply.

Our take: More clear guidelines on high-risk applications of AI should be welcomed, but the main debates might be on what exactly constitutes a high-risk application. Given AI’s fast development pace and wide implications, it will be tricky to write legislation that is both broad enough to be applicable to future AI technologies and specific enough to be enforceable and reasonable for compliance. Whatever EU passes will have significant impacts throughout the rest of the world, or at least the Western world, where international tech companies that need to comply with EU regulations will likely also carry over some of these measures to their other markets as well.

Other News

Research

Meta AI Introduces ‘Make-A-Scene’: A Deep Generative Technique Based On An Autoregressive Transformer For Text-To-Image Synthesis With Human Priors - "In recent years, the research related to text-to-image generation has been growing exponentially. Nevertheless, the current methods still lack at least three essential characteristics."

Researchers now able to predict battery lifetimes with machine learning - "Technique could reduce costs of battery development. Imagine a psychic telling your parents, on the day you were born, how long you would live."

This deep learning technique solves one of the tough challenges of robotics - "This article is part of our coverage of the latest in AI research. For humans, working with deformable objects is not significantly more difficult than handling rigid objects. We learn naturally to shape them, fold them, and manipulate them in different ways and still recognize them."

Applications

Kendrick Lamar uses deepfakes in latest music video - "American rapper Kendrick Lamar has made use of deepfakes for his latest music video. Deepfakes use generative neural network architectures – such as autoencoders or generative adversarial networks (GANs) – to manipulate or generate visual and audio content."

Artificial intelligence tapped to fight Western wildfires - "With wildfires becoming bigger and more destructive as the West dries out and heats up, agencies and officials tasked with preventing and battling the blazes could soon have a new tool to add to their arsenal of prescribed burns, pick axes, chain saws and aircraft."

How machine learning is speeding up Formula 1 - "Machine learning has become key to teams’ successes on the track, but its use in the sport extend beyond that."

‘Machine learning’ improves rainfall mapping, water plans - "Rainfall map accuracy is vital in climate and hydraulic modeling and supports environmental management decision making, water resource planning and weather forecasting."

China to build AI-powered 3D printed hydroelectric dam in Tibet - "A new project on the Tibetan plateau is aiming to build a hydroelectric dam by means of artificial intelligence (AI) and 3D printing in just two years."

Channelling the power of AI to improve hearing - "We know that people experience hearing loss differently and often with varying challenges depending on the environment. Chatting with a spouse in a quiet room may be quite different than following a conversation amid the background chatter of a busy office environment, for example."

Google unveils two new tools designed to fight skin color bias - "The tech giant released a new color scale built to capture the range of human skin tones. 3 minute ReadGoogle announced a set of initiatives Wednesday aimed at creating a more equitable product experience for people across the skin-tone spectrum."

Google Has a Plan to Stop Its New AI From Being Dirty and Rude - "The future of computing is AI-powered chatbots that have read the entire internet—if Google can figure out how to tame them."

Google ups the AI ante with new features for Voice, Assistant, Maps and more - "Google today announced several AI-powered features to Voice, Lens, Assistant, Maps and Translate. "

How Do Patients Feel About AI in Health Care? It Depends - "New research is exploring what patients think about the use of AI in health care. Yale University's Sanjay Aneja, MD, and colleagues surveyed a nationally representative group of 926 patients about their comfort with the use of the technology, what concerns they have, and on their overall opinions about AI."

Business

Brightseed’s first AI-detected ‘phytonutrient’ comes to market alongside a $68M A round - "Who knows what secrets await in the hearts of plants? Brightseed has found a couple of them, anyway, using an AI-based analysis method, and will be bringing its first products to market soon with the help of a $68 million B round."

Kinetix uses no-code AI to create 3D animations, secures $11M - "Kinetix, a Paris-based AI-assisted 3D animation platform, today announced it has secured $11 million in a funding round."

SLAMcore closes Series A to develop spatial understanding for robots - "SLAMcore, a developer of vision-based location and mapping for robots, autonomous machines and consumer electronics products, raised $16 million in Series A funding."

Robots, automation, AI, and the Pirate Parrot: Insights from our return to the epicenter of robotics - "Butler said they see in robotics, AI and automation what they saw in Internet 2.0, with even more potential."

Slack’s former head of machine learning wants to put AI in reach of every company - "“The point of Graft is to make the AI of the 1% accessible to the 99%.” "

Salesforce acquires Troops.ai to make useful Slack bots - "Proving that there’s an appetite for bots, Salesforce today announced that it will acquire Troops.ai, a bot-based service that integrates with Slack to make it easier for sales teams to retrieve and update data across Salesforce’s various systems."

Intel debuts new chips for AI workloads, data center acceleration and laptops - "Intel Corp. today expanded its product portfolio with several new chips, including an artificial intelligence processor that promises to  provide twice the performance of Nvidia Corp.’s flagship A100 graphics card."

The AI 50 2022 - "Our fourth annual AI 50 list, produced in partnership with Sequoia Capital, recognizes standouts in privately-held North American companies making the most interesting and effective use of artificial intelligence technology."

$62.5M Fundraise Values Self-Checkout System Mashgin at $1.5B - "Mashgin, a touchless self-checkout system powered by artificial intelligence (AI) and computer vision, has raised $62.5 million in a Series B funding round that values the business at $1.5 billion, according to a Monday (May 9) press release. "

A machine learning company pitches its plan to get ahead of the next pandemic - "Online mental health company Cerebral was hit with a grand jury subpoena from the U.S."

Voxel raises $15M for AI-powered workplace safety - "Voxel, a San Francisco-based artificial intelligence (AI)-powered, workplace safety company, has secured $15 million in new funding. The company uses computer vision and AI to identify hazards, risky behaviors and operational inefficiencies across workplaces. "

Ditch The Chalk: Tessa Lau’s Dusty Robotics Raises $45 Million To Automate Construction Plans - "Dusty Robotics built an autonomous printer on wheels for construction sites. The company is now worth $250 million."

Concerns

Clearview AI banned from selling its facial recognition software to most US companies - "On Monday, Clearview AI agreed to settle a 2020 lawsuit from the ACLU that accused the company of running afoul of an Illinois law banning the use of individuals’ biometric data without consent."

Justice Department and EEOC Warn Against Disability Discrimination - "The Department of Justice and the Equal Employment Opportunity Commission (EEOC) today each released a technical assistance document about disability discrimination when employers use artificial intelligence (AI) and other software tools to make employment decisions."

AI’s Threats to Jobs and Human Happiness Are Very Real - "But short-term job chaos will give way to long-term prosperity, says AI expert Kai-Fu Lee"

Analysis

The Turkish Drone That Changed the Nature of Warfare - "The Bayraktar TB2 has brought precision air-strike capabilities to Ukraine and other countries. It’s also a diplomatic tool, enabling Turkey’s rise."

Fun

A place to improve AI technology, together. - "AI Test Kitchen is an app where people can experience and give feedback on some of Google’s latest AI technologies. Our goal is to learn, improve, and innovate responsibly on AI together. We’ll be opening up to small groups of people gradually. Stay tuned for more updates."

Humans vs. DALL·E - "Quick: which one of these was generated by a human, and which by an AI? What about now?"

---

Copyright © 2022 Skynet Today, All rights reserved.