Mini Briefs

How the U.S. patent office is keeping up with AI

The United States Patent and Trademark Office (USPTO) recently released a report on the growth of AI patents, the unique challenges of reviewing AI patents, and what the agency is exploring to address those challenges. From 2002-2018, patent applications in AI have doubled from 30k to 60k. The report makes the conclusion that currently, “there is no universal definition of AI,” and that AI only works in narrow applications with human assistance. For the sake of processing patents, USPTO defines “AI as comprising one or more of eight component technologies,” like computer vision, speech recognition, and AI hardware.

While many experts agree that current patent and intellectual property (IP) laws can “deal with the evolution of AI,” AI still poses significant problems for patents and IP. For example, AI-powered language generation tools “may create an overwhelming amount of prior art.” Interestingly, USPTO is “exploring using AI to improve the accuracy and integrity of the trademark register,” using machine learning models to analyze patent abstracts and scan for prior art. The report acknowledges that adapting the office for AI is an ongoing process that will require:

further engagement with the public, additional guidance for stakeholders, continued training for examiners on emerging technologies, and consultations with foreign IP offices.

Splice doubles users, teaches AI to sell Similar Sounds

Splice, an audio sample marketplace that lets artists browse and sample sounds, has seen a lot of success with using AI to index and search for similar audio. Previously to search for specific types of sounds, artists relied on manually annotated tags that came with individual tracks. However, these tags are subjective and are often not descriptive enough. Instead, Splice’s AI-powered marketplace finds similarities across tracks that were not tagged before, and it allows artists to look for similar audios without relying on human annotations. It’s “like with Google Reverse Image Search…now you can do that for any sound.”

Podcast

Check out our weekly podcast covering these stories! Website | RSS | iTunes | Spotify | YouTube

News

Advances & Business

• MIT researchers say their AI model can identify asymptomatic COVID-19 carriers - Researchers at MIT say they’ve developed an algorithm that can diagnose COVID-19 by the sound of someone’s cough, even if that person is asymptomatic.

• In the UK, researchers push on with efforts to develop autonomous vehicles that repair potholes - As well as being an eyesore and general nuisance, potholes are potentially dangerous: they can damage vehicles, cause accidents and affect everyone from those who drive motor vehicles to cyclists and pedestrians.

• Natural and artificial intelligence networks process 3D fragments of visual images in same way - The brain detects 3D shape fragments (bumps, hollows, shafts, spheres) in the beginning stages of object vision – a newly discovered strategy of natural intelligence that Johns Hopkins University researchers also found in artificial intelligence networks trained to recognize visual objects.

• Researchers suggest AI can learn common sense from animals - AI researchers developing reinforcement learning agents could learn a lot from animals. That’s according to recent analysis by Google’s DeepMind, Imperial College London, and University of Cambridge researchers assessing AI and non-human animals.

• Skan raises $14 million to automate repetitive business processes with computer vision - Skan.ai, an AI-enabled process discovery and operational intelligence platform, today closed $14 million in funding. The company says the proceeds will be put toward accelerating Skan’s go-to-market and product R&D efforts.

• Motional partners with Via to launch a robo-taxi service in the U.S. - Motional, a joint self-driving research collaboration between Aptiv and Hyundai, today announced a partnership with Via that the companies say will serve as a blueprint for an on-demand, shared robo-taxi service.

• Startup Robust.AI, Founded By Cognitive Scientist Gary Marcus And Roboticist Rodney Brooks, Raises $15 Million To Make Building Smarter Robots Easier - Gary Marcus, a former NYU professor and founder of another startup called Geometric Intelligence that was bought by Uber, teamed up with roboticist Rodney Brooks, cofounder of iRobot and co-inventor of the Roomba, to start a new company.

• Outrider raises $65 million to make warehouses and distribution centers autonomous - Autonomous logistics startup Outrider today closed a $65 million funding round led by Koch Disruptive Technologies, bringing the company’s total raised to $118 million.

• AI agreement to enhance environmental monitoring, weather prediction - NOAA’s Satellite and Information Service (NESDIS) has signed an agreement with Google to explore the benefits of Artificial Intelligence (AI) and Machine Learning (ML) for enhancing NOAA’s use of satellite and environmental data.

• Mayo Clinic, Google Focus on AI Effort for Cancer Care - The organizations will leverage AI to use de-identified data to develop and validate an algorithm to automate contouring of health tissues and organs from tumors.

Concerns & Hype

• There’s No Turning Back on AI in the Military - In the digital arms race with China, the only thing worse than fearing AI itself is the fear of not having it at all.

• Will Robotic Trucks Be “Sweatshops on Wheels”? - The future of self-driving trucks is being shaped by today’s policy debates on safety, minimum wage, and truckers’ - not trucks’ - autonomy.

• The creators of South Park have a new weekly deepfake satire show - It’s the first example of a recurring production that will rely on deepfakes as part of its core premise. The fake news: A new weekly satire show from the creators of South Park is using deepfakes, or AI-synthesized media, to poke fun at some of the most important topics of our time.

• How to make a chatbot that isn’t racist or sexist - Tools like GPT-3 are stunningly good, but they feed on the cesspits of the internet. How can we make them safe for the public to actually use?

• Medical chatbot using OpenAI’s GPT-3 told a fake patient to kill themselves - We’re used to medical chatbots giving dangerous advice, but one based on OpenAI’s GPT-3 took it much further. If you’ve been living under a rock, GPT-3 is essentially a very clever text generator that’s been making various headlines in recent months.

• Inequality grows in AI research - A new working paper finds that as high-level AI research began to require huge amounts of computing resources, large tech firms and elite universities increasingly dominated the field’s most important conferences.

• Waymo’s driverless cars were involved in 18 accidents over 20 months - Waymo’s driverless cars have driven 6.1 million autonomous miles in Phoenix, Arizona, including 65,000 miles without a human behind the wheel from 2019 and the first nine months of 2020.

• Uber drivers union asks EU court to overrule ‘robo-firing’ by algorithm - The action, which has been filed in the District Court of Amsterdam where Uber’s European HQ is located, was initiated by three U.K.-based former Uber drivers with the backing of the App Drivers & Couriers Union (ADCU), a U.K.

• AI Camera Ruins Soccer Game For Fans After Mistaking Referee’s Bald Head For Ball - The AI camera appeared to mistake the man’s bald head for the ball for a lot of the match, repeatedly swinging back to follow the linesman instead of the actual game.

Analysis & Policy

• To see what makes AI hard to use, ask it to write a pop song - Welcome home welcome home oh oh oh the world is beautiful the world. They’re not the most catchy lyrics. But after I’ve listened to “Beautiful the World” half a dozen times, the chorus is stuck in my head and my foot is tapping.

• A Look at Google’s Efforts to Earn Public Trust Through ML Fairness and Responsible AI - Trust in AI systems is becoming, if not already, the biggest barrier for enterprises - as they start to move from exploring AI or potentially piloting or doing some proof of concept works into deploying AI into a production system

• AI teachers must be effective and communicate well to be accepted - The increase in online education has allowed a new type of teacher to emerge – an artificial one. But just how accepting students are of an artificial instructor remains to be seen.

Expert Opinions & Discussion within the field

• Yann LeCun on GPT-3 - Some people have completely unrealistic expectations about what large-scale language models such as GPT-3 can do.

Explainers

• How to create an AI that chats like you on WhatsApp - The goal of this guide is to build a system capable of chatting like you, using your own WhatsApp and Telegram chats as an machine learning dataset. So let’s start our adventure.

• How I used GPT-3 to hit Hacker News front page 5 times in 3 weeks - In three weeks, I got to the front page five times, received 1054 upvotes, and had 37k people come to my site.

That’s all for this week! If you are not subscribed and liked this, feel free to subscribe below!