Weekly roundup — The top AI stories you might have missed this week: - Researchers tested OpenAI's CLIP neural network for robots and found that it demonstrated racist and sexist stereotypes through its behaviors. Lead study author Andrew Hundt says humanity is at risk of creating "a generation of racist and sexist robots."
- Tesla laid off about 200 people, many of them Autopilot data labelers, at its offices in San Mateo, California. The annotation specialists were tasked with evaluating and labeling customer data to train Tesla's self-driving systems.
- California autonomous vehicle companies asked Gov. Gavin Newsom to allow self-driving truck operations in the state. California allows smaller AV operations, such as robotaxis, but has yet to pass regulations to allow the testing or deployment of autonomous trucks weighing more than 10,001 pounds.
- Google AI researchers introduced Minerva, a language model that can solve math and science questions using step-by-step reasoning. While GPT-3, BERT, and other NLP models have reached a high level of sophistication in generating text, they've fallen short of human-level performance in solving quantitative reasoning problems, according to the researchers.
- University of Chicago researchers said they developed an AI model that could predict the rate and location of certain crimes in the city weeks in advance at up to 90% accuracy. The model was trained on historical crime data.
| |
Data analytics group Govini says the U.S. government is not investing enough in AI, machine learning, and other emerging tech to succeed in future battlefield and other conflicts. While federal spending on these critical technologies rose from $60.7B in FY2017 to $117.2B in FY2021, the U.S. is still not moving fast enough to keep pace with China's advancements in these areas, according to a new Govini report. More: - Of the amount spent in FY21, $12B was invested in AI/ML and autonomous technologies, with the biggest investment increases going toward data acquisition and architecture storage.
- However, the U.S. still isn't pumping enough money into AI/ML to continue innovating in the area of battlefield technologies, Govini warned.
- It cited the Russia-Ukraine war, noting its use of autonomous drones, AI intelligence fusion, cyber battles, and computing devices designed to augment soldiers' awareness are the future of warfare.
- Notably, the Pentagon has requested its highest R&D budget in history for FY23, with many investments planned for AI.
Zoom-out: - Govini's analysis comes over a year after the U.S. National Security Commission on Artificial Intelligence released a final report stressing that the U.S. is unprepared to compete in the new age of artificial intelligence and risks falling behind China.
- The commission said that "big decisions" are needed to speed up AI innovation in favor of the U.S. and protect against malignant AI uses, such as cyber-attacks and China-backed disinformation campaigns.
BREAKING DEFENSE 
| |
 A message from ASHCROFT CAPITAL With the housing market the way it is these days, it seems impossible to invest in real estate right now – except for apartments. Why invest in multifamily apartments? Because most investors are looking for passive income or growth opportunities to add to their portfolio but multifamily private placements provide both in one investment. Multifamily allows investors a unique opportunity to collect monthly cash flow (passive income) and participate in the equity (growth) as the property appreciates in value. Here lies the advantages of investing in multifamily private placements: - you don’t have to manage a property
- you don’t have to worry about repairs or renovations
- you don’t even have to find your own deals
However, before you invest in any private placement, you still need to take the time to understand the terms. You can learn how to make informed decisions when investing in real estate by downloading this free Private Offering Guide. Download Now For accredited investors only. The guide is for informational purposes only and is not intended for any other use. Not an offer to sell securities.
| |
Social news site Reddit said it acquired MeaningCloud, a natural language processing firm, for an undisclosed sum. Launched in 2015, MeaningCloud's technology is focused on extracting meaning from web conversations and other unstructured content. More: - Reddit said the deal will allow it to strengthen its AI and machine learning capabilities and better understand unstructured data to surface relevant information for its users more quickly.
- The announcement aligns with Reddit's recent purchase of Spell, a platform to run resource-intensive ML experiments without the needed hardware. That deal is also expected to help Reddit surface more relevant posts for users and improve its spam detection and algorithm matching,
- Based in Spain, MeaningCloud's team is now supporting machine learning projects in Reddit's product, safety, and advertising teams.
- Reddit says it uses machine learning for its Discover Tab, introduced this year, along with personalized recommendations and targeted advertising.
TECHCRUNCH
| |
FIFA says it will use AI and semi-automated technologies to detect offside positions during the 2022 Men's World Cup. The tech will include sensors embedded in balls as well as limb-tracking technology for players, according to the soccer governing body. More: - The limb tracking tech relies on 12 cameras placed around a stadium to track 29 points on each player's body.
- It plans to implant IMU sensors in World Cup balls to gather inertial data and process it using AI software developed at MIT and elsewhere. The results will be communicated to on-field referees.
- In addition, a semi-automated video review system will be used to confirm or overturn calls for plays involving goals or penalties.
- While still reviewed by human referees, the system is meant to cut the length of such reviews from 70 seconds to 20.
- FIFA said it has tested the technology successfully at multiple tournaments.
THE GUARDIAN
| |
Chinese automaker FAW Group has invested an undisclosed amount in Horizon Robotics, which develops AI chips for robots and autonomous vehicles. Beijing-based Horizon said it will use the strategic funds to boost its engineering capabilities and for the R&D of its automotive-grade AI chips. More: - In addition to AV chips, Horizon produces software installed in vehicles, smart speakers, and other technologies.
- It has provided AI chips for more than 70 vehicle models at 20 car companies.
- Since 2020, it has had a strategic partnership with FAW Nanjing and FAW's Intelligent Connected Vehicle Development Institute.
- Last year, the company released its latest "Journey 5" AI chip, geared for Level 4 high autonomous driving. The chip will be installed in a new FAW car model.
- Horizon is also backed by Intel Capital, SK Hynix, Hillhouse Capital, SAIC, Yunfeng Capital, and others.
JUST AUTO
| |
IBM and Promare's crewless Mayflower vessel docked in Plymouth Harbor, Massachusetts, ending a 3,500-mile journey across the Atlantic after several setbacks. The autonomous ship used an "AI Captain," cameras, and sensors to navigate, collect ocean data, scan for hazards, and predict storms based on meteorological data. More: - The crewless vessel departed from Plymouth in the U.K. in April and was diverted to Portugal's Azores islands due to technical difficulties before reaching Nova Scotia, Canada, last month.
- The vessel was moved and docked next to a replica of the original Mayflower, which made the trip in 1620 with 101 people on board, in the town of Plymouth, Mass.
- The fully crewless vessel was built by IBM and Promare, a nonprofit marine research organization. The two co-engineered its software, which relied on real-time machine learning and analytics, sensor input, and a "decision engine" to navigate.
- Its AI "brain" was based on IBM's earlier work in AI and machine learning, including its chess master computer Deep Blue and "Jeopardy!" champion AI system Watson.
- In May, an autonomous cargo ship completed a near 500-mile voyage off Japan's East coast. The Suzaka relied on a navigation and vessel tracking system made by Orca AI, an Israeli computer vision startup.
BOSTON.COM
| |
QUICK HITS - Assets are getting larger, and end-users have you stuck between the cloud and a hard place. Deliver far beyond user expectations with Cox Edge's CDN.*
- Intel, the Aster Innovation and Research Center, and CARPL.ai are rolling out an AI-enabled health data platform in India.
- Compredict, a startup providing AI-enabled software solutions for predictive maintenance and vehicle health, raised $5.8M.
- Beverage alcohol giant Diageo acquired Vivanda, a startup using AI to match a person's taste and texture preferences to food and drinks.
- This AI-powered UI test automation tool gives developers an easy way to author self-healing cross-browser tests.*
*This is sponsored content.
| |
UPCOMING EVENTS - July 20 - The Rise of Metaverse Gaming (Register Here)
- August 17 - Build Your Diversified Portfolio: How To Invest in Bear Markets (Register Here)
- September 14 - Are Electric Cars Worth It? (Register Here)
| |
 |  | Beth is a writer and analyst at Inside.com covering artificial intelligence and daily tech news. Since starting work at Inside, she has covered various topics including virtual reality, drones, and e-commerce. Prior to that she was a public policy and investigative reporter for The Arizona Republic, where she won a Pulitzer Prize nomination and First Amendment Award for reporting on the rising costs of pensions. Reach her at Beth.Duckett@inside.com. |   | | Editor | Aaron Crutchfield is based in the high desert of California. Over the last two decades, he has spent time writing and editing at various local newspapers and defense contractors in California. When he's not working, he can often be found looking at the latest memes with his kids or working on his 1962 Ford. | |
| |
Comments
Post a Comment