Training AI through human interactions instead of datasets December 3, 2024 Duke University Researchers have developed a platform to help AI learn to perform complex tasks more like humans. Called 'GUIDE,' it works by allowing humans to observe AI's actions in real-time and provide ongoing, nuanced feedback. Rather than relying on huge datasets, human trainers offer detailed guidance that fosters incremental improvements and deeper understanding. In its debut study, GUIDE helps AI learn how best to play hide-and-seek. Facebook Twitter Pinterest LinkedIN Email During your first driving class, the instructor probably sat next to you, offering immediate advice on every turn, stop and minor adjustment. If it was a parent, they might have even grabbed the wheel a few times and shouted "Brake!" Over time, those corrections and insights developed experience and intuition, turning you into an independent, capable driver. Although advancements in artificial intelligence (AI) have made self-driving cars a reality, the teaching methods used to train them remain a far cry from even the most nervous side-seat driver. Rather than nuance and real-time instruction, AI learns primarily through massive datasets and extensive simulations, regardless of the application. Now, researchers from Duke University and the Army Research Laboratory have developed a platform to help AI learn to perform complex tasks more like humans. Nicknamed GUIDE for short, the AI framework will be showcased at the upcoming Conference on Neural Information Processing Systems (NeurIPS 2024), taking place Dec. 9-15 in Vancouver, Canada. "It remains a challenge for AI to handle tasks that require fast decision making based on limited learning information," explained Boyuan Chen, professor of mechanical engineering and materials science, electrical and computer engineering, and computer science at Duke, where he also directs the Duke General Robotics Lab. "Existing training methods are often constrained by their reliance on extensive pre-existing datasets while also struggling with the limited adaptability of traditional feedback approaches," Chen said. "We aimed to bridge this gap by incorporating real-time continuous human feedback." GUIDE functions by allowing humans to observe AI's actions in real-time and provide ongoing, nuanced feedback. It's like how a skilled driving coach wouldn't just shout "left" or "right," but instead offer detailed guidance that fosters incremental improvements and deeper understanding. In its debut study, GUIDE helps AI learn how best to play hide-and-seek. The game involves two beetle-shaped players, one red and one green. While both are controlled by computers, only the red player is working to advance its AI controller. The game takes places on a square playing field with a C-shaped barrier in the center. Most of the playing field remains black and unknown until the red seeker enters new areas to reveal what they contain. As the red AI player chases the other, a human trainer provides feedback on its searching strategy. While previous attempts at this sort of training strategy have only allowed for three human inputs -- good, bad or neutral -- GUIDE has humans hover a mouse cursor over a gradient scale to provide real-time feedback. The experiment involved 50 adult participants with no prior training or specialized knowledge, which is by far the largest-scale study of its kind. The researchers found that just 10 minutes of human feedback led to a significant improvement in the AI's performance. GUIDE achieved up to a 30% increase in success rates compared to current state-of-the-art human-guided reinforcement learning methods. "This strong quantitative and qualitative evidence highlights the effectiveness of our approach," said Lingyu Zhang, the lead author and a first-year PhD student in Chen's lab. "It shows how GUIDE can boost adaptability, helping AI to independently navigate and respond to complex, dynamic environments." The researchers also demonstrated that human trainers are only really needed for a short period of time. As participants provided feedback, the team created a simulated human trainer AI based on their insights within particular scenarios at particular points in time. This allows the seeker AI to continually train long after a human has grown weary of helping it learn. Training an AI "coach" that isn't as good as the AI it's coaching may sound counterintuitive, but as Chen explains, it's actually a very human thing to do. "While it's very difficult for someone to master a certain task, it's not that hard for someone to judge whether or not they're getting better at it," Chen said. "Lots of coaches can guide players to championships without having been a champion themselves." Another fascinating direction for GUIDE lies in exploring the individual differences among human trainers. Cognitive tests given to all 50 participants revealed that certain abilities, such as spatial reasoning and rapid decision-making, significantly influenced how effectively a person could guide an AI. These results highlight intriguing possibilities such as enhancing these abilities through targeted training and discovering other factors that might contribute to successful AI guidance. These questions point to an exciting potential for developing more adaptive training frameworks that not only focus on teaching AI but also on augmenting human capabilities to form future human-AI teams. By addressing these questions, researchers hope to create a future where AI learns not only more effectively but also more intuitively, bridging the gap between human intuition and machine learning, and enabling AI to operate more autonomously in environments with limited information. "As AI technologies become more prevalent, it's crucial to design systems that are intuitive and accessible for everyday users," said Chen. "GUIDE paves the way for smarter, more responsive AI capable of functioning autonomously in dynamic and unpredictable environments." The team envisions future research that incorporates diverse communication signals using language, facial expressions, hand gestures and more to create a more comprehensive and intuitive framework for AI to learn from human interactions. Their work is part of the lab's mission toward building the next-level intelligent systems that team up with humans to tackle tasks that neither AI nor humans alone could solve. This work is supported in part by Army Research Laboratory (W911NF2320182, W911NF2220113). Story Source: Materials provided by Duke University . Original written by Ken Kingery. Note: Content may be edited for style and length. Cite This Page :
Kash Patel Is the Right Remedy to 'Fix' FBI, Says Retired FBI LeaderAmericans support increasing government efficiency
Canadian dollar recoups losses as investors assess BoC minutesSAD or Winter Blues?
UAE airlines keep link to IsraelIs Secret Santa stressing you out? Here’s your holiday gift-exchange survival guideMajor stock indexes we mixed on Wall Street in afternoon trading Monday, marking a choppy start to a holiday-shortened week. The S&P 500 rose 0.6%. The Dow Jones Industrial Average slipped 21 points, or 0.1% as of 2:22 p.m. Eastern time. The tech-heavy Nasdaq composite rose 1%. Gains in technology and communications stocks helped outweigh losses in consumer goods companies and elsewhere in the market. Semiconductor giant Nvidia, whose enormous valuation gives it an outsize influence on indexes, rose 3.6%. Broadcom jumped 5.7% to also help support the broader market. Walmart fell 2.2% and PepsiCo slid 1.3%. Japanese automakers Honda Motor and Nissan said they are talking about combining in a deal that might also include Mitsubishi Motors. U.S.-listed shares in Honda jumped 12.1% , while Nissan fell 0.9%. Eli Lilly rose 3.3% after announcing that regulators approved Zepbound as the first and only prescription medicine for adults with sleep apnea. Department store Nordstrom fell 1.7% after it agreed to be taken private by Nordstrom family members and a Mexican retail group in a $6.25 billion deal. The Conference Board said that consumer confidence slipped in December. Its consumer confidence index fell back to 104.7 from 112.8 in November. Wall Street was expecting a reading of 113.8. The unexpectedly weak consumer confidence update follows several generally strong economic reports last week. One report showed the overall economy grew at a 3.1% annualized rate during the summer, faster than earlier thought. The latest report on unemployment benefit applications showed that the job market remains solid. A report on Friday said a measure of inflation the Federal Reserve likes to use was slightly lower last month than economists expected. Worries about inflation edging higher again had been weighing on Wall Street and the Fed. The central bank just delivered its third cut to interest rates this year, but inflation has been hovering stubbornly above its target of 2%. It has signaled that it could deliver fewer cuts to interest rates next year than it earlier anticipated because of concerns over inflation. Expectations for more interest rate cuts have helped drive a 25% gain for the S&P 500 in 2024. That drive included 57 all-time highs this year. Inflation concerns have added to uncertainties heading into 2025, which include the labor market's path ahead and shifting economic policies under an incoming President Donald Trump. "Put simply, much of the strong market performance prior to last week was driven by expectations that a best-case scenario was the base case for 2025," said Brent Schutte, chief investment officer at Northwestern Mutual Wealth Management Company Treasury yields rose in the bond market. The yield on the 10-year Treasury rose to 4.59% from 4.53% late Friday. European markets were mostly lower, while markets in Asia gained ground. Wall Street has several other economic reports to look forward to this week. On Tuesday, the U.S. will release its November report for sales of newly constructed homes. A weekly update on unemployment benefits is expected on Thursday. Markets in the U.S. will close at 1 p.m. Eastern on Tuesday for Christmas Eve and will remain closed on Wednesday for Christmas.
He's been helping Donald Trump’s most contentious Cabinet picks try to win confirmation in the Senate.Nephew only wants to play video games on vacationlululemon athletica inc. Announces Third Quarter Fiscal 2024 Results
EnerVenue Provides RWE with Long-Duration Energy Storage VesselsTM for Pilot Project – The global renewables giant is evaluating the metal-hydrogen batteries at its U.S. testing facility in WisconsinOcado is ridiculed by customers over 'atrocious' Christmas substitute choices items after retailer suffers delivery malfunction
House Ethics Committee accuses Gaetz of paying for sex, including with 17-year-old girlAmericans support increasing government efficiency
Image Small Image Mark Kaufman Body It should come as no surprise that we’ve shifted from an “information economy” to an “attention economy.” Computer scientists and psychologists have been studying attention spans for about 20 years, over which time the average time that a person can focus on any one thing has dropped from around 21⁄2 minutes to around 45 seconds – even less if it’s a lecture you’re listening to. When was the last time you had any time at all free from distractions? Our brains are wired to pick up external cues like sights and sounds, all of which can distract us from the task at hand. Add technology to the mix – the time we spend in front of screens – and you begin to wonder how we ever get anything done. All kinds of things can contribute to a diminished attention span: sleep deprivation, stress, a noisy or cluttered environment, and screen time. Making changes to any one of these factors would help; so would full immersion in a book. There’s little doubt that reading is an activity that you have to consciously make time for, when there are so many other things competing for our time and attention. The challenge, of course, is to find just the right book to capture your attention and imagination, and keep you engrossed. Online, you’ll find algorithms designed to make suggestions based on past purchases. Instead, spend time browsing a bookstore and talk with a bookseller (usually a voracious reader), who will recommend something based on your interests and tastes. You’ll find more than 50 categories from which to choose, from science-fiction to romance, nature to current affairs, personal growth to faith. As the holidays approach, books are the perfect gift that can be opened again and again. You’ll never have to worry about viruses, software updates or changing the batteries. Even if you’re shopping for that impossible-to-please relative, you’ll find something among the 10,000 or so books a bookstore carries, or hundreds of thousands more that can be ordered and available within a couple of days. Author Sandi Richmond (“Milepost 75”) celebrated her 80th birthday by hiking rim-to-rim at the Grand Canyon. Physical activity, sensible eating, and time management all contribute to living longer, healthier lives. You can be sure that Richmond has kept her brain just as healthy. Trading screen time for a good book is a great place to start. Mark Kaufman is the co-owner of Story & Song Bookstore Bistro. mark@storyandsongbookstore.com
School Superintendent Marcy Kelley told a federal judge that the symbol “XX” displayed on pink armbands worn by Bow parents should be banned from all sporting events going forward, whether or not a transgender athlete is playing. Kelley repeatedly defended the district’s decision to bar a group of parents and family members from wearing the armbands during a September soccer game in protest of a transgender girl’s participation in the game. She said the symbol “XX,” which refers to the chromosomes associated with females, is “anti-trans.” “I didn’t want to wait for her to see this and feel like it’s wrong for her to be trans,” Kelley testified, referring to the sophomore player on the visiting Plymouth Regional High School girls soccer team. Federal judge Steven McAuliffe will now weigh whether to temporarily block that ban on either First Amendment or viewpoint discrimination grounds, a decision which he described as legally “close” and which will hinge on previous court decisions in First Amendment cases related to school speech. McAuliffe’s assessment of the case followed two days of testimony from the plaintiffs – Bow parents Kyle Fellers, Anthony Foote, and Nicole Foote, and grandparent Eldon Rash – and from Bow administrators, including Kelley and high school principal Matt Fisk. The judge’s decision on the plaintiffs’ request for a preliminary injunction barring the school district from enforcing its protest restrictions likely won’t come until at least December, according to the schedule set in the case. Kelley spent more than an hour on the witness stand on Friday morning explaining why school leaders and ultimately the Bow police directed three of the plaintiffs in the case to remove the pink armbands during the soccer game. Two of the parents – Fellers and Foote – were later barred from attending future games, though those bans have since expired. Kelley also explained why she would take the same action if the armbands were ever displayed at a school event again, described her personal views on transgender girls’ participation on girls’ sports teams, and detailed her administrative team’s planning in the days preceding the Sept. 17 soccer game. Her testimony followed Fellers’, Foote’s and Rash’s accounts of what happened during the soccer game from their perspectives. Article continues after... Cross|Word Flipart Typeshift SpellTower Really Bad Chess Kelley vehemently defended her district’s approach to the incident, pointing to the school district’s responsibility to protect students from what her legal team has described as “harassment and intimidation” aimed at a particular student. “In school districts, when we suspect that there’s some sort of threat that something may happen, we don’t wait for it to happen to take action,” Kelley said. “And that’s what we did here.” Kelley described the “XX” marking, which was displayed prominently on the armbands, as “a pretty well-known anti-trans symbol” that she views as “exclusionary.” She said her personal views on the participation of transgender girls in girls sports were “nuanced and there are many factors to consider,” and said she disagreed with a blanket ban like the one enacted in New Hampshire this summer. But Kelley said her primary issue with the armbands was that they targeted a specific player. “If we were to allow harassment against a particular student, we would be liable,” Kelley said. The school district has contended that displaying the XX chromosomes violated a policy that requires “mutual respect, civility, and orderly conduct” during school events. The Bow administrators said the policy would be violated whether or not transgender athletes are present, because of the message of exclusion they send to other transgender students at Bow High. “We were previously worried about [the Plymouth player]. We are now more generally worried about the transgender community at this school and what this message says to them,” said Jonathan Shirley, an attorney for the Bow School district. In the afternoon, Fisk, the principal, testified about his experience as an adviser for his school’s gay-straight alliance, including the challenges he has witnessed transgender students navigate. “Schools should be places of education where students are able to learn in an environment in which they can take risks,” Fisk said. “And I don’t think you can do that if you see signs ... that you are not wanted, that you should be excluded.” Fisk said even the presence of the “XX” symbol in the school’s parking lot would prompt a conversation, though he didn’t outright say that “XX” bumper stickers are banned. When probing for the presence of viewpoint discrimination, Kelley acknowledged that a gay pride flag would be acceptable at the same sporting events that the “XX” symbol is banned. She said that is because the former is “inclusionary,” while the latter is “exclusionary.” Kelley also described in depth how the school district prepared for the September soccer protest in the days after they caught wind of it via social media. She said the district considered but ultimately rejected the option of closing the game to spectators completely. Ultimately, Kelley worked with Fisk, athletic director Michael Desilets, and the Bow police department to increase the presence of officials at the game. During the second half, the game was temporarily paused while Foote, Fellers, and Rash were asked to remove their armbands, which they ultimately agreed to do after a police officer got involved. Jeremy Margolis can be contacted at jmargolis@cmonitor.com .