Articles tagged with Artificial Intelligence
A new Yale University study found that even 'dumb' AI can help human groups by boosting their performance. The experiment, which used an online game with human players and robotic AI players, showed a significant improvement in the median time for groups to solve problems.
Georgia Tech researchers have developed a team of free-flying robots that can maintain formation and avoid collisions. The robots use virtual top hats to maneuver in the air. Meanwhile, autonomous blimps with face-detecting capabilities are being built to recognize hand gestures and interact with humans.
A new study reveals AI systems retain biases seen in humans, with European American names associated more with pleasant stimuli. The research uses statistical analysis of word associations to demonstrate these findings, which have implications for understanding cultural stereotypes.
Researchers from MIT's Computer Science and Artificial Intelligence Laboratory present a system called Flexplane that tests new traffic management protocols at realistic speeds without altering network hardware. The system uses a compact, efficient computational model of a network to schedule transmissions on the real network.
A £5m project by the University of Manchester aims to reduce offshore wind farm maintenance costs using AI, robotics, and advanced sensors. The initiative will enable predictive diagnostics, allowing for early issue detection and minimal human intervention in hazardous environments.
Researchers created an AI system called DeepStack that outperforms human players in heads-up no-limit poker by leveraging asymmetrical information about the game state. The system solves approximately 10^7 decision points in 5 seconds, enabling potential advances in real-world problems involving information asymmetry.
DeepStack, a computer program developed by the University of Alberta's Computer Poker Research Group, has defeated professional poker players in heads-up no-limit Texas hold'em poker. The program uses a technique called continual re-solving to determine the correct strategy for each situation, allowing it to play at human speed.
Researchers analyzed Wikipedia bots' interactions on 13 language editions over ten years, finding they behave differently in culturally distinct environments. Bots often engage in 'sterile fights' lasting years or reaching deadlock, with conflicts triggered later than human reverts.
UC researchers successfully employed fuzzy logic in a simulation to show it is an ideal system for navigating under dynamic conditions. The concept, also known as genetic-fuzzy, allows drones to make good navigational decisions amid statistical noise and has real-world applications such as autonomous delivery.
A new report reveals over 5,000 likely transshipments between 2012-2016, with potential for millions more, highlighting the need for improved regulation and monitoring of transshipment in the global seafood market. The study also identifies clusters of transshipment along EEZs of countries with limited monitoring capabilities.
A new wearable AI system developed by MIT researchers can predict the tone of conversations with high accuracy, analyzing audio and vital-sign data. The system has shown promise in understanding social interactions, particularly for people with anxiety or Asperger's syndromes.
A recent study on Reddit found that a small group of voters can significantly impact the content users see, with up-voted posts receiving a higher final score and being more likely to reach the front page. The study highlights the importance of understanding how social rating systems curate media in our daily lives.
A new system was presented at the Association for Computing Machinery's Symposium on Principles of Programming Languages that automatically handles database caching in web applications written in Ur/Web. This system offers twofold to 30-fold speedups in experiments involving two websites.
A Northwestern University team developed a new computational model that performs at human levels on the Raven's Progressive Matrices, a nonverbal standardized test. The model's ability to solve complex visual problems and understand analogies provides evidence for the importance of symbolic representations and analogy in visual reasoning.
Researchers at FAU's Computer Chemistry Center are developing an AI application to predict odor molecules, neutralizing undesirable smells in packaging and products. The project aims to improve product development efficiency, reducing trial-and-error experiments.
Researchers developed a method to read and amplify high confidence states using decoded neurofeedback, boosting participants' confidence without awareness. The effect was reversible, with confidence decreasing when the pattern was no longer detected.
The Internet of Things (IoT) will require new search mechanisms to support billions of connected devices, demanding machine-to-machine searches and secure data access. Complex IoT technologies like smart cities, autonomous cars, and environmental monitoring need efficient indexing and ranking mechanisms.
Researchers have developed simulated search agents that significantly outperformed human participants in computer search challenges, offering a promising solution for automating literature review and suggesting relevant sources. The agents can be configured to mimic human behavior, providing a credible simulation of a human researcher.
Researchers have trained a robot to imitate natural human actions, allowing for effective coordination during high-stakes events like surgeries. This development holds promise for improving safety and efficiency in medical procedures.
A new display technology, Cinema 3D, allows viewers to watch 3-D movies without glasses, using a special array of lenses and mirrors. The system enables high-resolution images to be replicated across the theater, offering an immersive cinematic experience from any seat.
UMass Amherst researchers develop a new framework for managing control transfer between humans and autonomous systems, enabling safe semi-autonomous vehicles. The approach leverages human strengths while capitalizing on machine capabilities.
A new anonymity scheme called Riffle provides strong security guarantees while using bandwidth efficiently. The system employs a mixnet with onion encryption and uses authentication encryption to verify the authenticity of messages, ensuring cryptographically secure communication as long as one server remains uncompromised.
Researchers from MIT's CSAIL present a new chip design called Swarm that makes parallel programs run many times faster and requires one-tenth the code. The chip has extra circuitry to prioritize tasks and handle synchronization between cores, making it easier for programmers to adapt sequential algorithms.
Researchers at MIT developed a new model to reduce unnecessary communication in human-robot teams, enabling them to work together more effectively. The system has been tested on computer simulations and shows promise for improving team performance.
The special issue of Current Opinion in Neurobiology delves into various aspects of brain plasticity, from language development at birth to cognitive flexibility associated with bilingualism. The journal features expert reviews that connect different fields of research and explore the role of artificial intelligence.
A study published in Frontiers in Psychology found that people are quicker to find a hard-to-see person in an image when the directions mention a prominent landmark first, as opposed to last. The researchers also discovered that listeners prefer to use these landmarks as 'mental bookmarks' to speed up the process.
The Leverhulme Centre for the Future of Intelligence aims to explore the opportunities and challenges of artificial intelligence, examining technical, practical, and philosophical questions. Researchers from various disciplines will work together to understand the potential benefits and risks of machine intelligence, with a focus on en...
Georgia Tech researchers trained IBM Watson AI to 'chat' with students, sparking creativity and providing answers to complex questions on biologically inspired design. The teams used Watson to learn from biology articles and ask questions on topics like desalination and solar cells, receiving rapid responses with relevant answers.
Researchers developed a system to analyze and improve running motions using big data and artificial intelligence, identifying key differences between high-rank marathon runners and beginners. The technology, called 'skill grouping,' converts movements into objective scores, enabling the development of healthcare tools and assisting tra...
Scientists at Tufts University developed an AI approach to illuminate cellular processes and suggest possible targets for aberrant development. The method was applied to tadpoles, where melanoma-like cells deviated from normal development, indicating a group dynamic rather than single-cell decisions.
GeoS uses a combination of computer vision and natural language processing to solve geometry questions. It achieved an SAT score of 500 out of 800, equivalent to the average test score for 2015.
Researchers in Navarre, Spain developed a system that uses soft computing to predict the survival of multiple trauma patients. The system achieved high accuracy by considering patient severity variables, age and other clinical parameters.
Researchers at UC Berkeley developed algorithms that enable robots to learn motor tasks through trial and error, mimicking human learning. The 'deep learning' technique allows robots to adapt to changing environments without pre-programmed details.
Researchers at Berkeley and UMass developed a new algorithm that enables autonomous systems to deal with uncertainty, allowing robots like PR2 to perform tasks without specific knowledge of what needs to be done. The framework combines several planning paradigms and optimizes them for efficient execution on modern hardware.
The system analyzes video data to extract clues such as heart rate and eye pupil radius, while also analyzing Twitter posts and keystroke rates. The program can accurately detect emotions, with a high correlation between its predictions and users' self-reported feelings.
Researchers developed robotic systems that can teach themselves by watching online cooking videos, allowing them to grasp intricate movements. The breakthrough enables self-learning and interaction with humans, paving the way for smart manufacturing environments and automated daily life tasks.
Researchers at Queen Mary University of London have created magic tricks using artificial intelligence, leveraging mathematical techniques to create new variants. The AI-generated tricks, such as a jigsaw puzzle and mind reading card trick, were well-received by audiences and are now available for download.
New research from Heriot-Watt University and the University of Southampton shows collective switching or group buying schemes can save UK consumers up to a third of their yearly electricity bills. The study proposes a model to help consumers form more efficient buying groups and minimize switching risks.
Researchers created a fully automated computer program called LEVAN that learns all possible variations of a concept by searching millions of books and images on the Web. The program displays results as a comprehensive, browsable list of images, helping users explore and understand topics quickly.
Researchers at MIT create a self-contained autonomous soft robot capable of rapid body motion, mimicking the escape maneuver of real fish. The robotic fish uses fluid flow through flexible channels to change direction quickly and explore new advantages in soft robotics.
A team of researchers created an autonomous robotic construction crew that builds complex structures without a central command, using collective intelligence and stigmergy. The TERMES system consists of simple robots that cooperate to modify their environment, achieving impressive results in building towers, castles, and pyramids.
Researchers suggest that evolvability can increase over generations without competitive pressure, leading to increased diversity. The study's findings challenge traditional theories on the origins of evolvability.
Researchers analyzed 2.5 million articles from 498 English-language outlets, finding that tabloid newspapers are more readable than broadsheets and use more sentimental language. The study also revealed a dominance of male voices in news content, with 'Sport' and 'Financial' articles being the most male-biased.
The University of Granada Research Group on Decision and Optimization Models won the Simulated Car Racing Competition with their Soft Computing-based modular architecture. This technology enables efficient control of cars in the presence of other competitors, allowing for a stable performance despite not being the fastest car.
The Autonomous Aerial Cargo Utility System (AACUS) aims to deliver supplies quickly and safely in hostile conditions using autonomous robotic helicopters. With a smartphone-like device, warfighters can control the aircraft's takeoff, flight, and landing.
A team of academics has developed an AI soccer manager that consistently ranks among the top 1% of players in the English Fantasy Football League. The software analyzes player statistics and performance to pick its team each week, outperforming human competitors.
Researchers at MIT CSAIL discovered brain organizes objects based on size, with distinct regions for big and small objects. This finding has major implications for fields like robotics and could lead to better understanding of brain organization and mapping information.
A new study by Sidney D'Mello and Art Graesser found that strategically inducing confusion in learning sessions on difficult conceptual topics can lead to better retention and application of knowledge. Subjects who were confused scored higher on a difficult post-test and could identify flaws in new case studies.
Bigham's proposal, Closed-Loop Crowd Support of People with Disabilities, aims to leverage the crowd to provide real-time feedback and support to individuals with visual impairments or other disabilities. The NSF CAREER Award will support his research over the next five years.
The study proposes an 'anytime' intelligence test that can be applied to any subject, system, or level of intelligence, providing a universal evaluation framework for human and artificial intelligence. This breakthrough has significant implications for cognitive sciences and the development of future intelligent systems.
Researchers at Rensselaer Polytechnic Institute have developed a system that analyzes and responds to the complex musical structures of improvised avant-garde performances, using interconnecting elements of cognition such as perception/action, reasoning, decision-making, planning, and memory.
A large-scale content-analysis of cross-linguistic text reveals national biases, cultural, economic and geographic links between countries influence the news agenda. Outlets from similar countries tend to cover similar stories, such as those related to trade or Eurovision.
North Carolina State University has received a four-year, $3.5 million grant to explore using artificial intelligence as a learning tool in fourth- and fifth-grade science classrooms. The project includes creating CyberPads with AI-based software that supports interactive scientific modeling for topics such as electricity and landforms.
The NRL's Navy Center for Applied Research in Artificial Intelligence has won two top awards at the IJCAI conference for its videos on transfer learning and gesture recognition. The winning videos showcase the lab's cutting-edge research in AI and robotics, with applications to naval missions.
Researchers developed an AI-powered system using RFID sensors and mobile devices to remind users of everyday tasks. The system recognizes user actions, assessing activities such as taking keys or leaving home, without modifying daily routines or requiring cameras or microphones.
Manuela Veloso, a professor at Carnegie Mellon University, has won the Autonomous Agents Research Award for her influential research in artificial intelligence and robotics. Her work on robot soccer has emerged as an important tool for studying autonomous agent cooperation in complex environments.
Children can influence and observe an artificially intelligent machine in a simple computer game called 'Gorge', learning about its behavioral patterns. Researchers aim to teach kids what AI is all about using this engaging game.
The Deep Green program aims to create a system that can help future combat commanders anticipate enemy moves and adjust plans. The system uses anticipatory planning with adaptive execution, generating options for all sides in an operation and predicting the likelihood of multiple futures.
Researchers developed a biomimetic jumping robot inspired by grasshoppers, capable of jumping 1.4 meters and navigating rough terrain with ease. The tiny robot can be fitted with sensors and solar cells to aid in search and rescue operations or explore remote areas on Earth or other planets.
Researchers solved checkers by creating a computer program that cannot be beaten, using 500 billion positions and superior processing capabilities. The program, Chinook, may play to a draw but will never lose.