Articles tagged with Artificial Intelligence
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.
Junior, Stanford University's new autonomous robot, joins the DARPA Urban Challenge with enhanced sensors and sophisticated AI software. Its brain is four times more powerful than its predecessor Stanley, collecting and processing vast amounts of data from sensors such as lasers, radar, and video cameras.
Researchers created digital humans that mimicked participants' head movements and voices to test their persuasive effects. The study found that people were more emotionally engaged when interacting with these digital humans compared to traditional computer interfaces.
The Intelligent Information Systems Institute aims to explore AI's applications in data management, game theory, and more. Cornell faculty members are collaborating with visiting scientists to inspire innovation.
Researchers at USC developed an animated health intervention called Carmen's Bright IDEAS to improve social problem-solving skills of mothers caring for kids with cancer. The system uses robotic thespians and AI to engage users in a positive experience, resulting in uniformly positive reviews from clinical trial participants.
The AI-powered Author system can generate readable prose, including dialogue and short stories, by analyzing a story plan and applying rules to create grammatically correct sentences. However, it still lacks the ability to discern fact from fiction, which is a significant hurdle before it can be used for hot news reporting.
The new A.I argumentation system, which uses defeasible reasoning, allows for a more natural and human-like approach to legal decision-making. This enables lawyers to build stronger arguments and make informed decisions based on competing perspectives.
A team of researchers at Sandia National Laboratories has developed an intelligent software agent that can defend itself and other computers against cyber attacks. The agent uses a multi-agent collective approach to detect and respond to unusual requests or commands, providing real-time protection against various types of threats.