Research on computing hardware and its future. Covers quantum error-correction, cyber defense, novel materials, and AI frameworks.
Researchers at Graz University of Technology discovered vulnerabilities in file notification systems of various operating systems, allowing tracking of user activities, keystrokes, and password prompts. Patches have been rolled out to address some of the vulnerabilities, but users are still advised to be cautious.
This technology introduces a method to enhance neural network security by applying weight perturbations during training, preventing unauthorized access and attacks. It strengthens model integrity across all layers, ensuring holistic protection against various attacks.
Researchers at Harvard and Georgia Tech have developed RLE-Bench, a benchmark that tests AI coding agents' ability to engineer physical robots. The benchmark contains 48 tasks that test the agent's ability to perform engineering work required to build and operate robotic systems, including control and perception algorithms, designing r...
Five UTDallas faculty members received CAREER awards for research on resilient networks, next-generation electronics, software security and mathematical understanding of complex systems. Researchers aim to improve network resilience, develop new semiconductor technologies and advance mathematical understanding of complex systems.
Team Kinetiz from NTU Singapore, K2, and Alp Autonomy won the world's first multi-car autonomous race in Europe, taking the lead at Imola Circuit. The team demonstrated exceptional autonomous driving skills, showcasing the potential of AI in motorsport.
The new MSc Computer Science with Artificial Intelligence programme equips learners with advanced technical knowledge and practical skills in AI. The programme combines Heriot-Watt's academic expertise with flexible online delivery, enabling learners worldwide to study alongside work and other commitments.
A new learning mechanism uses natural variability in neural activity to understand how synapses adapt and improve the learning capabilities of brain-inspired devices. The mechanism, called Spike-based Alignment Learning, solves the weight transport problem and matches the performance of existing approaches without unrealistic assumptions.
The system uses a 'reward/expectation' signal derived from the motor cortex to provide timely and precise control of prosthetic devices. It can adjust its policy based on evaluation signals, allowing for more natural control and improved outcomes.
This technology leverages predicted occupant counts to optimize HVAC control, accounting for misprediction costs to minimize energy consumption. By updating thermal states and misprediction distributions, the system achieves weighted occupant thermal comfort while minimizing total expected misprediction costs.
A new partnership will provide access to a constantly evolving training environment, enabling the development of practical cyber skills and strengthening connections between education, research, industry, and defence. The partnership aims to build Australia's sovereign cyber capability and ensure resilience in the face of evolving cybe...
A quantum LDPC code with a large minimum distance near 48 protects 4,612 logical qubits using 9,216 physical qubits, reducing hardware overhead. The code's design framework, inspired by classical LDPC codes, allows for strong evidence of a minimum distance and threshold phenomenon.
Researchers at the University of Warwick have created a new material combining magnetism and electrical polarisation, making it possible to switch magnetic information using an electric field. The material works at close to room temperature, a significant breakthrough for energy-efficient computer memory.
A study by Aalto University reveals pro-Russian hackers are using 'gamified' DDoS attacks to undermine public trust in governments and institutions, with cryptocurrency-based rewards for successful attacks. The researchers urge companies and governments to counteract the effects of propaganda and psychological intimidation.
A new AI framework, AdvWT, uses realistic fading, cracks, and corrosion to test vulnerabilities in AI vision systems. It achieves near-perfect attack success rates on lightweight CNNs and transformer-based models, and can also be used to restore naturally damaged traffic signs.
Researchers develop adaptive multi-expert framework for dynamic 3D reconstruction, combining strengths of multiple motion representations to improve reconstruction quality. The framework leverages complementary experts to handle heterogeneous dynamics, enabling more accurate reconstruction of complex scenes.
Research from the University of Birmingham and other institutions found that interacting with AI-powered customer service robots can reinforce or alter a consumer's self-perception. The study explores how mirroring and mimicry can lead to a 'robotoid humanness' where consumers become more like robots, raising ethical concerns.
Professor Fioretto's team develops AI-driven methods for autonomous power grid topology control, improving resilience and efficiency. The Genesis Mission Platform provides access to advanced AI models and high-performance computing resources.
A study of 3786 children found no significant associations between screen use and developmental difficulties, except for very high use of electronic games. Early high screen use was associated with greater difficulties at 10 years old, but not sustained at 15 years.
USC researchers have been selected for the U.S. Department of Energy's Genesis Mission to harness artificial intelligence for scientific discovery and innovation. Two projects led by USC will explore ways to develop faster and more energy-efficient computing hardware and better understand the natural concentration of critical minerals.
A Tulane University team is using AI to discover new superconductors, which could improve the nation's electrical grid, medical imaging, and quantum computing. The project combines high-fidelity calculations, physics-aware AI, and experimental measurements to accelerate discovery.
Peter J. Denning suggests that Turing's stance on artificial general intelligence and the imitation game has led to the AI mess, with a fundamental flaw in understanding tacit knowledge and its representation problem. He argues that machines cannot grasp human emotions, intuitions, or cultural context, making it impossible to achieve h...
A large-scale study of an online patient portal shows that AI-generated responses can introduce errors and extraneous details, leading to increased editing time for physicians. Adapting AI to individual physician communication styles can improve accuracy by 33% and reduce editing by 26%.
A new study by Rice University researcher Conner Joyce examines the evolution of cybercriminal communities into politically motivated networks. The research found that online interactions can create strong social bonds that facilitate the spread of political and extremist beliefs.
Researchers at Princeton University have developed a method to manufacture transistors with less risk of damage using a class of extremely thin materials called transition metal dichalcogenides. By pretreating the surface with oxygen or fluorine, it becomes easier to remove atoms from the top layer without damaging the lower layers.
Researchers developed a new magnetic memory material that can be rewritten using laser light, allowing for faster and more energy-efficient storage and processing of information. This breakthrough could help reduce power consumption in data centers and support future high-speed information systems.
A Japan-US collaborative team has developed the world's first integrated spintronic probabilistic bit on a silicon chip, paving the way for large-scale spintronic p-computers. The innovation addresses computational problems requiring parallel processing of enormous numbers of possible states.
Biocomputing harnesses human brain cells to perform computational functions, with applications in energy-efficient computing, drug discovery, and neuromorphic engineering. Early developments in biocomputing hold promise for biomedical research, despite ongoing bioethical concerns.
Researchers at Binghamton University developed a system that enables users to monitor real plants in real-time using virtual reality. This technology makes farming more accessible for older adults and people with disabilities, allowing them to observe plants without physically being present.
The University of Texas at Arlington has joined a statewide program to help address the shortage of K-12 computer science teachers in Texas. The WeTeach_CS Secondary Certification Preparation Hub provides training for teachers to implement computer science education, with UTA faculty providing mentoring and support.
ACM recognizes educators Yasmin Kafai and Mitchel Resnick for their contributions to creative computing and Scratch. Ed Felten receives the ACM Policy Award for his work on technology policy, while Jodi Tims is honored with the Outstanding Contribution to ACM Award for increasing women's participation in computing
A new AI system, Empirical Research Assistance (ERA), can automatically write scientific software programs that outperform human-written ones. ERA combines a large language model with search strategies to explore and refine thousands of pieces of code, reducing the time required for exploration from months to hours or days.
Monika Henzinger, an Austrian researcher, has made significant contributions to dynamic graph algorithms and web algorithms. She is recognized for her outstanding work in processing large datasets and mentoring the next generation of researchers.
University of Missouri researchers develop organic transistors that process information like biological neural networks, boosting brain-like computing and potentially leading to more energy-efficient artificial intelligence. The approach could lead to significant improvements in tasks such as pattern recognition and decision-making.
Researchers from MIT have developed a more user-friendly and efficient method to identify potential system failures in cloud computing algorithms. The 'MetaEase' technique analyzes an algorithm's source code directly to uncover hidden blind spots that might cause unexpected failures, reducing the risk of costly network outages.
Researchers at MIT developed a technique to overcome memory constraints and communication bottlenecks in federated learning, enabling faster and more accurate AI model training. The new framework, FTTE, uses a subset of model parameters and an asynchronous approach to reduce lag time and improve training performance.
A recent study from Lancaster University reveals that AI systems like ChatGPT can learn to mirror human impoliteness, potentially escalating into verbal violence. The research tested ChatGPT's ability to respond to real-life impolite interactions, finding it often produces more impolite behavior than humans.
Researchers developed a new AI-assisted tool called CacheMind to improve cache performance and reduce evictions. The tool uses causal reasoning to analyze fine-grained details about system behavior, enabling computer architects to identify patterns and implement fixes.
A team from Emory University developed a simple method to test the accuracy of protein language models, which are used to analyze complex biological data. By comparing how these models 'embed' natural proteins versus synthetic ones, researchers can estimate their reliability and improve their performance.
A seven-month study from Aalto University found that frequent micro checks and bursts of messaging are most strongly linked to feeling overwhelmed. The study's findings suggest that task-switching and constant checking contribute to digital overwhelm, rather than the total time spent on devices.
The ARU Arm AI Lab will provide researchers and students with access to advanced Arm AI technology, focusing on real-world applications in healthcare and life sciences. This partnership will also support emerging talent and drive innovation, building on existing collaborations and industry projects.
Researchers at the University at Albany and Rutgers University have developed an early-warning framework that can predict harmful social media interactions before they erupt. The model, called Comment Storm Severity (CSS), uses publicly available datasets from Reddit and Instagram to predict whether a thread will escalate into a toxic ...
A systematic review and meta-analysis found a small correlation between daily screen use and later sleep onset in youth. However, short-term fluctuations in screen time had limited effects on sleep duration, efficiency, or quality.
Study by Duke engineers reveals that a common device architecture used to test 2D transistors overstates their performance prospects up to sixfold. The back-gated architecture amplifies the transistor's performance using contact gating, but has physical limitations that prevent its use in commercial technologies.
Researchers at the University of East London identified a critical weakness in Industrial Internet of Things networks that can destabilize factories and safety-critical infrastructure. Attackers can exploit subtle 'delay attacks' to quietly disrupt coordination, making timing just as important as information in highly automated systems.
Researchers at UCSF and Wayne State University found that generative AI tools can perform orders of magnitude faster than human teams in analyzing health data. Junior researchers paired with AI generated viable prediction models in minutes, outperforming experienced programmers in hours or days.
Researchers will gather to discuss how to engineer AI systems that work in the real world, focusing on composition, optimization, verification, and evaluation. The conference aims to establish shared foundations for a new class of software, including methods for evaluating models and ensuring durability, efficiency, and dependability.
Researchers at Technical University of Denmark developed a groundbreaking nanolaser that can halve a computer's energy consumption. This technology has the potential to revolutionize various industries, including information technology and healthcare, by enabling ultra-small and energy-efficient lasers.
Researchers say tech companies need to adopt user-centered approaches to create authentication systems that are secure by design. They call for more inclusive and consistent methods, avoiding one-size-fits-all models that burden users with complex security rules.
Researchers have developed a passive, solar-powered orbital data center that can scale AI computing and reduce environmental impact. The system leverages decades of research on 'tethers' and could host thousands of computing nodes to replicate terrestrial data centers.
A UMass Amherst-led study suggests old-school analog computing can improve energy efficiency and computing speed in the IoT. Researchers have developed a brain-inspired sensing system that combines touch sensors and smart memory chips to reduce data transmission speeds.
The six finalists apply advanced analytics to tackle critical challenges in sustainability, supply chain optimization and technology. Their solutions have generated significant value and improved performance across industries.
Researchers at Politecnico di Milano developed a 'smart' chip that dramatically reduces energy consumption while accelerating data processing, achieving similar accuracy to digital systems with lower power consumption and faster performance.
A review of over 1,000 studies found that technology communication reduces engagement and positive emotional responses compared to in-person interactions. In-person communication has been shown to have health benefits, including reduced anxiety and improved educational outcomes.
The use of generative AI in software development can lead to increased productivity, but also creates technical debt and hasty coding if not managed properly. Companies must prioritize fixing technical debt, develop clear guidelines for AI-assisted coding, and train developers on the hazards of AI coding.
A new study from the University of East Anglia finds that AI tools excel at speed and clarity, but human translators remain crucial for tone, cultural nuance, and natural sound. Generative AI can handle parts of video translation, but consumers still notice when something feels off.
Researchers created a metric to quantify lattice flexibility and studied how it impacts proton transport. They ranked the importance of seven features, including hydrogen bond length and oxygen sublattice flexibility, finding that these are critical for efficient proton conduction.
A new study found that highly educated, Western European millennials have the most digital concerns, with individuals having higher levels of digital literacy being the most affected. The research also found that people's perceptions of digital harm are heightened by their level of digital exposure and knowledge.
A11yShape, an AI-assisted tool, enables visually impaired computer programmers to create, edit and verify complex 3D models without assistance. The tool uses GPT-4o to provide detailed descriptions of the model for blind and low-vision programmers.
A new AI system can accurately reconstruct hand muscle activity without sensors, enabling precise estimation of fine motor control. The technology has potential applications in sports science, rehabilitation, and human-machine interaction.
Researchers investigate how people with disabilities feel when their avatars depict them the way they see themselves. Most participants reported positive feedback when their avatar reflected their disability, and many felt comfortable expressing their identity in virtual reality.