Chipmakers and operating system developers have spent years building defenses. A new study from MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) shows that a key assumption behind many of them doesn't hold.
In recent years, computer scientists have developed a wide range of artificial intelligence (AI) models that can rapidly recognize patterns in data, generate content and solve other computational problems. Many of these AI systems are based on deep neural networks (DNNs), brain-inspired computational models that can make predictions based on specific data, or LLMs, models that can process human language, answer queries and generate text.
European energy operators are braced for next week's solar eclipse, which is expected to cause a temporary dip in power generation, and have taken steps to keep electricity flowing to businesses and consumers, an industry body said Friday.
Keele scientists have developed a new kind of electrolyzer that produces both hydrogen and the base materials for sustainable plastic production, opening avenues for a range of applications in green technology and sustainable development.
A legal battle between Apple and OpenAI over the ChatGPT maker's secretive device project escalated this week when the artificial intelligence (AI) developer called the iPhone maker's allegations "baseless" and asked for the lawsuit to be dismissed.
Amazon confirmed Friday that it is financing a massive private gas power plant in Texas that could become the single-largest source of greenhouse gas emissions in the United States.
The EU's tech chief called on Meta and TikTok to "act decisively" to stem disinformation at times of crisis, saying the European Commission held talks with both platforms Friday following last week's Ceuta migrant surge.
A team from the U.S. Department of Energy's (DOE) Argonne National Laboratory has successfully demonstrated an artificial intelligence (AI)-driven system to automate a powerful simulation method that predicts how atoms in materials interact. Known as atomistic simulations, this method can potentially accelerate the discovery of materials for areas such as batteries, aerospace and electronics. The research is published in the journal Digital Discovery.
URL scanning services are now a common component of modern security workflows. They help detect phishing websites or malware early and warn users before they visit a URL. However, the practice of some of these services—publishing scanned URLs—can inadvertently expose sensitive user data. CISPA researcher Ali Mustafa, together with colleagues from the Max Planck Institute for Security and Privacy and Ca' Foscari University of Venice, conducted the first systematic investigation of the risks arising from this practice.
Over the past two decades, a powerful but highly specialized branch of control engineering—known as negative imaginary (NI) systems theory—has quietly evolved into a key tool for stabilizing complex, vibration-prone systems, from flexible structures to advanced robotics.
AI semiconductors are becoming more programmable. KAIST researchers have developed a device whose response characteristics can be programmed to process data that changes at different speeds. The technology reduced prediction errors for time-varying data by up to 40-fold and is expected to enhance real-time AI performance in autonomous vehicles, robots and wearable devices.
On paper, a megawatt-hour is just a megawatt-hour; it will charge your phone or run your dishwasher the same way, regardless of the source. However, its value to, or impact on, the environment is not fixed. It depends almost entirely on what power plant it replaces. A new wind turbine only helps the climate and local air quality to the extent that it pushes a more polluting plant offline. If instead it is just competing with another clean energy source, the environmental benefit effectively disappears.
Meta said Thursday that one of its artificial intelligence models accessed the internet on its own and hacked another company, the latest in a series of disclosures about AI models going rogue.
As large language models (LLMs) become increasingly embedded in chatbots, virtual assistants, translation services, coding tools and other AI-powered applications, delivering responses quickly and efficiently has become a growing challenge. Because these models generate text one token at a time, inference can be slow and computationally expensive, particularly for larger models. While speculative decoding has emerged as a promising approach to accelerate inference, many existing methods either require additional model training or struggle to perform consistently across different hardware platforms.
The protection of personal data is now a key part of individual literacy in everyday life. Nevertheless, researchers regularly observe a contradiction among the general public, which they describe as the "privacy paradox." There is often a wide gap between the concerns expressed about protecting personal data and people's actual behavior when handling such data.
Researchers have developed a process for converting carbon dioxide and industrial waste into green hydrogen, reporting in a new study that turning trash into valuable mineral products could usher in a more efficient way to produce clean fuel.
The Yixing Zisha teapot sits neatly at the intersection of art, engineering and heritage. As such, it makes for an intriguing test case for a technology increasingly taking on creative tasks across almost every field of human endeavor: artificial intelligence, or AI. Research published in the International Journal of Systems, Control and Communications describes a computational system that can help AI generate new teapot designs without losing the cultural characteristics that make the craft distinctive.
Doyeon Kim, an undergraduate researcher in the Department of Electronic and Electrical Engineering, has published a paper in an academic journal. The research focuses on implementing drone artificial intelligence (AI) identification technology through low-power semiconductors. This study could dramatically improve the battery efficiency of systems that monitor unauthorized drones in real time in military and industrial environments. The research is published in the journal IEEE Transactions on Industrial Informatics.
Login
