A new framework allows AI models that have already been trained to learn new tasks without sacrificing performance when performing old tasks. The framework, called CHEEM, also improves an AI model’s operating efficiency by using fewer computational steps to perform simpler tasks.A new framework allows AI models that have already been trained to learn new tasks without sacrificing performance when performing old tasks. The framework, called CHEEM, also improves an AI model’s operating efficiency by using fewer computational steps to perform simpler tasks.[#item_full_content]

South Korean researchers have successfully developed a core technology that can fundamentally resolve “memory shortages,” a chronic bottleneck in large-scale artificial intelligence (AI) training. This technology is a next-generation memory expansion technology based on Ethernet, which is expected to drive infrastructural innovation across the entire AI and big data industries in the future.South Korean researchers have successfully developed a core technology that can fundamentally resolve “memory shortages,” a chronic bottleneck in large-scale artificial intelligence (AI) training. This technology is a next-generation memory expansion technology based on Ethernet, which is expected to drive infrastructural innovation across the entire AI and big data industries in the future.[#item_full_content]

Recent years have witnessed the unprecedented development of Industry 4.0 and the Industrial Internet of Things. These two technologies have significantly facilitated data collection from different sources for numerous tasks, such as reconstruction, classification, and prediction, for next-generation applications. However, the effective fusion and interpretation of these multi-source datasets remain challenging, making it a thriving area of research.Recent years have witnessed the unprecedented development of Industry 4.0 and the Industrial Internet of Things. These two technologies have significantly facilitated data collection from different sources for numerous tasks, such as reconstruction, classification, and prediction, for next-generation applications. However, the effective fusion and interpretation of these multi-source datasets remain challenging, making it a thriving area of research.[#item_full_content]

Computer scientists at UC Riverside have identified troubling flaws in a new generation of artificial intelligence (AI) agents designed to take over routine computer chores while users are away—sorting emails, organizing files, analyzing data, and handling other everyday digital tasks that might otherwise consume hours.Computer scientists at UC Riverside have identified troubling flaws in a new generation of artificial intelligence (AI) agents designed to take over routine computer chores while users are away—sorting emails, organizing files, analyzing data, and handling other everyday digital tasks that might otherwise consume hours.[#item_full_content]

New work explaining the inner workings of artificial intelligence could provide a way around the threat of AI “model collapse,” potentially averting growing numbers of AI hallucinations in the future.New work explaining the inner workings of artificial intelligence could provide a way around the threat of AI “model collapse,” potentially averting growing numbers of AI hallucinations in the future.[#item_full_content]

Researchers from The University of Osaka, Kyushu University, and the University of Victoria have developed a new method called Majority Voting SZZ (MV-SZZ) that accurately identifies defect-inducing software commits. By combining detailed code tracking with a majority voting system, the approach reduces false positives and outperforms existing techniques. This improvement could help developers debug software more efficiently and build more reliable systems. The work is published in the journal IEEE Transactions on Software Engineering.Researchers from The University of Osaka, Kyushu University, and the University of Victoria have developed a new method called Majority Voting SZZ (MV-SZZ) that accurately identifies defect-inducing software commits. By combining detailed code tracking with a majority voting system, the approach reduces false positives and outperforms existing techniques. This improvement could help developers debug software more efficiently and build more reliable systems. The work is published in the journal IEEE Transactions on Software Engineering.[#item_full_content]

Over the past few decades, computer scientists have developed increasingly advanced artificial intelligence (AI) systems that can tackle some tasks exceedingly well. These include computer vision models, systems that can rapidly analyze images and categorize them, recognize objects and faces, or make other accurate predictions.Over the past few decades, computer scientists have developed increasingly advanced artificial intelligence (AI) systems that can tackle some tasks exceedingly well. These include computer vision models, systems that can rapidly analyze images and categorize them, recognize objects and faces, or make other accurate predictions.[#item_full_content]

Researchers from MIT and elsewhere have developed a more user-friendly and efficient method to help networking engineers identify potential system failures before they cause major problems, like a cloud service outage that leaves millions of users unable to access applications.Researchers from MIT and elsewhere have developed a more user-friendly and efficient method to help networking engineers identify potential system failures before they cause major problems, like a cloud service outage that leaves millions of users unable to access applications.[#item_full_content]

Artificial intelligence systems based on neural networks—such as ChatGPT, Claude, DeepSeek or Gemini—are extraordinarily powerful, yet their internal workings remain largely a “black box.” To better understand how these systems produce their responses, a group of physicists at Harvard University has developed a simplified mathematical model of learning in neural networks that can be analyzed mathematically using the tools of statistical physics.Artificial intelligence systems based on neural networks—such as ChatGPT, Claude, DeepSeek or Gemini—are extraordinarily powerful, yet their internal workings remain largely a “black box.” To better understand how these systems produce their responses, a group of physicists at Harvard University has developed a simplified mathematical model of learning in neural networks that can be analyzed mathematically using the tools of statistical physics.[#item_full_content]

Today, artificial intelligence can describe images, recognize objects, and explain complex relationships. The pace of development is remarkable: So-called vision-language models (VLMs) combine text and image understanding in impressive ways. Yet, of all things, they struggle with a seemingly simple task—counting. Researchers at the Institute for Information Systems (iisys) at Hof University of Applied Sciences are now working to address this issue, with a paper posted to the arXiv preprint server.Today, artificial intelligence can describe images, recognize objects, and explain complex relationships. The pace of development is remarkable: So-called vision-language models (VLMs) combine text and image understanding in impressive ways. Yet, of all things, they struggle with a seemingly simple task—counting. Researchers at the Institute for Information Systems (iisys) at Hof University of Applied Sciences are now working to address this issue, with a paper posted to the arXiv preprint server.[#item_full_content]