This survey explores the application of hypergame theory to address challenges in multi-agent systems where participants operate with misaligned perceptions and incomplete information.
The authors propose a risk-budgeted verification framework to improve the safety and efficiency of real-time contingency screening in energy management systems.
Researchers conducted a large-scale empirical analysis of actor-critic algorithm design components to improve reliability and performance in real-world reinforcement learning applications.
A study of German companies explores how generative AI and predictive analytics are being integrated into human resource management to automate routine tasks.
Researchers propose a method for adapting generalist vehicle foundation models to specific platforms to improve high-speed autonomous control across diverse terrains.
The study demonstrates that model pruning can significantly reduce the computational requirements of diffusion-based text-to-audio generative models without sacrificing performance.
This paper discusses the challenges of computational reformulation for modern hardware when incumbent software outputs become the de facto industry specification.
Researchers developed a compact convolutional neural network that achieves state-of-the-art performance in handwritten Devanagari character recognition while significantly reducing parameter count.
A new framework utilizes AI to automate and accelerate various stages of professional upskilling and knowledge acquisition.
A new active learning strategy for audio classification improves efficiency by refining how segments are selected for annotation to reduce the costs associated with frame-level labeling.
A new multi-agent framework utilizes tool-augmented evidence to improve the profiling of urban regions by integrating diverse data sources.
A review of the Lenovo ThinkStation P3 Ultra SFF Gen 2 highlights its balance of compact size and expandability as a mini-PC workstation.
Researchers propose DeepLoop, a method for scaling sequential computation in Transformers by reusing physical blocks to increase effective depth without adding parameters.
Researchers introduced Attention Head Reweighting to improve the data efficiency of large language models during adaptation tasks.
Apple is taking legal action against OpenAI, claiming the artificial intelligence company's leadership directed the theft of proprietary trade secrets.
Unisplendour's subsidiary, Unis Computer, is raising 440 million RMB from new investors, which will reduce Unisplendour's ownership stake to 29.42%.
Shell Finance launched a content creator program to foster a new digital ecosystem.
A new podcast episode from The Register discusses how high token costs are causing corporate executives to reconsider their investments in artificial intelligence.
Several Chinese technology and sensor firms have jointly established a new sensor technology company in Wuhan.
A new state-backed equity investment partnership has been established in Hefei, China, with a capital contribution of 5 billion RMB to focus on industrial investments.
EdgeMode is divesting several data center development sites in Spain to Pure DC and Spark AI Foundry.
Competition in the AI agent smartphone market is intensifying as Nubia and StepFun introduce competing hardware platforms integrated with autonomous agent capabilities.
The Imaging-101 benchmark evaluates the ability of large language model coding agents to reconstruct scientific computational imaging pipelines across multiple domains.
Aviation technology firm Hanfa Changkong raised nearly 100 million yuan to develop 400kW-class engines for low-altitude economy applications.