Bridging the Language Gap in Global Crises
A new assessment reveals uneven progress in machine translation capabilities, leaving vulnerable communities at risk during emergencies.
Science
Forecasting the Future with Confidence: A New Foundation Model for Time Series
Researchers have developed a novel approach to time series forecasting that not only predicts future values but also rigorously quantifies the uncertainty surrounding those predictions.
Taming AI Economies: A Control System for Stable Tokenomics
![The system establishes a closed-loop feedback mechanism-the Dynamic-Control Buyback Mechanism-where deviations between a target price and real-time market values are processed by a PID controller to determine intervention intensity, subsequently constrained by solvency parameters and enacted through market buy-and-burn operations, ensuring asymptotic solvency even amidst volatile conditions and effectively stabilizing the decentralized AI economy via iterative price adjustments [latex] e_{k} [/latex].](https://arxiv.org/html/2601.09961v1/figures/overall.png)
New research demonstrates how applying control theory can dynamically stabilize the often-volatile token economies powering decentralized artificial intelligence networks.
AI Learns to Predict Thailand’s Monsoon
Researchers have harnessed artificial intelligence to identify a key climate indicator for more accurate long-term rainfall forecasting in Thailand.
Beyond Resilience: How Cities Can Thrive in the Face of Crisis
New research proposes a framework for urban antifragility, demonstrating that cities can not only withstand shocks but actively grow stronger through them.
The Weakest Links: Exposing Flaws in AI Agent Capabilities
A new large-scale study reveals that a significant portion of skills used by AI agents harbor security vulnerabilities that could lead to data breaches and unauthorized access.
Beyond Agent Alignment: Governing AI’s Collective Behavior
A new framework proposes shifting AI safety from internal constraints to robust, external governance structures designed for complex multi-agent systems.
How AI Might Save Us: Mapping Paths to Survival
A new analysis categorizes potential future scenarios where humanity successfully navigates the risks of increasingly powerful artificial intelligence.
Beyond the Lie: Taming falsehoods in AI Language Models
As large language models become increasingly integrated into critical applications, ensuring factual accuracy and mitigating the risk of fabricated information is paramount.
Beyond the Lab: AI Predicts Concrete Strength with Unprecedented Accuracy
A new study demonstrates the potential of artificial intelligence to rapidly and reliably assess concrete compressive strength, offering a path toward automated quality control in large-scale construction projects.