New research reveals that advanced reasoning systems aren’t just processing language, they’re dynamically reshaping their internal representations to grasp the underlying structure of problems.
New research demonstrates how vision and language models can empower humanoid robots to interpret complex tasks and execute them with grounded, spatial awareness.
A new review synthesizes the rapidly evolving landscape of AI agents designed to assist medical professionals and improve patient care.
New research explores how perceptions of authorship are shifting as generative AI tools become increasingly integrated into programming education.
New research introduces a comprehensive benchmark for evaluating AI assistants that learn from user behavior over time to provide truly helpful, proactive support.
![The study demonstrates that while idealized multi-agent flocking-measured by alignment [latex]\gamma\gamma[/latex] and inter-agent distances-maintains coherence, the introduction of even modest delays and noise predictably degrades performance, manifesting as deviations in centroid path length [latex]SS[/latex] and reduced overall flock stability.](https://arxiv.org/html/2602.04012v1/x9.png)
Researchers have developed a novel flocking algorithm that enhances group coordination by predicting the future movements of individuals within the swarm.
This review explores the emerging field of data agents-AI-powered systems designed to navigate and manage the complexities of the modern data landscape.
Researchers have developed a unified approach to modeling and controlling the complex interactions between robots and objects, paving the way for more robust and adaptable manipulation capabilities.
![Across the SIR, CR, and gases datasets, the average magnitude of forecasting error-measured as Mean Absolute Error [latex] MAE [/latex]-varied predictably with increasing noise levels, indicating a consistent sensitivity to data quality regardless of the specific time series being analyzed.](https://arxiv.org/html/2602.04114v1/test_Avg_MAE_per_noise_gases.png)
A new framework blends the strengths of mechanistic modeling and machine learning to achieve more accurate and interpretable predictions for complex systems.
![The system integrates streaming egocentric vision and audio via a real-time multimodal language model to not only generate spoken dialogue but also to proactively issue low-latency function calls-such as directing gaze to specific people, objects, or areas-that dynamically update perceptual context and drive active perception through external tools like [latex]Look\_at\_Person[/latex], [latex]Look\_at\_Object[/latex], and [latex]Use\_Vision[/latex].](https://arxiv.org/html/2602.04157v1/figs/main_fig.png)
Researchers are developing new systems that enable robots to engage in more natural and grounded conversations by combining real-time sensory input with advanced language processing.