Science – Page 90

Designing with Confidence: AI Meets Physics for Robust Metasurfaces

27.01.2026 by ebaster

A robust design optimization workflow leverages uncertainty quantification-specifically, physics-based rules-to efficiently balance exploration and exploitation, mitigating the risk of erroneous designs that can arise when single-fidelity surrogates deviate from the training distribution during metasurface design.

A new optimization framework leverages physics-based uncertainty to dramatically improve the reliability and efficiency of AI-driven metasurface design.

Cellular Chatter: Building Intelligent Networks with Molecular Signals

27.01.2026 by ebaster

The system diverges from conventional point-to-point Monte Carlo methods by embracing a bio-inspired network of excitable agents, suggesting an architectural shift towards resilience and adaptability in complex systems rather than simple linear progression.

Researchers are exploring how to harness the power of biological communication to create robust and adaptable networks for the Internet of Bio-Nano Things.

AI Chemists: Mastering Molecular Design with Reasoning and Tools

27.01.2026 by ebaster

ChemCRAFT demonstrates superior performance across diverse chemical reasoning benchmarks, achieving this through a training process-spanning supervised fine-tuning and reinforcement learning-that optimizes both task completion and efficient resource utilization, evidenced by a notably lower inference cost and more practical token length compared to existing chemical multi-agent systems.

A new approach combines artificial intelligence with strategic tool use to significantly advance the design and synthesis of complex molecules.

Seeing is Assisting: Eye Control for Smarter Robotic Arms

27.01.2026 by ebaster

Robust information transmission between an eye-tracking system and a robotic field of view is achieved through a defined program flow, enabling seamless coordination between visual focus and robotic action.

A new framework leverages eye-tracking to improve the accuracy and reliability of assistive robotic arms in everyday tasks.

Who Wrote That Code? Tracking AI’s Footprint on GitHub

27.01.2026 by ebaster

Agent-specific feature importance was determined through one-vs-rest XGBoost models, enabling the identification of key characteristics that differentiate individual agents within the system.

A new study reveals that AI coding assistants leave unique, detectable patterns in their contributions, even when submitted under human identities.

Robots That Think on Their Feet: Adapting to the Unexpected

27.01.2026 by ebaster

The framework integrates symbolic planning, reinforcement learning, and skill abstraction to achieve adaptive learning, effectively bridging the gap between high-level reasoning and low-level action execution through a neuro-symbolic approach.

A new approach combines the strengths of neural networks and symbolic reasoning to allow robots to overcome obstacles and learn in real-world, unpredictable environments.

The Ghostwriter in the Machine: When AI Rivals Human Prose

27.01.2026 by ebaster

New research reveals that AI, when trained on exceptional writing, can produce text indistinguishable from, and even favored over, human-authored work.

The Ghost in the Machine: Navigating Social Life with AI

27.01.2026 by ebaster

Social.AI enables users to refine follower demographics, publish content, and then receive automated replies from artificial intelligence agents - a streamlined interaction designed to personalize online engagement.

A new study explores how users interact with and perceive a social platform populated entirely by artificial intelligence agents, revealing surprising insights into the future of online connection.

Beyond Black Boxes: Understanding Autonomous AI Systems

27.01.2026 by ebaster

The progression from artificial intelligence agents to truly agentic systems unfolds across distinct eras, mirroring a fundamental tension: as autonomy increases, so too does the challenge of maintaining interpretability - a prophecy inherent in every architectural choice.

As AI agents become increasingly complex, current interpretability methods fall short of providing the accountability needed for safe and reliable deployment.

Robots That Work Together: The Rise of Multi-Agent Systems

27.01.2026 by ebaster

A multi-agent planning system dissects user instructions and visual scenes to orchestrate robotic action, achieved through a collaborative architecture-comprising activation, planning, and monitoring agents-each refined via supervised fine-tuning on datasets [latex]L_1[/latex] and [latex]L_2[/latex] derived from the VIKI benchmark.

A new benchmark challenge is pushing the boundaries of collaborative robotics, demanding increasingly sophisticated coordination and adaptability from teams of diverse machines.