Securing the IoT with SAILOR
A new RISC-V core family balances performance, energy efficiency, and robust security features for resource-constrained devices.
Science
Resolving Knowledge Conflicts in Visual Question Answering
A new framework dynamically prioritizes relevant information to improve accuracy when answering questions that require external knowledge.
Quantum Networks: Keeping Entanglement Alive in a Dynamic Sky
![The average age of entanglement decreases as the elementary link generation probability increases, exhibiting this relationship under conditions of mostly visible links-defined by visibility transition matrices of [latex]0.3\ 0.7; 0.3\ 0.7[/latex]-and a switching probability of 0.8.](https://arxiv.org/html/2602.23985v1/2602.23985v1/x1.png)
A new approach optimizes continuous entanglement distribution across satellite networks by intelligently managing intermittent link availability and decoherence.
Sharper Vision, Leaner Models: Adapting Quantization with Expert Networks
A new framework intelligently manages errors during model compression, enabling more efficient large-scale vision-language AI.
Shielding Quantum Code: A New Approach to Circuit Obfuscation
Researchers have developed a method combining logic and phase encryption to protect intellectual property within quantum circuits during the compilation process.
Safeguarding AI: A Dynamic Shield Against Emerging Threats
Researchers have developed a novel system that allows AI safety policies to be updated on the fly, without requiring costly and time-consuming model retraining.
Hidden Order, Stronger Magnetism: Designing Impurity Patterns for Quantum Control
New research reveals that carefully arranging magnetic impurities within quantum materials can dramatically improve and stabilize their magnetic properties.
Bridging the Language Gap in Semantic Understanding
New research demonstrates a method for automatically creating training data for semantic parsers in multiple languages, significantly reducing the need for costly manual annotation.
Unlocking Cancer’s Genetic Code: A Faster Path to Multi-Hit Combinations
![A binary matrix representing seven genes across five samples demonstrates the selection of two gene combinations - [latex] c_1c_1 [/latex] and [latex] c_2c_2 [/latex] - which, together, accurately identify two tumor samples ([latex] t_1t_1 [/latex] and [latex] t_2t_2 [/latex]) while also incorrectly flagging one normal sample ([latex] n_1n_1 [/latex]), highlighting a characteristic pattern of both sensitivity and limited specificity in this gene-based analysis.](https://arxiv.org/html/2602.22551v1/2602.22551v1/x1.png)
New optimization techniques are accelerating the identification of gene combinations that drive cancer development, bringing actionable insights within reach of standard computing resources.
Untangling Frustration: A New Phase Transition in Quantum Chains
![The spin-1 [latex]J_1-J_2-J_3[/latex] chain exhibits a complex phase diagram characterized by transitions between disordered, [latex]\mathbb{Z}_4[/latex] ordered, Haldane, and dimerized phases, with the nature of these transitions-ranging from Gaussian and Ashkin-Teller to first-order and Ising-shifting based on the relative strength of [latex]J_1[/latex], ultimately converging towards a multi-critical point at [latex]J_1 = 0[/latex] representing two copies of WZW SU(2)₂ critical theory and delineating a leg-dimerized phase from uncoupled Haldane chains.](https://arxiv.org/html/2602.23187v1/2602.23187v1/x2.png)
Researchers have uncovered a unique quantum phase transition in coupled frustrated chains, revealing unexpected critical behavior and expanding our understanding of complex quantum systems.