Precise Function Approximation with Lattice Algorithms
New research demonstrates how median lattice algorithms achieve near-optimal convergence rates for approximating periodic functions in high dimensions.
Science
Hiding in Plain Sound: A Watermark That Survives AI Audio Editing
![A system subtly shifts audio’s latent representation via an optimized perturbation δ before quantization, inducing a constrained movement-defined by the vector [latex]v = \mu\_B - \mu\_A[/latex] between cluster centroids-designed to survive the destructive cycles of neural codec pipelines and be reliably detected by a verification process [latex]\mathcal{E}[/latex].](https://arxiv.org/html/2603.05310v1/2603.05310v1/x3.png)
Researchers have developed a new audio watermarking technique that embeds information within the core structure of compressed audio, making it remarkably resilient to manipulation by advanced neural audio codecs.
Beyond the Hype: Sharpening AI Insights with Reliable Data
A new approach leverages multiple AI agents and advanced optimization techniques to carefully select the most trustworthy data for accurate sentiment analysis.
Spinning Out Insights: A New Approach to Particle Decay Analysis
Researchers have developed a novel method for constructing Lorentz-covariant amplitudes, offering a streamlined path to understanding particle interactions and decay processes.
Balancing Power and Precision in Stochastic Computing
![For CMOS chains, reliability against bit-flip errors increases exponentially with bias voltage but sub-exponentially with chain length, creating a performance envelope where a fixed dissipation budget is best utilized by minimizing chain length and maximizing bias voltage - a relationship governed by the circuit’s characteristic timescale [latex]\tau\_{0}=\frac{e}{I\_{0}}e^{V\_{\rm th}/(nV\_{T})}[/latex].](https://arxiv.org/html/2603.04658v1/2603.04658v1/x3.png)
New research reveals the fundamental tradeoffs between energy dissipation and bit stability in CMOS-based stochastic computing systems.
Bridging the Bangla QA Gap: A New Dataset for Educational Question Answering
Researchers have released a large-scale dataset designed to improve question answering systems for Bangla, addressing the critical need for resources in low-resource languages.
Twisted Lattices Unlock Exotic Superfluidity
![A triangular lattice configured with alternating single-particle hoppings-[latex]J_1[/latex] and [latex]J_2[/latex]-creates effective geometric frustration and a π-flux on each triangular plaquette, which is mathematically represented by a linearized one-dimensional chain Hamiltonian containing nearest-neighbor and next-nearest-neighbor couplings.](https://arxiv.org/html/2603.04498v1/2603.04498v1/x3.png)
Researchers have engineered a novel optical lattice system demonstrating chiral and pair superfluid phases through carefully controlled geometric frustration and correlated atom hopping.
Echoes of the Early Universe: Simulating Inflationary Ripples
A new lattice simulation study explores the spatial distribution of density fluctuations generated during hybrid inflation and their possible link to cosmic defects.
Matching Complexity: A Treewidth Limit
New research establishes a firm lower bound on the time complexity of finding optimal matchings in graphs with limited tree-like structure.
Splitting the Difference: How Lie Algebra Structure Shapes Commutative Subalgebras
This review explores the conditions under which specific decompositions of Lie algebras yield elegantly structured, polynomial-like subalgebras.