Topology Under Pressure: How Disorder Impacts Robust Edge States
![The study elucidates a model-the SSH model-governed by alternating hopping terms [latex]t_1[/latex] and [latex]t_2[/latex], alongside random corrections [latex]\tau_n[/latex] and on-site energies [latex]v_n[/latex], as described by tight-binding equations for local amplitudes [latex]\psi_n[/latex].](https://arxiv.org/html/2512.24738v1/SSH2.png)
A new study reveals the surprising resilience of topological edge states in the Su-Schrieffer-Heeger model, even when faced with significant disorder.
Science
Revealing Intersections: A New Era for Private Set Intersection
Researchers have developed more efficient and secure protocols for multi-party private set intersection, enabling identification of contributing parties while preserving data privacy.
Untangling Integrability: A Kagome Lattice Approach
New research explores the mathematical underpinnings of a solvable model on a Kagome lattice, revealing connections to quantum field theory and statistical mechanics.
The Repetition Curse: Overloading AI with Clever Prompts
New research reveals how carefully crafted, repetitive requests can cripple large language models by exploiting imbalances in their internal routing mechanisms.
Decoding the Twisted Path to Reliable Data
New algorithms significantly improve the efficiency of decoding complex error-correcting codes used in data storage and transmission.
Building Better Codes: A New Algebraic Approach
Researchers have developed a novel method for constructing hierarchical quasi-cyclic codes, leveraging the strengths of Reed-Solomon and polynomial evaluation techniques.
Beyond Accuracy: A Smarter Approach to Clinical Coding
![A hybrid system architecture manages the complexities of clinical code assignment by employing a tiered approach-semantic reasoning with a language model [latex]BioMistral-7B[/latex] initially attempts structured output, gracefully degrading to deterministic keyword matching when parsing fails, and finally subjecting all candidate codes to rigorous validation against a knowledge base of 257 ICD-10 codes and contextual evidence, ensuring a continuous pipeline that both generates and verifies outputs while meticulously logging reasons for rejections.](https://arxiv.org/html/2512.23743v1/x1.png)
A new multi-agent system, Hybrid-Code, combines the strengths of AI with robust data validation to deliver reliable and privacy-preserving diagnostic coding.
Pinpointing the Force Behind Quarks: A New Value for the Strong Coupling Constant
![The CT25 NNLO fit demonstrates that uncertainties in [latex]\alpha_{s}(M_{Z})[/latex] are modulated by the credibility level applied to both global and dynamical tolerance criteria, with the dynamic tolerance error band computed across a dataset of 49 published results establishing a quantifiable range for this sensitivity.](https://arxiv.org/html/2512.23792v1/x8.png)
A comprehensive global analysis of particle physics data yields a precise determination of αs, a fundamental parameter governing the strong nuclear force.
Unmasking Device Secrets: A Deep Dive into Power-Based Attacks
This review explores the growing threat of power side-channel attacks and the techniques used to exploit vulnerabilities in electronic devices.
Beyond Truthfulness: Gauging the Real Reliability of AI Models
As large language models become increasingly integrated into critical applications, a more nuanced understanding of their reliability-beyond simple accuracy-is essential.