Unlocking Hidden Charm: Predicting Tetraquark Decay Rates
![The study demonstrates how variations in Borel parameters influence hadronic coupling constants - specifically, the relationship between [latex]G_{\chi_{c1}\rho Z_{c}^{-}}[/latex] and [latex]G_{\chi_{c0}\pi Z_{c}^{+}}[/latex] - suggesting that even fundamental constants may be subject to shifts dependent on the framework used to observe them.](https://arxiv.org/html/2603.18877v1/x1.png)
New calculations using QCD sum rules offer theoretical predictions for the strong decays of exotic hidden-charm tetraquark states, paving the way for experimental validation.
Science
How Reasoning Unfolds: Predicting AI Reliability Step-by-Step
New research reveals that the pattern of uncertainty during an AI’s thought process is a more accurate indicator of a correct answer than the overall level of doubt.
Unlocking SIMON32: A New Approach to Lightweight IoT Security
Researchers have developed a refined differential cryptanalysis technique targeting the SIMON32 algorithm, improving the efficiency of security assessments for resource-constrained devices.
Beyond Confidence: A New Approach to Spotting AI’s Blind Spots
![CORE dissects input features into components aligned with classification-[latex]z_{\parallel}[/latex]-and orthogonal residuals-[latex]z_{\perp}[/latex]-the latter of which encodes a membership signal demonstrably more discerning than simple confidence metrics and inherently undetectable by methods relying solely on logits.](https://arxiv.org/html/2603.18290v1/x5.png)
Researchers have developed a novel method for reliably detecting when artificial intelligence systems encounter data outside of their training, improving safety and trustworthiness.
Hiding in Plain Sight: Adaptive Steganography with Fuzzy Logic
A new framework uses fuzzy logic to intelligently embed hidden data within images, balancing imperceptibility and resilience against detection.
Flowing Past the Sign Problem: A New Path for Simulating Materials
![The study demonstrates a nuanced correspondence between exact diagonalization and an annealing-based neural network ansatz-specifically, for the two-site Hubbard model at [latex] U=3 [/latex], [latex] \beta=5 [/latex], [latex] \mu=2 [/latex], and with [latex] N_t=20 [/latex] states-suggesting the network accurately captures the system’s underlying correlations despite the inherent complexity of fermionic systems.](https://arxiv.org/html/2603.18205v1/x5.png)
Researchers are employing deep generative models to overcome long-standing challenges in accurately simulating complex quantum materials at finite density.
Guiding AI Reasoning: A New Architecture for Reliable Answers
Researchers have developed a process-control framework designed to significantly improve the trustworthiness of large language models and reduce the risk of fabricated responses.
Finding Hidden Bugs in Sparse Tensor Compilation
A new fuzzing framework, TenSure, systematically uncovers errors in the increasingly complex process of compiling code for sparse tensors.
Beyond Accuracy: The Hidden Costs of Phishing Defense
New research reveals that high accuracy in phishing detection doesn’t necessarily translate to security, as attackers can exploit low-cost feature manipulation to bypass defenses.
Beyond Residuals: Guaranteeing Accuracy in Neural PDE Solutions
A new framework provides verifiable bounds on the error of neural network-based solvers for partial differential equations, moving beyond empirical observation to solution-space guarantees.