Chebyshev
PulseAugur coverage of Chebyshev — every cluster mentioning Chebyshev across labs, papers, and developer communities, ranked by signal.
4 day(s) with sentiment data
-
New DeepONet Surrogate Model Enhances Transport Problem Predictions
Researchers have developed a new surrogate model called the Rationally Enriched Chebyshev (REC) trunk for DeepONets, designed to handle high-Péclet transport problems with thin boundary layers. This REC trunk integrates…
-
QuantumPhaseNet extends Transformers with quantum-inspired theory
Researchers have introduced QuantumPhaseNet, a novel framework that extends Transformer models using gauge-covariant geometric and quantum-spectral principles. This approach models context-dependent semantic states as c…
-
New method enables label-free training for finite-element surrogate models
A new research paper proposes a label-free training objective for finite-element surrogate models, utilizing discrete energy minimization. This method eliminates the need for reference solutions, directly using the asse…
-
New scale law guides detection of distribution shifts in AI embeddings
Researchers have developed a new scale law for detecting distribution shifts in high-dimensional embeddings, which constrains moment-based statistical tests. This law, derived from Chebyshev's extremal problem, suggests…
-
New spectral loss method improves chaotic dynamics modeling on unstructured meshes
Researchers have developed a novel method for modeling chaotic dynamics on unstructured meshes by adapting binned spectral losses. This approach replaces traditional Fourier modes with graph-Laplacian frequency bands, e…
-
New KAN framework discovers kernels in integro-differential equations
Researchers have developed a new framework for discovering memory and nonlocal kernels in integro-differential equations using constrained Kolmogorov--Arnold Networks (KANs). This approach aims to overcome limitations o…
-
New SNLP method boosts FHE Transformer inference efficiency
Researchers have developed a new method called Layer-Parallel Inference (SNLP) to improve the efficiency of Transformer models when performing computations on encrypted data using fully homomorphic encryption (FHE). Tra…
-
New paper precisely maps tail probability under bounded kurtosis
Researchers have determined the precise worst-case tail probability for random variables with bounded kurtosis. This analysis defines a four-regime map that details how kurtosis bounds affect one-sided tail control, rev…
-
New SEDONet architecture enhances AI approximation for scientific computing
Researchers have developed a novel Spectral-Embedded Deep Operator Network (SEDONet) architecture to improve the approximation capabilities of DeepONets for complex problems in scientific computing. Unlike standard Deep…
-
New metrics assess hardware inference complexity of Kolmogorov-Arnold Networks
A new paper introduces hardware-oriented metrics for evaluating the inference complexity of Kolmogorov-Arnold Networks (KANs). These metrics, including Real Multiplications (RM), Bit Operations (BOP), and Number of Addi…
-
Graph Neural Networks Optimized for Driving Trajectory Prediction
A new research paper explores the effectiveness of various Graph Neural Network (GNN) layers for predicting driving trajectories. The study compares 19 different graph layer types, identifying five combinations that con…
-
New paper details SoS degree barriers in robust halfspace learning
A new research paper introduces a characterization of Sum-of-Squares (SoS) degree barriers within the Reweighted-Hinge method for robust halfspace learning. The study, which focuses on learning under malicious noise, es…
-
New method for multivariate time series prediction sets unveiled
Researchers have introduced filtered conformal ellipsoids, a novel method for joint prediction sets in multivariate time series. This approach utilizes a state-space filter to emit predictive means and covariances, whic…
-
FilterMoE enhances PPGNNs with joint node-channel adaptive filtering
Researchers have developed a new approach for pre-propagation graph neural networks (PPGNNs) called FilterMoE. This method addresses the puzzle of why more complex aggregators don't always outperform simpler ones in PPG…
-
Markov's Inequality Evolves Into Concentration-of-Measure Tools
This article explores the evolution of Markov's Inequality into a broader set of concentration-of-measure tools. It details how a single substitution within the inequality can lead to more powerful bounds like Chebyshev…