Neural & Evolutionary Computing
Research on neural network architectures, evolutionary algorithms, and bio-inspired computing.
cs.NE · 188 papersPrimitive Recursion without Composition: Dynamical Characterizations, from Neural Networks to Polynomial ODEs
This paper shows primitive recursion is equivalently characterized by recurrent ReLU networks, polynomial ODEs, and polynomial maps, revealing their computational strengths and weaknesses.
MAEO: Multiobjective Animorphic Ensemble Optimization for Scalable Large-scale Engineering Applications
MAEO is a parallel ensemble optimization framework that unifies multiple evolutionary algorithms to achieve superior multiobjective performance across complex engineering problems.
Necessary and sufficient conditions for universality of Kolmogorov-Arnold networks
This paper establishes necessary and sufficient conditions for the universal approximation property of Kolmogorov-Arnold Networks (KANs).
Learn&Drop: Fast Learning of CNNs based on Layer Dropping
Learn&Drop dynamically drops CNN layers during training to halve training time and reduce FLOPs without losing accuracy.
Why Architecture Choice Matters in Symbolic Regression
Architecture choice, not just expressiveness, critically determines the success of gradient-based symbolic regression in recovering target formulas.
A Multiplication-Free Spike-Time Learning Algorithm and its Efficient FPGA Implementation for On-Chip SNN Training
This paper introduces a multiplication-free, spike-time learning algorithm for efficient on-chip SNN training on FPGAs, achieving high accuracy and low resource use.
Collocation-based Robust Physics Informed Neural Networks for time-dependent simulations of pollution propagation under thermal inversion conditions on Spitsbergen
This paper introduces a robust, collocation-based PINN for time-dependent pollution simulation, revealing how thermal inversion increases PM.
Structure-Guided Diffusion Model for EEG-Based Visual Cognition Reconstruction
SGDM uses structural information to reconstruct visual cognition from EEG signals, outperforming existing methods in fidelity and generalization.
HubRouter: A Pluggable Sub-Quadratic Routing Primitive for Hybrid Sequence Models
HubRouter is a pluggable module that replaces O(n^2) attention with O(nM) hub-mediated routing, offering significant throughput gains.
A Co-Evolutionary Theory of Human-AI Coexistence: Mutualism, Governance, and Dynamics in Complex Societies
This paper proposes a co-evolutionary theory of human-AI coexistence based on conditional mutualism and governance, moving beyond obedience.
LTBs-KAN: Linear-Time B-splines Kolmogorov-Arnold Networks
LTBs-KAN introduces a linear-time B-spline Kolmogorov-Arnold Network, significantly speeding up KANs while reducing parameters.
L-System Genetic Encoding for Scalable Neural Network Evolution: A Comparison with Direct Matrix Encoding
Lsys genetic encoding dramatically improves neural network evolution over direct matrix encoding, showing superior performance, reliability, and generalization.
Multi-Task Optimization over Networks of Tasks
MONET is a new multi-task optimization algorithm that models task spaces as graphs, combining social and individual learning to outperform existing methods.
Neuromorphic Computing Based on Parametrically-Driven Oscillators and Frequency Combs
This paper explores neuromorphic computing with parametrically-driven oscillators, finding optimal performance in the parametric resonance regime.
Geometric Monomial (GEM): a family of rational 2N-differentiable activation functions
Introduces GEM, a family of C^2N-smooth, rational activation functions that outperform GELU on various benchmarks, improving deep learning optimization.
On the Role of Preprocessing and Memristor Dynamics in Reservoir Computing for Image Classification
Analyzes memristor dynamics and preprocessing in reservoir computing for image classification, achieving high MNIST accuracy and robustness.
Novelty-Based Generation of Continuous Landscapes with Diverse Local Optima Networks
This paper introduces a novel method to efficiently generate diverse continuous landscapes with tunable multimodality and their Local Optima Networks.
Trust-SSL: Additive-Residual Selective Invariance for Robust Aerial Self-Supervised Learning
Trust-SSL improves self-supervised learning robustness for aerial imagery by introducing a per-sample, per-factor trust weight and additive-residual objective.
Focus Session: Hardware and Software Techniques for Accelerating Multimodal Foundation Models
This paper presents a multi-layered hardware/software co-design methodology to efficiently accelerate multimodal foundation models, reducing computational and memory needs.
CO$_2$ sequestration hybrid solver using isogeometric alternating-directions and collocation-based robust variational physics informed neural networks (IGA-ADS-CRVPINN)
This paper introduces a hybrid IGA-ADS-CRVPINN solver for CO2 sequestration, achieving over 3x speedup compared to traditional methods.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.