archive
Every paper Pith has read. Search by title, abstract, or pith.
554 papers in cs.NE · page 2
-
Newborn units in growing nets get starved of gradients
On the Stability of Growth in Structural Plasticity
-
Taxonomy sorts SNN training rules and adds shared testbed
NeuroTrain: Surveying Local Learning Rules for Spiking Neural Networks with an Open Benchmarking Framework
-
Neural solvers reach energy parity after 158000 deployments
An Amortized Efficiency Threshold for Comparing Neural and Heuristic Solvers in Combinatorial Optimization
-
Neural solvers break even on energy after ~4500 deployments
An Amortized Efficiency Threshold for Comparing Neural and Heuristic Solvers in Combinatorial Optimization
-
Evolutionary merge beats trained models on reasoning benchmarks
Darwin Family: MRI-Trust-Weighted Evolutionary Merging for Training-Free Scaling of Language-Model Reasoning
-
Sparse autoencoders sort EEG concepts into three steerability regimes
Mechanistic Interpretability of EEG Foundation Models via Sparse Autoencoders
-
Sparse autoencoders expose entangled concepts in EEG models
Mechanistic Interpretability of EEG Foundation Models via Sparse Autoencoders
-
Steering EEG model features exposes three encoding regimes
Mechanistic Interpretability of EEG Foundation Models via Sparse Autoencoders
-
The paper reconstructs a zebrafish retinotectal circuit graph and uses ablation in a…
Dual-axis attribution of zebrafish tectal microcircuits for energy-efficient and robust neurocomputing
-
Neural automata self-heal textures and graft new ones without retraining
Texture Regenerating and Grafting Using Genome-Driven Neural Cellular Automata
-
Layered weights isolate first-front ascent in multiobjective sets
Nonsmooth Set-Gradient Ascent to the Pareto Front via Layered Hypervolume and Magnitude Indicators
-
Algorithm evolves mixed-type genes separately then assembles phenotypes
The Geno-Synthetic Algorithm: Type-Factored Coevolutionary Optimization for Heterogeneous Genotypes and Assembled Phenotypes
-
Trust signal guides sparse search in large-scale multi-objective problems
TRUST-TAEA: A trustworthiness-guided two-archive evolutionary algorithm with variable-grouping sparse search for large-scale multi-objective optimization
-
Biological neurons outperform AI agents in simulated navigation
Embodied Neurocomputation: A Framework for Interfacing Biological Neural Cultures with Scaled Task-Driven Validation
-
Spiking neurons split frequency choice from timing adjustment
FiTS: Interpretable Spiking Neurons via Frequency Selectivity and Temporal Shaping
-
LLM agent with chemical tools finds drug candidates with 10% better affinity
ToolMol: Evolutionary Agentic Framework for Multi-objective Drug Discovery
-
LLM agent with RDKit tools beats drug design baselines by over 10 percent
ToolMol: Evolutionary Agentic Framework for Multi-objective Drug Discovery
-
Attractor models beat larger transformers on language and puzzles
Solve the Loop: Attractor Models for Language and Reasoning
-
Quaternion DE converges faster than real DE on BBOB benchmarks
A Family of Quaternion-Valued Differential Evolution Algorithms for Numerical Function Optimization
-
Dependency-respecting merge search halves space and adds 6.7% accuracy
Black-Box Optimization of Mixed Binary-Continuous Variables: Challenges and Opportunities in Evolutionary Model Merging
-
Knowledge graphs source optimization problems via queries
Graph-Grounded Optimization: Rao-Family Metaheuristics, Classical OR, and SLM-Driven Formulation over Knowledge Graphs
-
Graph queries for optimization reveal hidden data flaws
Graph-Grounded Optimization: Rao-Family Metaheuristics, Classical OR, and SLM-Driven Formulation over Knowledge Graphs
-
Complex neurons beat simple ones under fixed parameters
Scaling Laws and Tradeoffs in Recurrent Networks of Expressive Neurons
-
Tunable conductances let spiking nets train directly and spike less
Multi-Timescale Conductance Spiking Networks: A Sparse, Gradient-Trainable Framework with Rich Firing Dynamics for Enhanced Temporal Processing
-
Contrastive video training yields MT-like direction maps
Self-organized MT Direction Maps Emerge from Spatiotemporal Contrastive Optimization
-
Local windows replace global attention in spiking transformers
Breaking Global Self-Attention Bottlenecks in Transformer-based Spiking Neural Networks with Local Structure-Aware Self-Attention
-
ECRAM transient dynamics enable short-term plasticity in neuromorphic circuits
Leveraging Non-Equilibrium ECRAM Dynamics for Short-Term Plasticity in Neuromorphic Circuits
-
Crossover speeds up NSGA-III on many-objective benchmarks
On the Impact of Crossover in Many-Objective Optimization: A Runtime Analysis of NSGA-III
-
Router rewrite alone accounts for all gains in evolutionary MoLoRA
Decomposing Evolutionary Mixture-of-LoRA Architectures: The Routing Lever, the Lifecycle Penalty, and a Substrate-Conditional Boundary
-
SRC neurons classify MNIST on FPGA at 0.45 mJ per digit
Energy-Efficient Implementation of Spiking Recurrent Cells on FPGA
-
FPGA SNN with SRC cells hits 96% MNIST accuracy at 0.45 mJ per digit
Energy-Efficient Implementation of Spiking Recurrent Cells on FPGA
-
Equilibrium in NeuroAI stabilizes neural
A Theory of Multilevel Interactive Equilibrium in NeuroAI
-
ReLU geometry yields exact causal explanations without distillation
Causal Explanations from the Geometric Properties of ReLU Neural Networks
-
Meta-policy guides search regions in constrained multi-objective optimization
Meta-Black-Box Optimization Can Do Search Guidance for Expensive Constrained Multi-Objective Optimization
-
Sparse coding and timing preserve knowledge across context switches
Joint sparse coding and temporal dynamics support context reconfiguration
-
Humans Compress Future Tasks When Building Abstractions
Prospective Compression in Human Abstraction Learning
-
LIF bandwidth matching boosts SNN accuracy on mmWave sensing
Frequency Matching in Spiking Neural Networks for mmWave Sensing
-
Evolving prompt embeddings raises LLM solution coverage and quality-diversity scores
Parameter-Efficient Neuroevolution for Diverse LLM Generation: Quality-Diversity Optimization via Prompt Embedding Evolution
-
Evolutionary optimization yields more diverse LLM alignments
EvoPref: Multi-Objective Evolutionary Optimization Discovers Diverse LLM Alignments Beyond Gradient Descent
-
Spike encoders recast as causal wavelets with error bounds
Encoding and Decoding Temporal Signals with Spiking Bandpass Wavelets
-
Search architecture upgrades beat bigger LLMs for evolutionary tasks
LEVI: Stronger Search Architectures Can Substitute for Larger LLMs in Evolutionary Search
-
AutoSINDy recovers true dynamics in 92.8 percent of noisy trials
Discovery of Nonlinear Dynamics with Automated Basis Function Generation
-
RDEx-CASK cuts time-to-target on CEC constrained problems
RDEx-CASK: Cauchy Mutation, Archive, and Stagnation Kick for RDEx-CSOP
-
Neuromorphic training cuts quadruped controller memory use by 4.3 times
Neuromorphic Reinforcement Learning for Quadruped Locomotion Control on Uneven Terrain
-
Sparse FFN routing shifts work into attention layers
Sparsity Moves Computation: How FFN Architecture Reshapes Attention in Small Transformers
-
Sparse FFN routing shifts computation to attention in Transformers
Sparsity Moves Computation: How FFN Architecture Reshapes Attention in Small Transformers
-
Co-evolving agent ensembles surpass static coding limits
Evolutionary Ensemble of Agents
-
Self-revising agent ensemble breaks static code limits
Evolutionary Ensemble of Agents
-
LLM-designed LSHADE records 510 wins on GNBG benchmark
ARES-LSHADE: Autoresearch-Enhanced LSHADE with Memetic Polish for the GNBG Benchmark
-
LLM research loop builds GNBG solver hitting machine precision on 18 functions
ARES-LSHADE: Autoresearch-Enhanced LSHADE with Memetic Polish for the GNBG Benchmark