A certified gradient-based method for contact-rich manipulation that quantifies smoothing-induced errors via set-valued discrepancies and incorporates them into analytical reachable sets for robust affine feedback policies.
hub Mixed citations
The CMA Evolution Strategy: A Tutorial
Mixed citation behavior. Most common role is background (50%).
65
Pith papers citing it
Background
50% of classified citations
abstract
This tutorial introduces the CMA Evolution Strategy (ES), where CMA stands for Covariance Matrix Adaptation. The CMA-ES is a stochastic, or randomized, method for real-parameter (continuous domain) optimization of non-linear, non-convex functions. We try to motivate and derive the algorithm from intuitive concepts and from requirements of non-linear, non-convex search in continuous domain.
hub tools
citation-role summary
background 4
method 2
citation-polarity summary
representative citing papers
Promptbreeder evolves both task prompts and the mutation prompts that improve them using LLMs, outperforming Chain-of-Thought and Plan-and-Solve on arithmetic and commonsense reasoning benchmarks.
Introduces homogeneous progress model and proves that for Z = N(-δ,1) with μ ≤ e^δ the growth rate R_μ equals (log^{1+o(1)} μ / μ) R_1.
Latent Heuristic Search performs continuous optimization over learned embeddings of heuristics, using normalizing flows and LLM prompting to discover competitive solvers for TSP, CVRP, KSP, and OBP.
A Q/D-space framework supplies sufficient order conditions for explicit Runge-Kutta methods and supports a recursive construction of even-order methods with stage count (p²-2p+8)/4.
EVA-0 is a zeroth-order test-time adaptation method that uses scale-invariant loss, anchor-guided optimization, and symmetric two-sided perturbations to enable inference and adaptation in two forward passes, outperforming prior methods on ImageNet-C with ViT-Base.
EGGROLL applies low-rank evolution strategies to train leaky integrate-and-fire spiking neural networks, reaching 79.21% accuracy on N-MNIST with 2.23 times lower per-generation time than full-rank ES.
EvoPref applies NSGA-II evolutionary optimization with archive-based diversity to populations of LoRA adapters, yielding 18% higher preference coverage and 47% lower collapse than gradient descent baselines while matching alignment quality.
ENMP prunes negative LoRA modules via evolutionary search to boost merging performance to new state-of-the-art levels across language and vision tasks.
Proprioceptive distribution matching adapts simulators for legged robot policies by comparing observation and action distributions, reducing sim-to-real gaps with minimal real data and no external sensing.
A bootstrap strategy for non-unitary CFTs uses statistical stability of OPE data across cross-ratios to optimize spectra, reproducing A-series minimal models and yielding candidate solutions for c>1.
Introduces observation traveling salesman distance and observation entropy to quantify exploration in Bayesian optimization acquisition functions and links them to empirical performance.
Unifies ES, CBO, and OVI black-box optimizers via two design axes and proposes hybrid methods that outperform base algorithms on benchmarks.
Reformulates caging as an eikonal minimum-time escape field, approximated via physics-informed neural network and embedded into whole-arm manipulation planning for improved robustness to contact-model mismatch.
Evolutionary optimization discovers developmental reward schedules that improve performance over extrinsic-only baselines on some MiniGrid tasks, with novelty emerging as the dominant early signal.
Optimizing environmental and demographic parameters in a JAX-based agent-based model with RNN controllers produces population dynamics resembling Lotka-Volterra cycles using a feature-based loss.
A learned linear multi-factor value model over seven cognitive psychology factors retains 0.770 gold evidence on LongMemEval blind regime versus 0.368 for recency and 0.518 for best single factor.
3D-URAM decouples radio map reconstruction into Bayesian UNet recovery with uncertainty and transformer-based PPO waypoint selection, reporting over 50% error reduction in simulations and real 300x200x100m field tests.
CoRP consolidates reward-weighted perturbations into a single model via low-rank structure, improving base LLMs by 8.1 points on average while using one-tenth the budget of prior ensembles and one forward pass.
Deep bibliography expansion in literature search achieves high recall while human citations are found to have only 51% moderate relevance compared to 86-88% for AI methods.
ECo-MoE co-optimizes latent robot genotypes and a gated mixture of control experts to improve evolvability in robot body-controller co-design.
KSOS-BO improves acquisition function optimization in Bayesian optimization by casting it as a kernel sum of squares semidefinite program, outperforming Sobol, DE, and CMA-ES baselines on 10/15 benchmarks with 81% average regret reduction.
Recasts sampling-based nonconvex optimization as smoothed gradient descent to obtain non-asymptotic convergence guarantees and introduces the DIDA annealed algorithm that converges to the global optimum.
Deep Boltzmann Quantum States with natural-gradient optimization and annealing-like training match exact or best-known solutions for large infinite-range Ising spin glasses and solve job shop scheduling instances.
citing papers explorer
-
Evolutionary Negative Module Pruning for Better LoRA Merging
ENMP prunes negative LoRA modules via evolutionary search to boost merging performance to new state-of-the-art levels across language and vision tasks.