Related Experiment Video
Updated: Jan 30, 2026
08:40
Isokinetic Robotic Device to Improve Test-Retest and Inter-Rater Reliability for Stretch Reflex Measurements in Stroke Patients with Spasticity
Published on: June 12, 2019
7.9K
Improving CMA-ES convergence speed, efficiency, and reliability in noisy robot optimization problems
Russell M Martin1, Steven H Collins2
1Department of Mechanical Engineering, Stanford University, Stanford, 94305, USA rumartin@stanford.edu.
Evolutionary Computation
|January 28, 2026
Summary
Adaptive Sampling CMA-ES (AS-CMA) optimizes robot policies by dynamically allocating evaluation time, improving speed and reducing costs compared to standard methods. This novel approach enhances efficiency in noisy environments with minimal setup complexity.
Area of Science:
- Robotics
- Optimization Algorithms
- Machine Learning
Background:
- Robot policy optimization is time-intensive, with evaluation time impacting speed-accuracy trade-offs.
- Current methods face challenges with noise and efficiency in complex optimization landscapes.
Purpose of the Study:
- Introduce Adaptive Sampling CMA-ES (AS-CMA), an enhancement to CMA-ES for improved optimization efficiency.
- Enable consistent precision by assigning sampling time based on predicted sorting difficulty.
Main Methods:
- Developed AS-CMA, a novel algorithm supplementing CMA-ES with adaptive sampling time allocation.
- Compared AS-CMA against CMA-ES with static sampling times and Bayesian optimization in simulated cost landscapes.
- Validated AS-CMA performance in a real-world exoskeleton optimization experiment.
Main Results:
- AS-CMA achieved convergence in 98% of runs without parameter tuning.
- AS-CMA demonstrated 24-65% faster convergence and 29-76% lower total cost than optimized CMA-ES.
- AS-CMA showed superior efficiency and reliability in complex landscapes compared to Bayesian optimization.
Conclusions:
- AS-CMA enhances optimization efficiency and reliability, particularly in noisy or complex environments.
- The adaptive sampling strategy offers a practical improvement over static sampling times.
- AS-CMA minimally increases setup complexity and tuning requirements for optimization tasks.
Related Concept Videos
01:54Convergent Evolution
32.8K
Evolution shapes the features of organisms over time, ensuring that they are suited for the environments in which they live. Sometimes, selection pressure leads to the rise of similar but unrelated adaptations in organisms with no recent common ancestors, a process known as convergent evolution.
32.8K
01:29Reliability and Validity
13.9K
Reliability and validity are two important considerations that must be made with any type of data collection. Reliability refers to the ability to consistently produce a given result. In the context of psychological research, this would mean that any instruments or tools used to collect data do so in consistent, reproducible ways.
13.9K
01:17Region of Convergence
922
The z-transform is a powerful mathematical tool used in the analysis of discrete-time signals and systems. It is a crucial tool in the analysis of discrete-time systems, but its convergence is limited to specific values of the complex variable z. This range of values, known as the Region of Convergence (ROC), is fundamental in determining the behavior and stability of a system or signal. The ROC defines the region in the complex plane where the z-transform converges, which can take various...
922
01:21Convergence of Fourier Series
400
The Fourier series is a powerful mathematical tool for representing periodic signals as an infinite sum of complex exponentials. In practice, this infinite series is truncated to a finite number of terms, yielding a partial sum. This truncation makes the approximation of the signal feasible but introduces certain challenges, particularly near discontinuities, known as the Gibbs phenomenon.
The Gibbs phenomenon refers to the persistent oscillations and overshoots that occur near discontinuities...
The Gibbs phenomenon refers to the persistent oscillations and overshoots that occur near discontinuities...
400
01:25Distribution Reliability and Automation
507
Distribution reliability in electrical power systems is critical for ensuring an uninterrupted power supply to consumers at minimal cost. According to IEEE Standard Terms, reliability is the probability that a device will function without failure over a specified time period or amount of usage. For electric power distribution, this translates to maintaining continuous power supply and addressing customer concerns over power outages. Several indices, as defined by IEEE Standard 1366-2012, are...
507
01:20Region of Convergence of Laplace Tarnsform
1.1K
The Region of Convergence (ROC) is a fundamental concept in signal processing and system analysis, particularly associated with the Laplace transform. The ROC represents an area in the complex plane where the Laplace transform of a given signal converges, determining the transform's applicability and utility.
Consider a decaying exponential signal that begins at a specific time. When deriving its Laplace transform, the time-domain variable is replaced with a complex variable. This...
Consider a decaying exponential signal that begins at a specific time. When deriving its Laplace transform, the time-domain variable is replaced with a complex variable. This...
1.1K