Bayesian optimization of peripheral intraneural stimulation protocols to evoke distal limb movements

2021 ◽  
Author(s):  
Elena Losanno ◽  
Marion Badi ◽  
Sophie Wurth ◽  
Simon Borgognon ◽  
Gregoire Courtine ◽  
...  
Keyword(s):  
Search Space ◽  
Bayesian Optimization ◽  
Objective Functions ◽  
Limb Movements ◽  
Distal Limb ◽  
Motor Functions ◽  
Manual Search ◽  
Online Test ◽  
Stimulation Parameters ◽  
Peripheral Motor

Abstract Objective. Motor neuroprostheses require the identification of stimulation protocols that effectively produce desired movements. Manual search for these protocols can be very time-consuming and often leads to suboptimal solutions, as several stimulation parameters must be personalized for each subject for a variety of target movements. Here, we present an algorithm that efficiently tunes peripheral intraneural stimulation protocols to elicit functionally relevant distal limb movements. Approach. We developed the algorithm using Bayesian Optimization and defined multi-objective functions based on the coordinated recruitment of multiple muscles. We implemented different multi-output Gaussian Processes to model our system and compared their functioning by applying the algorithm offline to data acquired in rats for walking control and in monkeys for hand grasping control. We then performed a preliminary online test in a monkey to experimentally validate the functionality of our method. Main results. Offline, optimal intraneural stimulation protocols for various target motor functions were rapidly identified in both experimental scenarios. Using the model that performed best, the algorithm converged to stimuli that evoked functionally consistent movements with an average number of actions equal to 20% (13% unique) and 20% (16% unique) of the search space size in rats and monkeys, respectively. Online, the algorithm quickly guided the observations to stimuli that elicited functional hand gestures, although more selective motor outputs could have been achieved by refining the objective function used. Significance. These results demonstrate that Bayesian Optimization can reliably and efficiently automate the tuning of peripheral neurostimulation protocols, establishing a translational framework to configure peripheral motor neuroprostheses in clinical applications. The proposed method can potentially be applied to optimize motor functions using other stimulation modalities.

A Memetic Optimization Strategy Based on Dimension Reduction in Decision Space

2015 ◽  
Vol 23 (1) ◽  
pp. 69-100 ◽  
Author(s):  
Handing Wang ◽  
Licheng Jiao ◽  
Ronghua Shang ◽  
Shan He ◽  
Fang Liu
Keyword(s):  
Dimension Reduction ◽  
Optimization Problems ◽  
State Of The Art ◽  
Search Space ◽  
Optimization Strategy ◽  
Objective Functions ◽  
Multi Objective Optimization ◽  
Decision Space ◽  
Multi Objective ◽  
Decision Variables

There can be a complicated mapping relation between decision variables and objective functions in multi-objective optimization problems (MOPs). It is uncommon that decision variables influence objective functions equally. Decision variables act differently in different objective functions. Hence, often, the mapping relation is unbalanced, which causes some redundancy during the search in a decision space. In response to this scenario, we propose a novel memetic (multi-objective) optimization strategy based on dimension reduction in decision space (DRMOS). DRMOS firstly analyzes the mapping relation between decision variables and objective functions. Then, it reduces the dimension of the search space by dividing the decision space into several subspaces according to the obtained relation. Finally, it improves the population by the memetic local search strategies in these decision subspaces separately. Further, DRMOS has good portability to other multi-objective evolutionary algorithms (MOEAs); that is, it is easily compatible with existing MOEAs. In order to evaluate its performance, we embed DRMOS in several state of the art MOEAs to facilitate our experiments. The results show that DRMOS has the advantage in terms of convergence speed, diversity maintenance, and portability when solving MOPs with an unbalanced mapping relation between decision variables and objective functions.

scholarly journals Abandoning Objectives: Evolution Through the Search for Novelty Alone

2011 ◽  
Vol 19 (2) ◽  
pp. 189-223 ◽  
Author(s):  
Joel Lehman ◽  
Kenneth O. Stanley
Keyword(s):  
Fitness Function ◽  
Search Space ◽  
Objective Functions ◽  
Natural Evolution ◽  
Inherent Limitation ◽  
Novelty Search ◽  
Maze Navigation ◽  
New Perspective ◽  
Underlying Pathology ◽  
Real World Problems

In evolutionary computation, the fitness function normally measures progress toward an objective in the search space, effectively acting as an objective function. Through deception, such objective functions may actually prevent the objective from being reached. While methods exist to mitigate deception, they leave the underlying pathology untreated: Objective functions themselves may actively misdirect search toward dead ends. This paper proposes an approach to circumventing deception that also yields a new perspective on open-ended evolution. Instead of either explicitly seeking an objective or modeling natural evolution to capture open-endedness, the idea is to simply search for behavioral novelty. Even in an objective-based problem, such novelty search ignores the objective. Because many points in the search space collapse to a single behavior, the search for novelty is often feasible. Furthermore, because there are only so many simple behaviors, the search for novelty leads to increasing complexity. By decoupling open-ended search from artificial life worlds, the search for novelty is applicable to real world problems. Counterintuitively, in the maze navigation and biped walking tasks in this paper, novelty search significantly outperforms objective-based search, suggesting the strange conclusion that some problems are best solved by methods that ignore the objective. The main lesson is the inherent limitation of the objective-based paradigm and the unexploited opportunity to guide search through other means.

An Approach for the Fast Calculation Method of Pareto Solutions of a Two-objective Network

2015 ◽  
Vol 22 (01) ◽  
pp. 1550005 ◽  
Author(s):  
Natsumi Takahashi ◽  
Tomoaki Akiba ◽  
Shuhei Nomura ◽  
Hisashi Yamamoto
Keyword(s):  
Shortest Path ◽  
Optimization Problem ◽  
Computing Time ◽  
Search Space ◽  
Shortest Path Problem ◽  
Objective Functions ◽  
Pareto Solutions ◽  
Fast Calculation ◽  
Multi Objective ◽  
Other Information

The shortest path problem is a kind of optimization problem and its aim is to find the shortest path connecting two specific nodes in a network, where each edge has its distance. When considering not only the distances between the nodes but also some other information, the problem is formulated as a multi-objective shortest path problem that involves multiple conflicting objective functions. The multi-objective shortest path problem is a kind of optimization problem of multi-objective network. In the general cases, multi-objectives are rarely optimized by a solution. So, to solve the multi-objective shortest path problem leads to obtaining Pareto solutions. An algorithm for this problem has been proposed by using the extended Dijkstra's algorithm. However, this algorithm for obtaining Pareto solutions has many useless searches for paths. In this study, we consider two-objective shortest path problem and propose efficient algorithms for obtaining the Pareto solutions. Our proposed algorithm can reduce more search space than existing algorithms, by solving a single-objective shortest path problem. The results of the numerical experiments suggest that our proposed algorithms reduce the computing time and the memory size for obtaining the Pareto solutions.

scholarly journals BOKEI: Bayesian Optimization Using Knowledge of Correlated Torsions and Expected Improvement for Conformer Generation​

2019 ◽  
Author(s):  
Lucian Chan ◽  
Geoffrey Hutchison ◽  
Garrett Morris
Keyword(s):  
Random Search ◽  
Search Space ◽  
Empirical Method ◽  
Fixed Number ◽  
Low Energy ◽  
Intramolecular Interactions ◽  
Bayesian Optimization ◽  
Superior Performance ◽  
Expected Improvement ◽  
Wide Range

<div>A key challenge in conformer sampling is to find low-energy conformations with a small number of energy evaluations. We have recently demonstrated Bayesian optimization as an effective method to search for energetically favorable conformations. This approach balances between <i>exploitation</i> and <i>exploration</i>, and lead to superior performance when compared to exhaustive or random search methods. In this work, we extend strategies on proteins and oligopeptides (e.g. Ramachandran plots of secondary structure) to study the correlated torsions in small molecules. We use a bivariate von Mises distribution to capture the correlations, and use it to constrain the search space. We validate the performance of our Bayesian optimization with prior knowledge (BOKEI) on a dataset consisting of 533 diverse small organic molecules, using a force field (MMFF94) and a semi empirical method (GFN2). We compare BOKEI with Bayesian optimization with expected improvement (BOA-EI), and a genetic algorithm (GA), using a fixed number of energy evaluations. In 70(± 2.1)% of the cases examined, BOKEI finds lower energy conformations than global optimization with BOA-EI or GA. More importantly, these patterns find correlated torsions in 10-15% of molecules in larger data sets, 3-8 times more frequently than previous work. We also find that the BOKEI patterns not only describe steric clashes, but also reflect favorable intramolecular interactions, including hydrogen bonds and π-π stacking. Further understanding of the conformational preferences of molecules will help find low energy conformers efficiently for a wide range of computational modeling applications.</div>

scholarly journals Design of Mechanical Metamaterials via Constrained Bayesian Optimization

2018 ◽  
Author(s):  
Conner Sharpe ◽  
Carolyn Conner Seepersad ◽  
Seth Watts ◽  
Dan Tortorelli
Keyword(s):  
Design Space ◽  
Engineering Problem ◽  
Bayesian Optimization ◽  
Probabilistic Constraints ◽  
Objective Functions ◽  
Bulk Properties ◽  
New Class ◽  
Naturally Occurring ◽  
Design Variables ◽  
Mechanical Metamaterials

Advances in additive manufacturing processes have made it possible to build mechanical metamaterials with bulk properties that exceed those of naturally occurring materials. One class of these metamaterials is structural lattices that can achieve high stiffness to weight ratios. Recent work on geometric projection approaches has introduced the possibility of optimizing these architected lattice designs in a drastically reduced parameter space. The reduced number of design variables enables application of a new class of methods for exploring the design space. This work investigates the use of Bayesian optimization, a technique for global optimization of expensive non-convex objective functions through surrogate modeling. We utilize formulations for implementing probabilistic constraints in Bayesian optimization to aid convergence in this highly constrained engineering problem, and demonstrate results with a variety of stiff lightweight lattice designs.

Bayesian Optimization in Weakly Specified Search Space

2017 ◽  
Author(s):  
Vu Nguyen ◽  
Sunil Gupta ◽  
Santu Rane ◽  
Cheng Li ◽  
Svetha Venkatesh
Keyword(s):  
Search Space ◽  
Bayesian Optimization

scholarly journals Machine-learning-assisted fabrication: Bayesian optimization of laser-induced graphene patterning using in-situ Raman analysis

2020 ◽  
Author(s):  
Hud Wahab ◽  
Vivek Jain ◽  
Alexander Scott Tyrrell ◽  
Michael Alan Seas ◽  
Lars Kotthoff ◽  
...  
Keyword(s):  
Machine Learning ◽  
Irradiation Time ◽  
Electronic Devices ◽  
Parameter Tuning ◽  
Search Space ◽  
Underlying Mechanism ◽  
Single Step ◽  
Bayesian Optimization ◽  
High Quality ◽  
Optimization Search

The control of the physical, chemical, and electronic properties of laser-induced graphene (LIG) is crucial in the fabrication of flexible electronic devices. However, the optimization of LIG production is time-consuming and costly. Here, we demonstrate state-of-the-art automated parameter tuning techniques using Bayesian optimization to advance rapid single-step laser patterning and structuring capabilities with a view to fabricate graphene-based electronic devices. In particular, a large search space of parameters for LIG explored efficiently. As a result, high-quality LIG patterns exhibiting high Raman G/D ratios at least a factor of four larger than those found in the literature were achieved within 50 optimization iterations in which the laser power, irradiation time, pressure and type of gas were optimized. Human-interpretable conclusions may be derived from our machine learning model to aid our understanding of the underlying mechanism for substrate-dependent LIG growth, e.g. high-quality graphene patterns are obtained at low and high gas pressures for quartz and polyimide, respectively. Our Bayesian optimization search method allows for an efficient experimental design that is independent of the experience and skills of individual researchers, while reducing experimental time and cost and accelerating materials research.

scholarly journals Personalizing Dual-Target Cortical Stimulation with Bayesian Parameter Optimization Successfully Treats Central Post-Stroke Pain: A Case Report

2021 ◽  
Vol 12 (1) ◽  
pp. 25
Author(s):  
Evan M. Dastin-van Rijn ◽  
Seth D. König ◽  
Danielle Carlson ◽  
Vasudha Goel ◽  
Andrew Grande ◽  
...  
Keyword(s):  
Case Report ◽  
Memory Task ◽  
Cortical Stimulation ◽  
Bayesian Optimization ◽  
Systematic Evaluation ◽  
Central Pain ◽  
Post Stroke ◽  
Stimulation Parameters ◽  
Dorsolateral Prefrontal ◽  
Personalized Approach

Central pain disorders, such as central post-stroke pain, remain clinically challenging to treat, despite many decades of pharmacological advances and the evolution of neuromodulation. For treatment refractory cases, previous studies have highlighted some benefits of cortical stimulation. Recent advances in new targets for pain and the optimization of neuromodulation encouraged our group to develop a dual cortical target approach paired with Bayesian optimization to provide a personalized treatment. Here, we present a case report of a woman who developed left-sided facial pain after multiple thalamic strokes. All previous pharmacologic and interventional treatments failed to mitigate the pain, leaving her incapacitated due to pain and medication side effects. She subsequently underwent a single burr hole for placement of motor cortex (M1) and dorsolateral prefrontal cortex (dlPFC) paddles for stimulation with externalization. By using Bayesian optimization to find optimal stimulation parameters and stimulation sites, we were able to reduce pain from an 8.5/10 to a 0/10 during a 5-day inpatient stay, with pain staying at or below a 2/10 one-month post-procedure. We found optimal treatment to be simultaneous stimulation of M1 and dlPFC without any evidence of seizure induction. In addition, we found no worsening in cognitive performance during a working memory task with dlPFC stimulation. This personalized approach using Bayesian optimization may provide a new foundation for treating central pain and other functional disorders through systematic evaluation of stimulation parameters.

scholarly journals High Dimensional Bayesian Optimization via Supervised Dimension Reduction

2019 ◽  
Author(s):  
Miao Zhang ◽  
Huiqi Li ◽  
Steven Su
Keyword(s):  
Dimension Reduction ◽  
Reduction Method ◽  
Bayesian Optimization ◽  
High Dimensional ◽  
Inverse Regression ◽  
Objective Functions ◽  
High Dimensions ◽  
Expensive Problems ◽  
Regret Bounds ◽  
Supervised Dimension Reduction

Bayesian optimization (BO) has been broadly applied to computational expensive problems, but it is still challenging to extend BO to high dimensions. Existing works are usually under strict assumption of an additive or a linear embedding structure for objective functions. This paper directly introduces a supervised dimension reduction method, Sliced Inverse Regression (SIR), to high dimensional Bayesian optimization, which could effectively learn the intrinsic sub-structure of objective function during the optimization. Furthermore, a kernel trick is developed to reduce computational complexity and learn nonlinear subset of the unknowing function when applying SIR to extremely high dimensional BO. We present several computational benefits and derive theoretical regret bounds of our algorithm. Extensive experiments on synthetic examples and two real applications demonstrate the superiority of our algorithms for high dimensional Bayesian optimization.

Sign in / Sign up

Export Citation Format

Share Document