Data-driven experimental design and model development using Gaussian process with active learning

Chang, Jiangfang; Kim, Jiseob; Zhang, Byoung-Tak; Pitt, Mark A.; Myung, Jay I.

doi:10.1016/j.cogpsych.2020.101360

Cited by 8 publications

(2 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…GPO builds a surrogate model of the real loss scriptL that incorporates noise − and allows for nonlocal search , (see section VII in the Supporting Information). As an active learning technique, it guides the sampling of new parameters, improving optimization efficiency. − …”

Section: Resultsmentioning

confidence: 99%

Accurate Memory Kernel Extraction from Discretized Time-Series Data

Tepper,

Dalton,

Netz

2024

J. Chem. Theory Comput.

View full text Add to dashboard Cite

Memory effects emerge as a fundamental consequence of dimensionality reduction when low-dimensional observables are used to describe the dynamics of complex many-body systems. In the context of molecular dynamics (MD) data analysis, accounting for memory effects using the framework of the generalized Langevin equation (GLE) has proven efficient, accurate, and insightful, particularly when working with high-resolution time series data. However, in experimental systems, highresolution data are often unavailable, raising questions about the impact of the data resolution on the estimated GLE parameters. This study demonstrates that direct memory extraction from time series data remains accurate when the discretization time is below the memory time. To obtain memory functions reliably, even when the discretization time exceeds the memory time, we introduce a Gaussian Process Optimization (GPO) scheme. This scheme minimizes the deviation of discretized two-point correlation functions between time series data and GLE simulations and is able to estimate accurate memory kernels as long as the discretization time stays below the longest time scale in the data, typically the barrier crossing time.

show abstract

Section: Resultsmentioning

confidence: 99%

Accurate Memory Kernel Extraction from Discretized Time-Series Data

Tepper,

Dalton,

Netz

2024

J. Chem. Theory Comput.

View full text Add to dashboard Cite

show abstract

“…non-parametric stimulus selection using Gaussian process regression (Schulz et al 2018) or Gaussian process classification (Chang et al 2021) can be used. Although this paper assumed the situation where we know the true model for simplicity, the effect of individual-level stimulus selection methods in the situation where researchers use the above advanced methods should be investigated in a future study.…”

Section: Limitation and Future Researchmentioning

confidence: 99%

The effect of individual-level adaptive stimulus selection on the group-level parameters for cognitive models

2023

View full text Add to dashboard Cite

In psychological experiments, researchers' primary interests often gravitate towards group-level parameters such as the mean difference between experimental conditions. However, existing studies on adaptive stimulus selection have essentially focused on individual-level parameters. Hence, the present paper investigated the effect of adaptive stimulus selection on group-level parameters. Two simulation studies indicated that the individual-level adaptive stimulus selection led to higher estimation precision of group-level parameters than random stimulus selection on elaborated decision-making models. Based on these results, it is suggested that individual-level adaptive stimulus selection methods be adopted even cases where the research interest leans toward group-level parameters.

show abstract

A systematic review of unique methods for measuring discount rates

Lipman,

Attema

2024

J Risk Uncertain

View full text Add to dashboard Cite

Discount rates play a pivotal role in various fields, capturing the trade-off between present and future, influencing predictions of unhealthy behaviors. Despite substantial variations in discount rates across studies, methods, and individuals, our goal was to systematically overview diverse methods for measuring discount rates. We conducted a comprehensive literature review across multiple databases, targeting English studies introducing and utilizing discounting measurement methods in human subjects. Two authors screened titles and abstracts, with full text review split between them. Extracted data encompassed bibliographic details, theoretical aspects (e.g., discount function), and operational features (e.g., elicitation procedure). After deduplication, 4976, 218, and 83 records underwent title and abstract screening, full text screening, and inclusion, respectively. A total of 86 unique methods were identified. Predominantly, methods were devised for money (75%) and health (22%) discounting. Network analysis on citations indicated limited cross-disciplinary overlap. Only about one fourth, one third, and one third of methods possessed theoretically desirable traits: i) allowance for negative discount rates, ii) application of multiple discount functions, and iii) correction for non-linear utility of outcomes. Diverse discounting measurement methods exist, showcasing varied theoretical and operational characteristics. These differences may stem from the isolation of development in fields like psychology, experimental economics, and health economics. Our systematic review aids readers in choosing methods aligned with their priorities when measuring discount rates.

show abstract

Data-driven experimental design and model development using Gaussian process with active learning

Cited by 8 publications

References 33 publications

Accurate Memory Kernel Extraction from Discretized Time-Series Data

Accurate Memory Kernel Extraction from Discretized Time-Series Data

The effect of individual-level adaptive stimulus selection on the group-level parameters for cognitive models

A systematic review of unique methods for measuring discount rates

Contact Info

Product

Resources

About