An official website of the United States government
Aggregating person-level data across multiple clinical study sites is often constrained by privacy regulations, necessitating the development of decentralized modeling approaches in biomedical research. To address this requirement, a federated nonlinear regression algorithm based on the Choquet integral has been introduced for outcome prediction. This approach avoids reliance on prior statistical assumptions about data distribution and captures feature interactions, reflecting the non-additive nature of biomedical data characteristics. This work represents the first theoretical application of Choquet integral regression to multisite longitudinal trial data within a federated learning framework. The Multiple Imputation Choquet Integral Regression with LASSO (MIChoquet-LASSO) algorithm is specifically designed to reduce overfitting and enable variable selection in federated learning settings. Its performance has been evaluated using synthetic datasets, publicly available biomedical datasets, and proprietary longitudinal randomized controlled trial data. Comparative evaluations were conducted against benchmark methods, including ordinary least squares (OLS) regression and Choquet-OLS regression, under various scenarios such as model misspecification and both linear and nonlinear data structures in non-federated and federated contexts. Mean squared error was used as the primary performance metric. Results indicate that MIChoquet-LASSO outperforms compared models in handling nonlinear longitudinal data with missing values, particularly in scenarios prone to overfitting. In federated settings, Choquet-OLS underperforms, whereas the federated variant of the model, FEDMIChoquet-LASSO, demonstrates consistently better performance. These findings suggest that FEDMIChoquet-LASSO offers a reliable solution for outcome prediction in multisite longitudinal trials, addressing challenges such as missing values, nonlinear relationships, and privacy constraints while maintaining strong performance within the federated learning framework.
more »
« less
A Choquet-Integral Based Approach To Identify Weight Loss Component Subsets
Obesity is a common, serious, and costly chronic disease in the United States, a risk factor for several major cancers, linked to higher rates of illness and death. It is thus a critical issue that needs attention from health care professionals and the public alike. We use a novel approach to target nonstandard variations to better understand the variables associated with weight loss. We introduce a new methodology using the Choquet Integral with fuzzy measure, an approach that accounts for interactions between measured features. The Choquet Integral has limited sourced applications to the biomedical field despite widespread use in theoretical mathematics and economics. Our technique applies it to health data to show a robust method to target and optimize weight loss parameters. We identify data versus noise, optimally choose a reduced version of the powerset for computability purposes, and identify the sub-additive cooperative learning bound of the Choquet Integral. We show that the proposed technique targets heretofore unknown variations in predictive weight loss studies with broad potential applications.
more »
« less
- Award ID(s):
- 2140729
- PAR ID:
- 10651527
- Publisher / Repository:
- IEEE
- Date Published:
- Page Range / eLocation ID:
- 202 to 203
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Energy and Accuracy Characterization of a Burst-Mode Range Sensing Approach for Smart Contact LensesIn this paper, we present the characterization of a new range sensing approach for use in emerging smart contact lens applications. Smart contact lenses offer a promising approach to treating the most common form of vision loss by using a tunable lens to accommodate for focal errors. A range sensor is an integral component of the system because it estimates an object's distance from the user in order to determine the target focal length. We performed an empirical study with custom fabricated coils in a mock eyeball setup to understand the energy-accuracy trade-offs of a burst-mode sensing approach based on transmission and reception of square pulses between the coils. We wirelessly transmitted square pulses between the coils and estimated the range of an object by sensing the received voltages and inferring the angular relationship between the two contacts. We demonstrate a functioning range sensing approach that can be implemented with energy as low as 1.8 nJ per measurement with at least 95% accuracy.more » « less
-
Ko, Hanseok (Ed.)Backpropagation (BP) has been a successful optimization technique for deep learning models. However, its limitations, such as backward- and update-locking, and its biological implausibility, hinder the concurrent updating of layers and do not mimic the local learning processes observed in the human brain. To address these issues, recent research has suggested using local error signals to asynchronously train network blocks. However, this approach often involves extensive trial-and-error iterations to determine the best configuration for local training. This includes decisions on how to decouple network blocks and which auxiliary networks to use for each block. In our work, we introduce a novel BP-free approach: a block-wise BP-free (BWBPF) neural network that leverages local error signals to optimize distinct sub-neural networks separately, where the global loss is only responsible for updating the output layer. The local error signals used in the BP-free model can be computed in parallel, enabling a potential speed-up in the weight update process through parallel implementation. Our experimental results consistently show that this approach can identify transferable decoupled architectures for VGG and ResNet variations, outperforming models trained with end-to-end backpropagation and other state-of-the-art block-wise learning techniques on datasets such as CIFAR-10 and Tiny-ImageNet. The code is released at https://github.com/Belis0811/BWBPF.more » « less
-
This study focuses on developing and examining the effectiveness of Transfer Learning (TL) for structural health monitoring (SHM) systems that transfer knowledge about damage states from one structure (i.e., the source domain) to another structure (i.e., the target domain). Transfer Learning (TL) is an efficient method for knowledge transfer and mapping from source to target domains. In addition, Proper Orthogonal Modes (POMs), which help classify behavior and health, provide a promising tool for damage identification in structural systems. Previous investigations show that damage intensity and location are highly correlated with POM variations for structures under unknown loads. To train damage identification algorithms based on POMs and ML, one generally needs to use multiple simulations to generate damage scenarios. The developed process is applied to a simply supported truss span in a multi-span railway bridge. TL is first used to obtain relationships between POMs for two modeled bridges: one being a source model (i.e., labeled) and the other being the target modeled bridge (i.e., unlabeled). This technique is then implemented to develop POMs for a damaged, unknown target using TL that links source and target POMs. It is shown that the trained knowledge from one bridge was effectively generalized to other, somewhat similar, bridges in the population.more » « less
-
Lately, the industry has recognized immense potential in wearables (particularly, smartwatches) being an attractive alternative/supplement to the smartphone. To this end, there has been recent activity in making the smartwatch ‘self-sufficient’ i.e. using it to make/receive calls, etc. independently of the phone. This marked shift in the way wearables will be used in future calls for changes in the core micro- architecture of smartwatch processors. In this work, we first identify ten key target applications for the smartwatch users that the processor must be able to quickly and efficiently execute. We show that seven of these workloads are inherently parallel, and are compute- and data-intensive. We therefore propose to use a multi-core processor with simple out- of-order cores (for compute performance) and augment them with a light-weight software-assisted hardware prefetcher (for memory performance). This simple core with the light-weight prefetcher, called WearCore, is 2.9x more energy-efficient and 2.8x more area- efficient over an in-order core. The improvements are similar with respect to an out-of-order core.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1109/CHASE60773.2024.00039
