An official website of the United States government
Abstract Machine learning influences numerous aspects of modern society, empowers new technologies, from Alphago to ChatGPT, and increasingly materializes in consumer products such as smartphones and self-driving cars. Despite the vital role and broad applications of artificial neural networks, we lack systematic approaches, such as network science, to understand their underlying mechanism. The difficulty is rooted in many possible model configurations, each with different hyper-parameters and weighted architectures determined by noisy data. We bridge the gap by developing a mathematical framework that maps the neural network’s performance to the network characters of the line graph governed by the edge dynamics of stochastic gradient descent differential equations. This framework enables us to derive a neural capacitance metric to universally capture a model’s generalization capability on a downstream task and predict model performance using only early training results. The numerical results on 17 pre-trained ImageNet models across five benchmark datasets and one NAS benchmark indicate that our neural capacitance metric is a powerful indicator for model selection based only on early training results and is more efficient than state-of-the-art methods.
more »
« less
This content will become publicly available on May 21, 2026
Designing soft and tough multiple-network elastomers: impact of reversible radical deactivation on filler network architecture and fracture toughness
Reversible radical deactivation during gelation narrows the chain length distribution of polymer networks, compromising stress delocalization, energy dissipation, and fracture toughness.
more »
« less
- Award ID(s):
- 2225167
- PAR ID:
- 10614031
- Publisher / Repository:
- RSC
- Date Published:
- Journal Name:
- Soft Matter
- Volume:
- 21
- Issue:
- 20
- ISSN:
- 1744-683X
- Page Range / eLocation ID:
- 4029 to 4042
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract SummaryNetwork biology is an interdisciplinary field bridging computational and biological sciences that has proved pivotal in advancing the understanding of cellular functions and diseases across biological systems and scales. Although the field has been around for two decades, it remains nascent. It has witnessed rapid evolution, accompanied by emerging challenges. These stem from various factors, notably the growing complexity and volume of data together with the increased diversity of data types describing different tiers of biological organization. We discuss prevailing research directions in network biology, focusing on molecular/cellular networks but also on other biological network types such as biomedical knowledge graphs, patient similarity networks, brain networks, and social/contact networks relevant to disease spread. In more detail, we highlight areas of inference and comparison of biological networks, multimodal data integration and heterogeneous networks, higher-order network analysis, machine learning on networks, and network-based personalized medicine. Following the overview of recent breakthroughs across these five areas, we offer a perspective on future directions of network biology. Additionally, we discuss scientific communities, educational initiatives, and the importance of fostering diversity within the field. This article establishes a roadmap for an immediate and long-term vision for network biology. Availability and implementationNot applicable.more » « less
-
Abstract Double-network (DN) hydrogels, consisting of two contrasting and interpenetrating polymer networks, are considered as perhaps the toughest soft-wet materials. Current knowledge of DN gels from synthesis methods to toughening mechanisms almost exclusively comes from chemically-linked DN hydrogels by experiments. Molecular modeling and simulations of inhomogeneous DN structure in hydrogels have proved to be extremely challenging. Herein, we developed a new multiscale simulation platform to computationally investigate the early fracture of physically-chemically linked agar/polyacrylamide (agar/PAM) DN hydrogels at a long timescale. A “random walk reactive polymerization” (RWRP) was developed to mimic a radical polymerization process, which enables to construct a physically-chemically linked agar/PAM DN hydrogel from monomers, while conventional and steered MD simulations were conducted to examine the structural-dependent energy dissipation and fracture behaviors at the relax and deformation states. Collective simulation results revealed that energy dissipation of agar/PAM hydrogels was attributed to a combination of the pulling out of agar chains from the DNs, the disruption of massive hydrogen bonds between and within DN structures, and the strong association of water molecules with both networks, thus explaining a different mechanical enhancement of agar/PAM hydrogels. This computational work provided atomic details of network structure, dynamics, solvation, and interactions of a hybrid DN hydrogel, and a different structural-dependent energy dissipation mode and fracture behavior of a hybrid DN hydrogel, which help to design tough hydrogels with new network structures and efficient energy dissipation modes. Additionally, the RWRP algorithm can be generally applied to construct the radical polymerization-produced hydrogels, elastomers, and polymers.more » « less
-
In the small town of Sitka, Alaska, frequent and often catastrophic landslides threaten residents. One challenge associated with disaster preparedness is access to timely and reliable risk information. As with many small but diverse towns, who or what is a trustworthy source of information is often contested. To help improve landslide communication in Sitka, we used a community-partnered approach to social network analysis to identify (1) potential key actors for landslide risk communication and (2) structural holes that may inhibit efficient and equitable communication. This short take describes how we built trust and developed adaptive data collection methods to build an approach that was acceptable and actionable for Sitka, Alaska. This approach could be useful to other researchers for conducting social network analysis to improve risk communication, particularly in rural and remote contexts.more » « less
-
Abstract One of the goals of open science is to promote the transparency and accessibility of research. Sharing data and materials used in network research is critical to these goals. In this paper, we present recommendations for whether, what, when, and where network data and materials should be shared. We recommend that network data and materials should be shared, but access to or use of shared data and materials may be restricted if necessary to avoid harm or comply with regulations. Researchers should share the network data and materials necessary to reproduce reported results via a publicly accessible repository when an associated manuscript is published. To ensure the adoption of these recommendations, network journals should require sharing, and network associations and academic institutions should reward sharing.more » « less
