The Graduate Research Identity Development program (GRID) is an initiative in the College of Engineering at North Carolina A&T State University, sponsored by the National Science Foundation since 2019. The program offers seminar-type lectures supplemented with activities designed to help graduate students develop critical skills for research-based careers. The program is focused on graduate engineering students but is open to graduate students from all programs. Students also choose mentors from within and outside the university with the goal of increasing their sense of belonging to the field and their identities as research engineers. As part of this program, a pilot study is in progress, aimed at performing a full-scale network analysis of student interactions. A web-based survey was administered to collect information about students in and outside the College of Engineering who participate in the GRID program sessions. The survey was designed to collect information on the relationship networks (or lack thereof) that students are involved in as they matriculate through their graduate program. It assesses things such as how and where the students interact with one another, members of faculty and staff, and with contacts from intramural and extramural organizations. Several items are also used to assess students’ perceptions of themselves as research engineers. In this paper, we focus on the interactions of students in the classroom. More specifically, we form networks based on the student answers about the classes they have taken in different departments. We then analyze the resultant networks and contrast certain graph theoretic properties to students’ scores on the research engineer identity items. Do students that are in the periphery, or students that have more connections attain higher research engineer identity scores? Do students that form complete subnetworks (cliques) or core-periphery structures (induced stars) have higher scores than others? This paper presents the findings from this pilot study from the network analysis on this cohort of students. In summary, we find that students with high eigenvector centrality scores and those who form larger cliques possess significantly higher research engineer identity scores.
more »
« less
Joint training of a predictor network and a generative adversarial network for time series forecasting: A case study of bearing prognostics
- Award ID(s):
- 1919265
- PAR ID:
- 10339290
- Date Published:
- Journal Name:
- Expert Systems with Applications
- Volume:
- 203
- Issue:
- C
- ISSN:
- 0957-4174
- Page Range / eLocation ID:
- 117415
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
null (Ed.)We demonstrate digital predistortion (DPD) using a novel, neural-network (NN) method to combat the nonlinearities in power amplifiers (PAs), which limit the power efficiency of mobile devices, increase the error vector magnitude, and cause inadequate spectral containment. DPD is commonly done with polynomial-based methods that use an indirect-learning architecture (ILA) which can be computationally intensive, especially for mobile devices, and overly sensitive to noise. Our approach using NNs avoids the problems associated with ILAs by first training a NN to model the PA then training a predistorter by backpropagating through the PA NN model. The NN DPD effectively learns the unique PA distortions, which may not easily fit a polynomial-based model, and hence may offer a favorable tradeoff between computation overhead and DPD performance. We demonstrate the performance of our NN method using two different power amplifier systems and investigate the complexity tradeoffs.more » « less
-
Knuth, Kevin; Drozdz, Stanislaw (Ed.)Mapping network nodes and edges to communities and network functions is crucial to gaining a higher level of understanding of the network structure and functions. Such mappings are particularly challenging to design for covert social networks, which intentionally hide their structure and functions to protect important members from attacks or arrests. Here, we focus on correctly inferring the structures and functions of such networks, but our methodology can be broadly applied. Without the ground truth, knowledge about the allocation of nodes to communities and network functions, no single network based on the noisy data can represent all plausible communities and functions of the true underlying network. To address this limitation, we apply a generative model that randomly distorts the original network based on the noisy data, generating a pool of statistically equivalent networks. Each unique generated network is recorded, while each duplicate of the already recorded network just increases the repetition count of that network. We treat each such network as a variant of the ground truth with the probability of arising in the real world approximated by the ratio of the count of this network’s duplicates plus one to the total number of all generated networks. Communities of variants with frequently occurring duplicates contain persistent patterns shared by their structures. Using Shannon entropy, we can find a variant that minimizes the uncertainty for operations planned on the network. Repeatedly generating new pools of networks from the best network of the previous step for several steps lowers the entropy of the best new variant. If the entropy is too high, the network operators can identify nodes, the monitoring of which can achieve the most significant reduction in entropy. Finally, we also present a heuristic for constructing a new variant, which is not randomly generated but has the lowest expected cost of operating on the distorted mappings of network nodes to communities and functions caused by noisy data.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