Leveraging WiFi Network Logs to Infer Student Collocation and its Relationship with Academic Performance
Authors:
V. Das Swain,
H. Kwon,
S. Sargolzaei,
B. Saket,
M. Bin Morshed,
K. Tran,
D. Patel,
Y. Tian,
J. Philipose,
Y. Cui,
T. Plötz,
M. De Choudhury,
G. D. Abowd
Abstract:
A comprehensive understanding of collocation can help understand performance outcomes. For university cohorts, this needs data that describes large groups over a long period. Harnessing user devices to infer this, while tempting, is challenged by privacy concerns, power consumption, and maintenance issues. Alternatively, embedding new sensors in the environment is limited by the expense of coverin…
▽ More
A comprehensive understanding of collocation can help understand performance outcomes. For university cohorts, this needs data that describes large groups over a long period. Harnessing user devices to infer this, while tempting, is challenged by privacy concerns, power consumption, and maintenance issues. Alternatively, embedding new sensors in the environment is limited by the expense of covering the entire campus. We investigate the feasibility of leveraging WiFi association logs for this purpose. While these provide coarse approximations of location, these are easily obtainable and depict multiple users on campus over a semester. We explore how these coarse collocations are related to individual performance. Specifically, we inspect the association between individual performance and the collocation behaviors of project group members. We study 163 students (in 54 project groups) over 14 weeks. After describing how we determine collocation with the WiFi logs, we present a study to analyze how collocation within groups relates to a student's final score. We find collocation behaviors show a significant correlation (Pearson's r = 0.24) with performance -- better than both peer feedback or individual behaviors like attendance. Finally, we discuss how repurposing WiFi logs can facilitate applications for domains like mental wellbeing and physical health.
△ Less
Submitted 5 May, 2021; v1 submitted 22 May, 2020;
originally announced May 2020.
Entropy-Based Modeling for Estimating Soft Errors Impact on Binarized Neural Network Inference
Authors:
Navid Khoshavi,
Saman Sargolzaei,
Arman Roohi,
Connor Broyles,
Yu Bi
Abstract:
Over past years, the easy accessibility to the large scale datasets has significantly shifted the paradigm for develo** highly accurate prediction models that are driven from Neural Network (NN). These models can be potentially impacted by the radiation-induced transient faults that might lead to the gradual downgrade of the long-running expected NN inference accelerator. The crucial observation…
▽ More
Over past years, the easy accessibility to the large scale datasets has significantly shifted the paradigm for develo** highly accurate prediction models that are driven from Neural Network (NN). These models can be potentially impacted by the radiation-induced transient faults that might lead to the gradual downgrade of the long-running expected NN inference accelerator. The crucial observation from our rigorous vulnerability assessment on the NN inference accelerator demonstrates that the weights and activation functions are unevenly susceptible to both single-event upset (SEU) and multi-bit upset (MBU), especially in the first five layers of our selected convolution neural network. In this paper, we present the relatively-accurate statistical models to delineate the impact of both undertaken SEU and MBU across layers and per each layer of the selected NN. These models can be used for evaluating the error-resiliency magnitude of NN topology before adopting them in the safety-critical applications.
△ Less
Submitted 21 April, 2020; v1 submitted 10 April, 2020;
originally announced April 2020.
