Commit Journal

Analyzing the Effects of Combining Gradient Conflict Mitigation Methods in Multi-Task Learning

Abstract
Multi-task machine learning approaches involve training a single model on multiple tasks at once to increase performance and efficiency over multiple singletask models trained individually on each task. When such a multi-task model is trained to perform multiple unrelated tasks, performance can degrade significantly since unrelated tasks often have gradients that vary widely in direction. These conflicting gradients may destructively interfere with each other, causing weights learned during the training of some tasks to become unlearned during the training of others. The research selects three existing methods to mitigate this problem: Project Conflicting Gradients (PCGrad), Modulation Module, and Language-Specific Subnetworks (LaSS). It explores how the application of different combinations of these methods affects the performance of a convolutional neural network on a multi-task image classification problem. The image classification problem used as a benchmark utilizes a dataset of 4,503 leaf images to create two separate tasks: the classification of plants and the detection of disease from leaf images. Experiment results on this problem show performance benefits over singular mitigation methods, with a combination of PCGrad and LaSS obtaining a task-averaged F1 score of 0.84686. This combination outperforms individual mitigation approaches by 0.01870, 0.02682, and 0.02434 for PCGrad, Modulation Module, and LaSS, respectively in terms of F1 score.

Keywords: Gradient Conflict Mitigation Methods, Multi-Task Learning, Project Conflicting Gradients (PCGrad), Modulation Module, Language-Specific Subnetworks (LaSS)

Read full article: https://journal.binus.ac.id/index.php/commit/article/view/8905

Visit our website: https://journal.binus.ac.id/index.php/commit/index

Insights into Mobile Government Adoption Factors: A Comprehensive Analysis of Peduli Lindungi Application in Indonesia

Abstract
Information and Communication Technology (ICT) progression has notably impacted the shift from traditional public services to digital alternatives. Among the various digital services, m-government services, provided by smartphone technology, have gained widespread popularity. Unfortunately, the broader adoption of digital technology encounters several challenges, including insufficient user interest and acceptance, as well as concerns regarding security and user privacy. The primary goal of the research is to address the existing gap in the literature by examining the factors that contribute to the effective implementation of m-government services. A mix of key components is employed, incorporating the Information Systems (IS) Success model and Technology Acceptance Model (TAM) as research variables. The research applies a quantitative approach in the form of an online survey. Furthermore, a Partial Least Square- Structure Equational Modeling (PLS-SEM) analytic approach is performed to evaluate 230 data points. The research findings support five hypotheses while rejecting three hypotheses. Significantly, the findings suggest that perceived usefulness and ease of use influence behavioral intention considerably. Additionally, constructions related to service quality significantly impact behavioral intention. Meanwhile, both system quality and information quality do not contribute to affecting behavioral intention. Furthermore, information quality exerts a substantial impact on perceived usefulness, but it does not influence perceived ease of use. Finally, it is observed that system quality significantly affects the perceived ease of use.

Keywords: M-Government Service, Public Health Service, Peduli Lindungi Application, User Acceptance

Read full article: https://journal.binus.ac.id/index.php/commit/article/view/9024
Visit our website: https://journal.binus.ac.id/index.php/commit/index