As an emerging interpretable technique, Generalized Additive Models (GAMs) adopt neural networks to individually learn non-linear functions for each feature, which are then combined through a linear model for final predictions. Although GAMs can explain deep neural networks (DNNs) at the feature level, they require large numbers of model parameters and are prone to overfitting, making them hard to train and scale. Additionally, in real-world datasets with many features, the interpretability of feature-based explanations diminishes for humans. To tackle these issues, recent research has shifted towards concept-based interpretable methods. These approaches try to integrate concept learning as an intermediate step before making predictions, explaining the predictions in terms of human-understandable concepts. However, these methods require domain experts to extensively label concepts with relevant names and their ground-truth values. In response, we propose CAT, a novel interpretable Concept-bAsed Taylor additive model to simplify this process. CAT does not require domain experts to annotate concepts and their ground-truth values. Instead, it only requires users to simply categorize input features into broad groups, which can be easily accomplished through a quick metadata review. Specifically, CAT first embeds each group of input features into one-dimensional high-level concept representation, and then feeds the concept representations into a new white-box Taylor Neural Network (TaylorNet). The TaylorNet aims to learn the non-linear relationship between the inputs and outputs using polynomials. Evaluation results across multiple benchmarks demonstrate that CAT can outperform or compete with the baselines while reducing the need of extensive model parameters. Importantly, it can effectively explain model predictions through high-level concepts. Source code is available at github.com/vduong143/CAT-KDD-2024.
23 September 2025 in Pervasive and Mobile Computing
Distance education is the most important part of modern education ideas, and the experiment manipulation tache is the key factor which restricts the more perfect of distance education mode. The paper analyses the application feasibility of virtual laboratory based on Web in distance education firstly, then makes a simply introduction of virtual laboratory, and describes the system structure and function flow of virtual laboratory based on Web in succession, and expatiates the realization of technologies of virtual laboratory, in the end it analyses the virtual laboratory's application in distance education.
This paper presents a dynamic analysis of an energy efficiency dehumidifier for building applications. This new structured dehumidifier is working with a controllable desiccant concentration, and the dynamic behavior of the heat and mass transfer process inside the dehumidifier becomes rather critical. A model is developed to describe the dynamic behavior of the process air with experimental validation. The numerical results of the dynamic response are clarified to conclude that decreasing the solution inlet temperature has a more significant impact on improving the dehumidification ability than increasing solution mass flow rate. The outcomes provide a comprehensive understanding of the dynamic behavior and the results will benefit the humidity control of the indoor air to improve the living quality of the occupants.
14 June 2022 in Concurrency and Computation: Practice and Experience