As virtual home assistants are becoming more popular, there is an emerging need for supporting languages other than English. While more wide-spread or popular languages such as Spanish, French or Hindi are already integrated into existing home assistants like Google Home or Alexa, integration of other less-known languages such as Romanian is still missing. This paper explores the problem of Natural Language Understanding (NLU) applied to a Romanian home assistant. We propose a customized capsule neural network architecture that performs intent detection and slot filling in a joint manner and we evaluate how well it handles utterances containing various levels of complexity. The capsule network model shows a significant improvement in intent detection when compared to models built using the well-known Rasa NLU tool. Through error analysis, we observe clear error patterns that occur systematically. Variability in language when expressing one intent proves to be the biggest challenge encountered by the model.
1 November 2024 in Knowledge and Information Systems
Clustering algorithms are essential in data analysis, but evaluating their performance is challenging when the true labels are not available, especially for non-convex clusters. Traditional performance evaluation metrics struggle to identify clustering quality, often assigning higher scores for linearly separated clusters than the true clusters. We propose an original approach to distance computation that accounts for the data structure, thus improving the clustering quality evaluation for non-convex clusters without affecting other shapes of clusters. We also showcase the applicability of this method through a modified version of K-Means using the proposed method that is capable of correctly separating non-convex clusters. The validation included the analysis of performance and time complexity of 3 traditional clustering quality evaluation metrics and the K-Means clustering algorithm against their augmented versions with the proposed approach. This analysis conducted on 7 benchmark synthetic datasets and 6 real datasets with various numbers of examples and features of diverse characteristics and joint complexities: simple convex clusters, overlapped and imbalanced clusters, and non-convex clusters. Through these analyses, we show the ineffectiveness of traditional methods and that the proposed approach overcomes the weaknesses of traditional methods.
This paper presents a novel approach for electricity consumption profiling in households through the fusion of usage data for individual smart devices. The novelty of the approach consists of leveraging the data representing the usage of individual appliances rather than using direct measurements of energy consumption. Our methodology focuses on merging signals representing the interaction of the user with the device to compute patterns in the total energy consumption per household. Subsequently, we apply data mining techniques—specifically, unsupervised clustering—to analyze the resulting time-series data representing daily energy consumption. Through this approach, we aim to identify and characterize patterns in energy usage within households, enabling insights for energy optimization strategies and resource allocation. This information can be further used in practical tasks, such as flattening energy consumption. The proposed approach offers an alternative to the direct measurement of energy usage, considering the potential for sensor failure or malfunction. This underscores the importance of implementing a complementary method for verifying and validating energy consumption data.
In the era of data-driven technologies, the need for diverse and high-quality datasets for training and testing machine learning models has become increasingly critical. In this article, we present a versatile methodology, the Generic Methodology for Constructing Synthetic Data Generation (GeMSyD), which addresses the challenge of synthetic data creation in the context of smart devices. GeMSyD provides a framework that enables the generation of synthetic datasets, aligning them closely with real-world data. To demonstrate the utility of GeMSyD, we instantiate the methodology by constructing a synthetic data generation framework tailored to the domain of event-based data modeling, specifically focusing on user interactions with smart devices. Our framework leverages GeMSyD to create synthetic datasets that faithfully emulate the dynamics of human–device interactions, including the temporal dependencies. Furthermore, we showcase how the synthetic data generated using our framework can serve as a valuable resource for machine learning practitioners. By employing these synthetic datasets, we perform a series of experiments to evaluate the performance of a neural-network-based prediction model in the domain of smart device interaction. Our results underscore the potential of synthetic data in facilitating model development and benchmarking.
Vehicular traffic in urban areas faces congestion challenges that negatively impact our lives. The infrastructure associated with intelligent transportation systems provides means for addressing the associated challenges in urban areas. This study proposes an effective and scalable vehicular traffic congestion avoidance methodology. It introduces a traffic thresholding mechanism to predict and avoid vehicular traffic congestion during route computation. Our methodology was evaluated and validated by employing four road network topologies, three vehicular traffic density levels and various traffic light configurations, resulting in 26 urban traffic scenarios. Using our approach, the number of vehicles that can run in free flow can be increased by up to 200%, whereas for traffic congestion scenarios, the time spent in traffic may be reduced by up to 69% and CO2 emissions by up to 61%. To the best of our knowledge, in the vehicular traffic flow prediction domain, this is the first approach that covers a set of road network topologies and a large and representative set of scenarios for simulated urban traffic congestion testing. Moreover, the comparative analysis with different other solutions in the domain, showed that we obtained the best driving time and CO2 emission reduction.