RPRodica Potolea

Rodica Potolea

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Rodica Potolea has a total of 35 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Understanding Cultural Alignment in Multilingual LLMs via Natural Debate Statements

    by Vlad-Andrei Negru, Camelia Lemnaru, Mihai Surdeanu, Rodica Potolea

    In this work we investigate the sociocultural values learned by large language models (LLMs). We introduce a novel open-access dataset, Sociocultural Statements, constructed from natural debate statements using a multi-step methodology. The dataset is synthetically labeled to enable the quantization of sociocultural norms and beliefs that LLMs exhibit in their responses to these statements, according to the Hofstede cultural dimensions. We verify the accuracy of synthetic labels using human quality control on a representative sample. We conduct a comparative analysis between two groups of LLMs developed in different countries (U.S. and China), and use as a comparative baseline patterns observed in human measurements. Using this new dataset and the analysis above, we found that culturally-distinct LLMs reflect the values and norms of the countries in which they were developed, highlighting their inability to adapt to the sociocultural backgrounds of their users.

  • Open Access

    MorphNLI: A Stepwise Approach to Natural Language Inference Using Text Morphing

    by Vlad Andrei Negru, Robert Vacareanu, Camelia Lemnaru, Mihai Surdeanu, Rodica Potolea

    We introduce MorphNLI, a modular step-by-step approach to natural language inference (NLI). When classifying the premise-hypothesis pairs into {entailment, contradiction, neutral}, we use a language model to generate the necessary edits to incrementally transform (i.e., morph) the premise into the hypothesis. Then, using an off-the-shelf NLI model we track how the entailment progresses with these atomic changes, aggregating these intermediate labels into a final output. We demonstrate the advantages of our proposed method particularly in realistic cross-domain settings, where our method always outperforms strong baselines with improvements up to 12.6% (relative). Further, our proposed approach is explainable as the atomic edits can be used to understand the overall NLI label.

  • Open Access

    MorphNLI: A Stepwise Approach to Natural Language Inference Using Text Morphing

    by Vlad Andrei Negru, Robert Vacareanu, Camelia Lemnaru, Mihai Surdeanu, Rodica Potolea

    Vlad Andrei Negru, Robert Vacareanu, Camelia Lemnaru, Mihai Surdeanu, Rodica Potolea. Findings of the Association for Computational Linguistics: NAACL 2025. 2025.


  • Open Access

    A path-based distance computation for non-convexity with applications in clustering

    by Eugen-Richard Ardelean, Raluca Laura Portase, Rodica Potolea, Mihaela Dînșoreanu

    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.


  • Open Access

    From Individual Device Usage to Household Energy Consumption Profiling

    by Ramona Tolas, Raluca Portase, Rodica Potolea

    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.


  • Open Access

    From Sensors to Insights: An Original Method for Consumer Behavior Identification in Appliance Usage

    by Raluca Laura Portase, Ramona Tolas, Rodica Potolea

    In light of the energy crisis, extensive research is being conducted to enhance load forecasting, optimize the targeting of demand response programs, and advise building occupants on actions to enhance energy performance. Cluster analysis is increasingly applied to usage data across all consumer types. More accurate consumer identification translates to improved resource planning. In the context of Industry 4.0, where comprehensive data are collected across various domains, we propose using existing sensor data from household appliances to extract the usage patterns and characterize the resource demands of consumers from residential households. We propose a general pipeline for extracting features from raw sensor data alongside global features for clustering device usages and classifying them based on extracted time series. We applied the proposed method to real data from three different types of household devices. We propose a strategy to identify the number of existent clusters in real data. We employed the label data obtained from clustering for the classification of consumers based on data recorded on different time ranges and achieved an increase in accuracy of up to 15% when we expanded the time range for the recorded data on the entire dataset, obtaining an accuracy of over 99.89%. We further explore the data meta-features for a minimal dataset by examining the necessary time interval for the recorded data, dataset dimensions, and the feature set. This analysis aims to achieve an effective trade-off between time and performance.


  • Open Access

    SmartLaundry: A Real-Time System for Public Laundry Allocation in Smart Cities

    by Raluca Laura Portase, Ramona Tolas, Rodica Potolea

    Smart cities facilitate the comprehensive management and operation of urban data generated within a city, establishing the foundation for smart services and addressing diverse urban challenges. A smart system for public laundry management uses artificial intelligence-based solutions to solve the challenges of the inefficient utilization of public laundries, waiting times, overbooking or underutilization of machines, balancing of loads across machines, and implementation of energy-saving features. We propose SmartLaundry, a real-time system design for public laundry smart recommendations to better manage the loads across connected machines. Our system integrates the current status of the connected devices and data-driven forecasted usage to offer the end user connected via a mobile application a list of recommended machines that could be used. We forecast the daily usage of devices using traditional machine learning techniques and deep learning approaches, and we perform a comparative analysis of the results. As a proof of concept, we create a simulation of the interaction with our system.


  • Open Access

    Random forest evaluation using multi-key homomorphic encryption and lookup tables

    by Diana-Elena Petrean, Rodica Potolea

    In recent years, machine learning (ML) has become increasingly popular in various fields of activity. Cloud platforms have also grown in popularity, as they offer services that are more secure and accessible worldwide. In this context, cloud-based technologies emerged to support ML, giving rise to the machine learning as a service (MLaaS) concept. However, the clients accessing ML services in order to obtain classification results on private data may be reluctant to upload sensitive information to cloud. The model owners may also prefer not to outsource their models in order to prevent model inversion attacks and to protect intellectual property. The privacy-preserving evaluation of ML models is possible through multi-key homomorphic encryption (MKHE), that allows both the client data and the model to be encrypted under different keys. In this paper, we propose an MKHE evaluation method for decision trees and we extend the proposed method for random forests. Each decision tree is evaluated as a single lookup table, and voting is performed at the level of groups of decision trees in the random forest. We provide both theoretical and experimental evaluations for the proposed method. The aim is to minimize the performance degradation introduced by the encrypted model compared to a plaintext model while also obtaining practical classification times. In our experiments with the proposed MKHE random forest evaluation method, we obtained minimal (less than 0.6%) impact on the main ML performance metrics considered for each scenario, while also achieving reasonable classification times (of the order of seconds).