http://hdl.handle.net/20.500.14044/33957 Sun, 26 Jul 2026 00:01:24 GMT 2026-07-26T00:01:24Z http://hdl.handle.net/20.500.14044/39245 Training Experimental Language Models with Low Resources, for the Hungarian Language Yang, Zijian Győző; Váradi, Tamás In recent years, natural language processing tasks, like sentiment analysis, can be solved with high performance techniques, if a pre-trained language model is fine-tuned. However, in most cases, the pre-training of language models require huge computational resources and training corpora. Our paper addresses the issue of developing deep neural network language models for low resourced languages, such as Hungarian. Pre-training language models like BERT, requires a prohibitive amount of computational power and huge amount of training data. Unfortunately, neither of these prerequisites are commonly available for low resource languages. The question is how well the system can perform with limited resources (both in data and hardware). We focus our research on five transformer models: ELECTRA, ELECTRIC, RoBERTa, BART and GPT-2. To evaluate our models, we fine-tuned the models in six different natural language processing tasks: sentence-level sentiment analysis, named entity recognition, noun phrase chunking, extractive summarization and abstractive summarization. Our results suggest that while our experimental models obviously cannot surpass the performance of the state-of-the-art Hungarian BERT model, they require a smaller carbon footprint, may bring neural network technology to mobile applications and, finally, they may lower the threshold to engaging with neural network technology in low resourced languages, which has been an obstacle so far, in the synergistic co-development of cognitive info-communication systems and its related disciplines. Sun, 01 Jan 2023 00:00:00 GMT http://hdl.handle.net/20.500.14044/39245 2023-01-01T00:00:00Z http://hdl.handle.net/20.500.14044/39244 Sentiment Analysis of Souls-like Role-Playing Video Game Reviews Guzsvinecz, Tibor The emotions and experiences of players are complex as they can vary due to many factors, such as the difficulty of a video game. However, after finishing video games, most players write reviews of them in which their experiences are described. To understand these better regarding challenging video games, sentiment and textual analyses were conducted on Steam reviews. 993932 reviews were scraped from its website and were imported into R. Using several packages available in R (syuzhet, dplyr, tidytext, ggpubr) and the NRC Emotion Lexicon, the data were investigated. Reviews were grouped into negative and positive ones. According to the results, the following can be concluded regarding player experience in challenging video games: 1) Significant differences exist in case of all emotions and sentiments between the two groups; 2) The emotional valence among both groups significantly differs as well: of negative reviews, it is below zero, and of positive ones is above it; 3) In case of positive reviews, the commonly used words better detail player experiences than in case of negative ones; 4) Regarding the number of votes on and words in both negative and positive reviews, no correlation can be found between them. However, negative reviews are more likely to be agreed with. Sun, 01 Jan 2023 00:00:00 GMT http://hdl.handle.net/20.500.14044/39244 2023-01-01T00:00:00Z http://hdl.handle.net/20.500.14044/39241 Vehicle Automation Impact on Traffic Flow and Stability: A Review of Literature Mohammed, Dilshad; Horváth, Balázs Vehicle automation technologies open new avenues for improving present transportation networks. It is predicted that Autonomous vehicles (AVs) and Connected Vehicles (CVs) will improve the traffic flow though increasing road capacities and reducing travel time and congestion to a great extent. According to the studies, AVs and CVs benefits will likely to more increase with their penetration rates and level of automation and cooperation. Eventually it might accelerate the expected evolution in car mobility in which the gradual transition from regular human driving to automated driving will occur. This paper summarizes the recent studies on the impact of vehicle automation in traffic flow. Although many factors are considered to affect traffic flow under the use of vehicle automation. This study has selected four main factors, these are travel behavior factor, the effect of platooning, travel time factor, and the effect of intersection control. The reviewed studies relevant to these keywords have been extensively argued and thoroughly discussed in this paper. Furthermore, the different applied models and the analytical frameworks which have been used to achieve various results and outcomes, are described in an illustrative table. Finally, all findings which have been demonstrated from this paper emphasize a great contribution of the developed vehicle automation technology to the future of our transportation system. Sun, 01 Jan 2023 00:00:00 GMT http://hdl.handle.net/20.500.14044/39241 2023-01-01T00:00:00Z http://hdl.handle.net/20.500.14044/39240 Sentiment Analysis with Neural Models for Hungarian Laki, László János; Zijian, Győző Yang Sentiment analysis is a powerful tool to gain insight into the emotional polarity of opinionated texts. Computerized applications can contribute to the establishment of next- generation models that can provide us with data of unprecedented quantity and quality. However, these models often require substantial amount of resources in order to meet the desired performance expectations. Therefore, numerous research efforts are targeted to achieve high-quality results while lowering the resource needs by improving the structure and function of the models used. From a cognitive perspective, it is important to understand the mental state of users when they engage in activities that potentially reflect their feelings and emotions. With the emergence of the widespread use of digital solutions, users post opinionated texts on social media, which can be used as a valuable source to detect their underlying sentiments. Therefore, these platforms offer an unparalleled opportunity to perform sentiment analysis. In recent years, natural language processing tasks, like sentiment analysis, can be solved with high performance, if a pre-trained language model is fine-tuned. Herein we present the first neural transformer-based sentiment analysis model for Hungarian, which achieved state-of-the-art performance. Several limitation factors can occur during fine-tuning, such as the lack of training corpora with appropriate size or the complete absence of usable training material. In our experiment, we use data augmentation methods, specifically machine translation and cross-lingual transfer, to increase the size of our training corpora. Here, we demonstrate our experimentation with 9 different language models. Our work provides evidence for the increased efficiency of the trained models if translation text is added to the training corpora. Furthermore, using the augmentation technique, we could further increase the performance of our models. Consequently, our findings represent an important milestone in the advancement of sentence-level and aspect- based sentiment analysis in the Hungarian language. Sun, 01 Jan 2023 00:00:00 GMT http://hdl.handle.net/20.500.14044/39240 2023-01-01T00:00:00Z