Janos Pal Vadasz

Dr. Janos Pal Vadasz

National University of Public Service, Budapest, Hungary
orcid

Research Focus & Metrics

Academic Career Length

11Years (2016-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
13 August 2021 in Symmetry
Verified
Journal Article
Open access
PDF file
GCGergely Csányi
Gergely Csányi
DNDaniel Nagy
Daniel Nagy
RVRenátó Vági
Renátó Vági
JVJanos Pal Vadasz
Janos Pal Vadasz
Tamás OroszTamás Orosz
Tamás Orosz
Data sharing is a central aspect of judicial systems. The openly accessible documents can make the judiciary system more transparent. On the other hand, the published legal documents can contain much sensitive information about the involved persons or companies. For this reason, the anonymization of these documents is obligatory to prevent privacy breaches. General Data Protection Regulation (GDPR) and other modern privacy-protecting regulations have strict definitions of private data containing direct and indirect identifiers. In legal documents, there is a wide range of attributes regarding the involved parties. Moreover, legal documents can contain additional information about the relations between the involved parties and rare events. Hence, the personal data can be represented by a sparse matrix of these attributes. The application of Named Entity Recognition methods is essential for a fair anonymization process but is not enough. Machine learning-based methods should be used together with anonymization models, such as differential privacy, to reduce re-identification risk. On the other hand, the information content (utility) of the text should be preserved. This paper aims to summarize and highlight the open and symmetrical problems from the fields of structured and unstructured text anonymization. The possible methods for anonymizing legal documents discussed and illustrated by case studies from the Hungarian legal practice.

29 December 2021 in Applied Sciences
Verified
Journal Article
Open access
PDF file
Tamás OroszTamás Orosz
Tamás Orosz
RVRenátó Vági
Renátó Vági
GCGergely Csányi
Gergely Csányi
DNDaniel Nagy
Daniel Nagy
IÜIstván Üveges
István Üveges
AMAndrea Megyeri
Andrea Megyeri
Many machine learning-based document processing applications have been published in recent years. Applying these methodologies can reduce the cost of labor-intensive tasks and induce changes in the company’s structure. The artificial intelligence-based application can replace the application of trainees and free up the time of experts, which can increase innovation inside the company by letting them be involved in tasks with greater added value. However, the development cost of these methodologies can be high, and usually, it is not a straightforward task. This paper presents a survey result, where a machine learning-based legal text labeler competed with multiple people with different legal domain knowledge. The machine learning-based application used binary SVM-based classifiers to resolve the multi-label classification problem. The used methods were encapsulated and deployed as a digital twin into a production environment. The results show that machine learning algorithms can be effectively utilized for monotonous but domain knowledge- and attention-demanding tasks. The results also suggest that embracing the machine learning-based solution can increase discoverability and enrich the value of data. The test confirmed that the accuracy of a machine learning-based system matches up with the long-term accuracy of legal experts, which makes it applicable to automatize the working process.

29 January 2022 in Applied Sciences
Verified
Journal Article
Open access
PDF file
GCGergely Csányi
Gergely Csányi
RVRenátó Vági
Renátó Vági
DNDaniel Nagy
Daniel Nagy
IÜIstván Üveges
István Üveges
JVJanos Pal Vadasz
Janos Pal Vadasz
Tamás OroszTamás Orosz
Tamás Orosz
One of the most time-consuming parts of an attorney’s job is finding similar legal cases. Categorization of legal documents by their subject matter can significantly increase the discoverability of digitalized court decisions. This is a multi-label classification problem, where each relatively long text can fit into more than one legal category. The proposed paper shows a solution where this multi-label classification problem is decomposed into more than a hundred binary classification problems. Several approaches have been tested, including different machine-learning and text-augmentation techniques to produce a practically applicable model. The proposed models and the methodologies were encapsulated and deployed as a digital-twin into a production environment. The performance of the created machine learning-based application reaches and could also improve the human-experts performance on this monotonous and labor-intensive task. It could increase the e-discoverability of the documents by about 50%.

10 December 2024 in Big Data and Cognitive Computing
Verified
Journal Article
Open access
PDF file
GCGergely Csányi
Gergely Csányi
DLDorina Petra Lakatos
Dorina Petra Lakatos
IÜIstván Üveges
István Üveges
AMAndrea Megyeri
Andrea Megyeri
JVJanos Pal Vadasz
Janos Pal Vadasz
RVRenátó Vági
Renátó Vági
This research paper presents findings from an investigation in the semantic similarity search task within the legal domain, using a corpus of 1172 Hungarian court decisions. The study establishes the groundwork for an operational semantic similarity search system designed to identify cases with comparable facts using preliminary legal fact drafts. Evaluating such systems often poses significant challenges, given the need for thorough document checks, which can be costly and limit evaluation reusability. To address this, the study employs manually created fact drafts for legal cases, enabling reliable ranking of original cases within retrieved documents and quantitative comparison of various vectorization methods. The study compares twelve different text embedding solutions (the most recent became available just a few weeks before the manuscript was written) identifying Cohere’s embed-multilingual-v3.0, Beijing Academy of Artificial Intelligence’s bge-m3, Jina AI’s jina-embeddings-v3, OpenAI’s text-embedding-3-large, and Microsoft’s multilingual-e5-large models as top performers. To overcome the transformer-based models’ context window limitation, we investigated chunking, striding, and last chunk scaling techniques, with last chunk scaling significantly improving embedding quality. The results suggest that the effectiveness of striding varies based on token count. Notably, employing striding with 16 tokens yielded optimal results, representing 3.125% of the context window size for the best-performing models. Results also suggested that from the models having 8192 token long context window the bge-m3 model is superior compared to jina-embeddings-v3 and text-embedding-3-large models in capturing the relevant parts of a document if the text contains significant amount of noise. The validity of the approach was evaluated and confirmed by legal experts. These insights led to an operational semantic search system for a prominent legal content provider.

23 September 2023 in Information
Verified
Journal Article
Open access
PDF file
GCGergely Csányi
Gergely Csányi
RVRenátó Vági
Renátó Vági
AMAndrea Megyeri
Andrea Megyeri
AFAnna Fülöp
Anna Fülöp
DNDaniel Nagy
Daniel Nagy
IÜIstván Üveges
István Üveges
Few-shot learning is a deep learning subfield that is the focus of research nowadays. This paper addresses the research question of whether a triplet-trained Siamese network, initially designed for multi-class classification, can effectively handle multi-label classification. We conducted a case study to identify any limitations in its application. The experiments were conducted on a dataset containing Hungarian legal decisions of administrative agencies in tax matters belonging to a major legal content provider. We also tested how different Siamese embeddings compare on classifying a previously non-existing label on a binary and a multi-label setting. We found that triplet-trained Siamese networks can be applied to perform classification but with a sampling restriction during training. We also found that the overlap between labels affects the results negatively. The few-shot model, seeing only ten examples for each label, provided competitive results compared to models trained on tens of thousands of court decisions using tf-idf vectorization and logistic regression.

Verified
Journal Article
Open access
PDF file
GCGergely Csányi
Gergely Csányi
IÜIstván Üveges
István Üveges
DLDorina Petra Lakatos
Dorina Petra Lakatos
DRDóra Ripszám
Dóra Ripszám
KKKornélia Kozák
Kornélia Kozák
JVJanos Pal Vadasz
Janos Pal Vadasz
This paper presents an in-production Rhetorical Role Labeling (RRL) classifier developed for Hungarian judicial decisions. RRL is a sequential classification problem in Natural Language Processing, aiming to assign functional roles (such as facts, arguments, decision, etc.) to every segment or sentence in a legal document. The study was conducted on a human-annotated sentence-level RRL corpus and compares multiple neural architectures, including BiLSTM, attention-based networks, and a support vector machine as baseline. It further investigates the impact of late chunking during vectorization, in contrast to classical approaches. Results from tests on the labeled dataset and annotator agreement statistics are reported, and performance is analyzed across architecture types and embedding strategies. Contrary to recent findings in retrieval tasks, late chunking does not show consistent improvements for sentence-level RRL, suggesting that contextualization through chunk embeddings may introduce noise rather than useful context in Hungarian legal judgments. The work also discusses the unique structure and labeling challenges of Hungarian cases compared to international datasets and provides empirical insights for future legal NLP research in non-English court decisions.