MSMiklós Szócska

Miklós Szócska

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Miklós Szócska has a total of 60 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Health Former: Dual-Level Time-Aware Transformers for Irregular Electronic Health Record Events

    by Péter Kőrösi Szabó, Gábor Kovács, Botond Forrai, Judit Laki, Tamás Kováts, Miklós Szócska, Adrián Csiszárik

    Background and Objectives: Longitudinal electronic health records (EHRs) form irregular event sequences that mix multiple clinical coding systems and care settings. Learning transferable patient representations requires modeling both within-encounter code composition and long-range temporal dependencies. We aim to develop a pretraining framework that preserves event structure and explicitly uses elapsed time, while remaining straightforward to fine-tune for new supervised endpoints without task-specific feature engineering. Methods: We propose HealthFormer, a dual-level Transformer for event-centric EHR modeling. An Intra-Event Encoder aggregates heterogeneous domain tokens within each typed clinical event into an event embedding via code-specific embedding modules and attention pooling. Event embeddings are combined with a Date Encoder and a continuous-time attention bias based on attention with linear biases (ALiBI) inside an Inter-Event Encoder. We pretrain on Hungarian national administrative health records from a large-scale nationwide longitudinal cohort (spanning millions of individuals over a decade) using multi-task self-supervision with (i) per-domain masked token prediction (masked language modeling, MLM), (ii) event-type prediction under full-event masking (Event-level MLM), (iii) next-event type prediction, and (iv) time-to-next-event (Δt) regression.Results: Pretraining induces hierarchy-consistent organization in learned diagnosis (ICD-10) embedding geometry conducive to analysis and interpretation. On incident cancer prediction, end-to-end fine-tuning achieves test AUCs of 0.81/0.75/0.73 for colorectal cancer (CRC) and 0.94/0.87/0.84 for prostate cancer across 30/60/90-day horizons on balanced cohorts, outperforming logistic-regression baselines, including time-decayed bag-of-codes.Conclusions: HealthFormer provides an event-centric, time-aware representation that transfers via standard fine-tuning without endpoint-specific designs. Using ICD-10 diagnoses and ATC codes can facilitate adoption beyond Hungary. Learned diagnosis embeddings align with the hierarchy, enabling clinical inspection. Broader benchmarking across endpoints remains needed.

  • Open Access

    HealthFormer: Dual-level time-aware Transformers for irregular electronic health record events

    by Péter Kőrösi-Szabó, Gábor Kovács, Botond Forrai, Judit Laki, Tamás Kováts, Miklós Szócska, Adrián Csiszárik

    Longitudinal electronic health records (EHRs) form irregular event sequences that mix multiple clinical coding systems and care settings. Learning transferable patient representations requires modeling both within-encounter code composition and long-range temporal dependencies. We aim to develop a pretraining framework that preserves event structure and explicitly uses elapsed time, while remaining straightforward to fine-tune for new supervised endpoints without task-specific feature engineering. We propose HealthFormer, a dual-level Transformer for event-centric EHR modeling. An Intra-Event Encoder aggregates heterogeneous domain tokens within each typed clinical event into an event embedding via code-specific embedding modules and attention pooling. Event embeddings are combined with a Date Encoder and a continuous-time attention bias based on attention with linear biases (ALiBI) inside an Inter-Event Encoder . We pretrain on Hungarian national administrative health records from a large-scale nationwide longitudinal cohort (spanning millions of individuals over a decade) using multi-task self-supervision with (i) per-domain masked token prediction (masked language modeling, MLM), (ii) event-type prediction under full-event masking (Event-level MLM), (iii) next-event type prediction, and (iv) time-to-next-event (Δ t ) regression. Pretraining induces hierarchy-consistent organization in learned diagnosis (ICD-10) embedding geometry conducive to analysis and interpretation. On incident cancer prediction, end-to-end fine-tuning achieves test AUCs of 0.81/0.75/0.73 for colorectal cancer (CRC) and 0.94/0.87/0.84 for prostate cancer across 30/60/90-day horizons on balanced cohorts, outperforming logistic-regression baselines, including time-decayed bag-of-codes. HealthFormer provides an event-centric, time-aware representation that transfers via standard fine-tuning without endpoint-specific designs. Using ICD-10 diagnoses and ATC codes can facilitate adoption beyond Hungary. Learned diagnosis embeddings align with the hierarchy, enabling clinical inspection. Broader benchmarking across endpoints remains needed.

  • Lessons from the applications and regulation of health sector data in the EU

    by Miklós Szócska, Krisztina Davidovics

    The growth of computing capacities has increased the opportunities to use health-related data for better system capabilities and, ultimately, better services for patients. To fulfill the promises of emerging AI-based technologies, many issues must be resolved. The pacing problem, the Collingridge Dilemma, the Code of Law phenomenon, the Precautionary Principle, and the data-related regulation of health systems are challenges for all rule-of-law-based legal systems. Concerning health care, globalized tech company-generated business models could overrule solidarity-based and national risk-sharing communities. As data-driven solutions are complex systems, traditional legal instruments might not be able to maintain the trust of society. The ongoing EU-level regulatory packages represent efforts to address potential legislative challenges relating to digital transformation. This chapter discusses concepts for a better understanding of existing medical issues and real-life experiences in Europe concerning health-related digital technologies. It is hoped that this discussion will provide valuable lessons for addressing ongoing and future challenges, both within the health sector and concerning wider developments in the rapidly evolving information ecosystem.

  • Open Access

    Are Artificial Intelligence-Assisted Three-Dimensional Histological Reconstructions Reliable for the Assessment of Trabecular Microarchitecture?

    by János Báskay, Dorottya Pénzes, Endre Kontsek, Adrián Pesti, András Kiss, Bruna Katherine Guimarães Carvalho, Miklós Szócska, Bence Tamás Szabó, Csaba Dobó-Nagy, Dániel Csete, et al.

    Objectives: This study aimed to create a three-dimensional histological reconstruction through the AI-assisted classification of tissues and the alignment of serial sections. The secondary aim was to evaluate if the novel technique for histological reconstruction accurately replicated the trabecular microarchitecture of bone. This was performed by conducting micromorphometric measurements on the reconstruction and comparing the results obtained with those of microCT reconstructions. Methods: A bone biopsy sample was harvested upon re-entry following sinus floor augmentation. Following microCT scanning and histological processing, a modified version of the U-Net architecture was trained to categorize tissues on the sections. Detector-free local feature matching with transformers was used to create the histological reconstruction. The micromorphometric parameters were calculated using Bruker’s CTAn software (version 1.18.8.0, Bruker, Kontich, Belgium) for both histological and microCT datasets. Results: Correlation coefficients calculated between the micromorphometric parameters measured on the microCT and histological reconstruction suggest a strong linear relationship between the two with p-values of 0.777, 0.717, 0.705, 0.666, and 0.687 for BV/TV, BS/TV, Tb.Pf Tb.Th, and Tb.Sp, respectively. Bland–Altman and mountain plots suggest good agreement between BV/TV measurements on the two reconstruction methods. Conclusions: This novel method for three-dimensional histological reconstruction provides researchers with a tool that enables the assessment of accurate trabecular microarchitecture and histological information simultaneously.


  • Open Access

    Impact of regulatory tightening of the Hungarian tobacco retail market on availability, access and cigarette smoking prevalence of adolescents

    by Tamás Joó, Kristie Foley, Zoltán Brys, Todd Rogers, Miklós Szócska, József Bodrogi, Péter Gaál, Melinda Pénzes

    Introduction Policies that reduce tobacco retail density to decrease tobacco use among the youth are critical for the tobacco endgame. This paper reviews a Hungarian tobacco regulatory measure, which, since 2013, has confined the sale of tobacco products exclusively to so-called National Tobacco Shops, summarises the changes in the national tobacco retail marketplace and reports on analyses of the impact of this intervention on illegal sales to minors and adolescent smoking behaviour. Methods We reviewed the available national statistical data on the structure and dynamics of the tobacco retail market. Changes in lifetime and current (past 30 days) use of cigarettes among Hungarian adolescents aged 13–17 years were assessed using data from international youth surveys on health behaviours collected in 2010–2020. Results Since the start of policy implementation, the density of tobacco shops in Hungary decreased by 85%, from 4.1 to 0.6 per 1000 persons. The prevalence of lifetime and current cigarette smoking among adolescents declined by 13–24 percentage points (pp) and by 4.8–15 pp, respectively. The rate of illegal sales of tobacco products to minors decreased by 27.6 pp, although the prevalence of compensatory access strategies, especially asking others to buy cigarettes for minors, increased. Conclusions After a significant decrease in the nationwide availability of licensed tobacco retailers, Hungary experienced short-term reductions in youth smoking prevalence. However, the sporadic implementation of complementary, evidence-based tobacco control strategies might limit further declines in youth smoking initiation and tobacco product use.


  • Open Access

    Bizonyítékalapú pszichoszociális kockázatértékelés a munkahelyen: a depressziós tünetegyüttes prediktorainak vizsgálta

    by Katalin Nistor, Gergely Tóth, Miklós Szócska

    Becslések szerint a depresszió 2030-ra a betegségterhek vezető okává válik. Az elmúlt több mint két évtized nemzetközi és hazai kutatási eredményei rávilágítottak arra, hogy egyes munkahelyi pszichoszociális tényezők fokozhatják a hangulatzavarok, ezen belül a depresszió kialakulásának kockázatát, a kapcsolódó tünetetek súlyosbodását, hosszútávon negatív hatást gyakorolva a munkaképességre és szervezeti teljesítményre. Ezért keresztmetszeti nagymintás (N=13104) reprezentatív vizsgálatunk célja a COPSOQ II kérdőívvel vizsgált dimenziónkénti modellépítés keretében a depressziós tünetegyüttes prediktorainak meghatározása, a főbb munkahelyi pszichoszociális kockázati és védő tényezők azonosítása. Eredményeink alapján még a COVID-19 világjárványt megelőző időszakban is magas volt a klinikai depresszió prevalenciája a magyar munkavállalók körében (30,9%). A modellépítések által kontroll alatt tartva a szociodemográfiai tényezőket a depressziós tüneteggyüttes huszonegy indikátorát, azaz nyolc szignifikáns munkahelyi pszichoszociális kockázati- és tizenhárom védő tényezőt sikerült azonosítani. A depressziós tüneteggyüttes azonosított főbb kockázati tényezői a következőek: szekálás (OR=2,92), munka- magánélet konfliktus (OR=2,27), mennyiségi elvárás (OR=1,67), szerepkonfliktus (OR=1,38), erőszakos fenyegetés (OR=1,38), érzelmi megterhelés (OR=1,33), szexuális zaklatás (OR=1,27), munkaütem (OR=1,1). Az azonosított protektív faktorok pedig a következőek: munkahely iránti elkötelezettség (OR=0,44), munkahelyi elégedettség (OR=0,45), munkahelyi közösség (OR=0,64), igazságosság és tisztelet (OR=0,66), vezetés iránti bizalom (OR=0,68), jutalmazás (OR=0,72), munkatársak közötti bizalom (OR=0,83), hatáskör, kontroll (OR=0,85), munkakör egyértelműsége (OR=0,85), támogatás a felettestől (OR=0,89), a munka értelmessége (OR=0,90), előreláthatóság (OR=0,92), támogatás a munkatársaktól (OR=0,94). A depressziós tünetegyüttest prediktáló hét modell közül a Munka-magánélet egyensúly tényezőire, illetve az Együttműködés és vezetés dimenziókra építő modellek bizonyultak a legjobban illeszkedőeknek. A szervezeti életre és produktivitásra, valamint a munkavállalók egészségvédelmére vonatkoztatva az eredmények értékelésének és a legerősebb indikátorok kiemelésének gyakorlati implikációi vannak. Ezekkel útmutatást kívánunk nyújtani a tématerülettel gyakorlati szinten foglalkozó vezetőknek, a HR- és a foglalkozás-egészségügyi szakembereknek arra vonatkozóan, hogy a munkahelyi pszichoszociális környezet mely dimenzióira és tényezőire érdemes kiemelt hangsúlyt fektetni a mentális egészségvédelem, ezen belül is a depressziós tünetegyüttes megelőzésének, illetve csökkentésének érdekében.

  • Open Access

    HunCRC: annotated pathological slides to enhance deep learning applications in colorectal cancer screening

    by Bálint Ármin Pataki, Alex Olar, Dezső Ribli, Adrián Pesti, Endre Kontsek, Benedek Gyöngyösi, Ágnes Bilecz, Tekla Kovács, Kristóf Attila Kovács, Zsófia Kramer, et al.

    Histopathology is the gold standard method for staging and grading human tumors and provides critical information for the oncoteam’s decision making. Highly-trained pathologists are needed for careful microscopic analysis of the slides produced from tissue taken from biopsy. This is a time-consuming process. A reliable decision support system would assist healthcare systems that often suffer from a shortage of pathologists. Recent advances in digital pathology allow for high-resolution digitalization of pathological slides. Digital slide scanners combined with modern computer vision models, such as convolutional neural networks, can help pathologists in their everyday work, resulting in shortened diagnosis times. In this study, 200 digital whole-slide images are published which were collected via hematoxylin-eosin stained colorectal biopsy. Alongside the whole-slide images, detailed region level annotations are also provided for ten relevant pathological classes. The 200 digital slides, after pre-processing, resulted in 101,389 patches. A single patch is a 512 × 512 pixel image, covering 248 × 248 μm2 tissue area. Versions at higher resolution are available as well. Hopefully, HunCRC, this widely accessible dataset will aid future colorectal cancer computer-aided diagnosis and research.


  • Open Access

    Kiberbiztonság a koronavírus idején – a COVID–19 nemzetbiztonsági aspektusai

    by Tamás Palicz, Balázs Bencsik, Miklós Szócska

    Összefoglaló. A COVID–19 pandémia az információbiztonság területén új kihívásokat jelentett. A távolról végzett munka különböző formái jelentős mértékben növelték az online tér biztonsági kockázatát. Nőtt a hálózatok nagysága, az adatforgalom, és azon felhasználók száma, akiknek nem volt érdemi tapasztalatuk az online térben. A járvány ideje alatt a kibertérből érkező támadások szektoronként és időszakonként eltérő intenzitásúak voltak, a támadások típusa a phishingtől a malwareken keresztül az információs zavarkeltésig széles spektrumban változott. Számos jelenségnek nemzetbiztonsági vonatkozásai is voltak. Összefoglaló cikkünkben a fenti jelenségek nemzetközi és hazai tapasztalatait összegezzük, különös figyelmet szentelve az egészségügyi rendszernek, illetve a vakcinafejlesztés kibertérből érkező fenyegetéseinek. Summary. During the COVID-19 pandemic, new challenges emerged in the field of information security and cyber security. Home office, home schooling and distance learning, or even telemedicine hit some organizations unprepared. Security risks in online space have increased significantly: the number of network endpoints and the number of computers, laptops and mobile devices have increased with network data traffic as well as the number of users who had no significant experience in online space. They appeared as a significant risk factor. This has been exacerbated, especially in healthcare, by the extremely high workload, which has made systems highly vulnerable. During the epidemic, attacks from cyberspace varied in intensity from sector to sector and period to period. Statistics from international and national organizations have shown that from the end of the first quarter of 2020, the number of cyber security incidents jumped sharply and then remained high even after a small decline. The types of attacks had an extremely wide range: from phishing through malware to misinformation, almost all types of attacks occurred. Many phenomena also had national security implications. Ransomware virus attacks on health have affected almost all health systems and reached high levels by the end of 2020 in particular. It was during the first period that, in an emergency case, there is thought to be an association between a ransomware virus attack and the death of a patient who was not admitted because of the attack. In addition to distance measures and the associated increase in cyber threats, the emerging threats related to vaccination, which is central to the fight against the epidemic, should also be highlighted. This period has shed light on how many vulnerabilities there are, from vaccine development through drug trials to delivery to vaccines and the organization of vaccines, that cybercriminals are able to attack. In order to prevent and combat these threats and attacks, and to respond appropriately, complex, multidisciplinary collaborations are needed in which security science has a privileged place. In our review article, we summarize the international and national experiences of the above phenomena, paying special attention to the health care system and the threats coming from cyberspace in vaccine development.


  • Open Access

    Előszó

    by Miklós Szócska

    "Előszó" published on 30 Jul 2021 by Akadémiai Kiadó.

  • Open Access

    Adattudományi innováció az egészségügy környezeti kihívásainak kezelésében: a nagy adatállományok hasznosításának jelentősége és lehetőségei a járványkezelésben

    by Péter Gaál, Tamás Joó, Tamás Palicz, Péter Pollner, István Schiszler, Miklós Szócska

    Összefoglaló. A COVID-19 járvány rámutatott arra, hogy az egészségügy kiemelt nemzetbiztonsági terület. Az egészségbiztonsági kockázati tényezőkkel szemben ellenálló egészségügyi rendszerek adattudományi innováció nélkül nem képzelhetők el. A közlemény két esettanulmányon keresztül mutatja be, hogy a nagy, működés során generálódó adatbázisok elemzése hogyan segítheti a járványokkal szembeni védekezést. A mobilcella információk elemzése a leghatékonyabb eszköz a tömeges népességmozgások nyomon követésére, így a vesztegzár intézkedések hatásának döntéstámogatási célú vizsgálatára, az oltásellenes közösségimédia-aktivitás hálózatelemzése pedig segíti az immunizációs kampányok tervezését és megvalósítását. Tanulmányunkban amellett érvelünk, hogy az egészségügy információ- és kommunikációtechnológia fejlődésére építő digitalizációja a kulcsa egy környezeti változásokkal megbirkózni képes egészségügy kialakításának. Summary. The COVID-19 pandemic has shown that health and health care should be considered a top priority area of national security. Health security risks can only be addressed with resilient health systems, which are not possible to be established without innovation in health data science. This publication introduces two examples to illustrate this point, both in the field of the management of epidemics. The first case provides a summary of our previous publication about how mobile phone Call Detail Records can be used to trace population movement to evaluate the effectiveness of movement restriction measures, such as the lock down, which was implemented in Hungary during the first phase of the COVID-19 pandemic. Our analysis shows that the collation and processing of Call Detail Records is an effective and inexpensive method to monitor mass population movement, and complements well the GPS-based smartphone method, which is more suitable for contact tracing and controlling of home quarantine of individuals. Our CDR-based method could be used by other countries, as well as to monitor movement between countries at the European level or internationally, with minimal adaptation effort. The second case introduces a study to gain insight into and better understanding of the potential impact of antivaccination social media activism on the Human Papilloma Virus vaccination campaign in Hungary in 2014. The network analysis of Facebook antivaccination posts and comments showed that during this period, the activists in this network were unable to reach a wider population and were not able to disturb the implementation of the expansion of the well functioning Hungarian public vaccination programme. Unfortunately, this is not the case regarding the COVID-19 vaccination campaign in progress, which suggests that the antivaccination activism is a real and serious security threat to be dealt with. In conclusion, we argue in this paper that the digital transformation of health care, based on the explosive development of information and communication technologies, is of key importance to the establishment of resilient health systems, which are able to cope efficiently with the challenges posed by the rapid environmental changes generated by societal transformation of the 21st century.