https://scholar.dgist.ac.kr/handle/20.500.11750/10139 2026-10-03T21:37:12Z 2026-10-03T21:37:12Z Son, Joonwoo Park, Myoungouk Jeong, Seok Chan Kim, Sung-Hee https://scholar.dgist.ac.kr/handle/20.500.11750/60856 2026-09-21T09:10:17Z 2026-03-31T15:00:00Z

Title: End-to-End 자율주행 AI의 안전성 확보를 위한 기술 동향과 국제 안전 표준 관점의 분석 Author(s): Son, Joonwoo; Park, Myoungouk; Jeong, Seok Chan; Kim, Sung-Hee Abstract: End-to-end (E2E) learning-based autonomous driving has emerged as a promising paradigm that directly maps sensor inputs to vehicle control commands using data-driven models. Compared to conventional modular architectures, E2E approaches offer advantages in terms of architectural simplification and holistic learning of complex driving contexts from large-scale data. However, since E2E systems rely on probabilistic decision-making, exhibit limited explainability, and remain vulnerable to distribution shifts, edge cases, and long-tail scenarios, those benefits actually introduce fundamental challenges in safety assurance. This paper reviews recent technological developments in E2E autonomous driving and systematically analyzes key AI safety issues from a system-level perspective. Core challenges including rare operational scenarios, uncertainty under out-of-distribution conditions, and limitations in traceability and accountability are examined in relation to existing automotive safety frameworks. In particular, the paper investigates how ISO 21448 (Safety of the Intended Functionality, SOTIF) and UL 4600 can be reinterpreted and applied to learning-based autonomous driving systems to complement traditional failure-based functional safety standards. To address the structural mismatch between E2E architectures and existing safety standards, this paper discusses rule-based safety shields and scenario-based validation as practical and standard-compatible mechanisms for mitigating non-failure-based risks and constructing evidence-driven safety arguments. The analysis demonstrates that instead of E2E autonomous driving invalidating existing safety frameworks, it actually necessitates their complementary and systematic integration to achieve robust safety assurance in learning-based autonomous driving systems.

2026-03-31T15:00:00Z Ramu, Dandugudumula Hajra, Sugato Panda, Swati Kaja, Kushal Ruthvik Mishra, Yogendra Kumar Kim, Hoe Joon Kim, Eunjoo https://scholar.dgist.ac.kr/handle/20.500.11750/60674 2026-08-25T07:10:18Z 2026-06-30T15:00:00Z

Title: TENG-Driven Electrotherapy: A Self-Powered Approach to Inducing Cancer Cell Apoptosis Author(s): Ramu, Dandugudumula; Hajra, Sugato; Panda, Swati; Kaja, Kushal Ruthvik; Mishra, Yogendra Kumar; Kim, Hoe Joon; Kim, Eunjoo Abstract: Most of the cancer-related deaths are caused by metastasis, which also remains a significant obstacle to successful clinical management. Even though several anti-metastatic treatments have been put forth, systemic toxicity, low cellular responsiveness, and drug resistance typically undermine their therapeutic efficacy. Although triboelectric nanogenerators (TENGs) have become highly effective self-powered electrical therapies for biomedical applications, their potential as an active treatment tool for metastasis suppression has not yet been fully investigated. Here, we present a self-powered TENG-driven electrotherapeutic approach that suppresses early lung cancer cell migration in vitro by carefully regulated electrical stimulation. Electrical stimulation at 60 V and 760 nA for 5 min disrupted redox homeostasis and induced caspase-3-mediated apoptotic death of A549 cells. TENG-based electrical therapy serves as a self-powered electrical stimulation source that triggers apoptosis in cancer cells by activating the caspase-3/PARP pathway. These results raise possibilities for TENGs not merely as energy-harvesting devices but as active, mechanistic electrotherapeutic platforms, converting mechanical energy into controlled electrical signals that initiate apoptotic cell death pathways in cancer cells.

2026-06-30T15:00:00Z Yun, JinHyo Joseph Zhao, Xiaofei Koo, Inhyouk Del Gaudio, Giovanna Della Corte, Valentina Turon, Katarzyna Yigitcanlar, Tan https://scholar.dgist.ac.kr/handle/20.500.11750/60463 2026-07-22T02:10:16Z 2026-05-31T15:00:00Z

Title: Social economy and social innovation across diverse contexts: City-level findings from Korea, Italy, and Poland Author(s): Yun, JinHyo Joseph; Zhao, Xiaofei; Koo, Inhyouk; Del Gaudio, Giovanna; Della Corte, Valentina; Turon, Katarzyna; Yigitcanlar, Tan Abstract: This study investigates the contextual foundations that shape the characteristics and dynamics of the social economy and its approach to social open innovation across three countries-South Korea, Italy, and Poland. Specifically, it examines three core dimensions: the agenda pursued, the organizational types involved, and the sustainability of social economy actors. Addressing significant gaps in the literature, the study poses two research questions: (a) How do political, economic, and cultural contexts influence the development of the social economy and social open innovation in each country? (b) In what ways do these contextual differences affect organizational forms, strategic agendas, and long-term sustainability? Employing a comparative, multi-method qualitative approach-including in-depth interviews and participatory observation-the findings reveal stark contrasts across the three countries. Italy's social economy is mature and economically grounded, Poland's is culturally driven yet emergent, and South Korea's is politically shaped but organizationally fragile. These differences suggest that context plays a critical role in defining the evolution, focus, and resilience of the social economy and its potential for fostering socially oriented innovation. The study contributes to grounded theory development by offering a comparative framework that links contextual foundations with the trajectory of social innovation across diverse urban settings.

2026-05-31T15:00:00Z Park, Young-Jin Cho, Hui-Sup https://scholar.dgist.ac.kr/handle/20.500.11750/60455 2026-07-20T18:01:13Z 2026-06-30T15:00:00Z

Title: Entropy-Gated Prediction Agreement for Two-View Video Action Recognition Author(s): Park, Young-Jin; Cho, Hui-Sup Abstract: Human action recognition (HAR) often struggles to capture important temporal cues distributed across an entire video when relying solely on a single sampled clip. To overcome this limitation, this study proposes a framework that constructs two temporal views from the same video and explicitly learns the prediction consistency between them. Specifically, the prediction-level agreement (AG) loss was introduced to align the class probability distributions of the two views. In addition, conditional gating was applied to adaptively control the contribution of AG loss according to the sample-wise prediction confidence, thereby reducing unstable alignment in temporally ambiguous or information-insufficient segments. The proposed framework was evaluated using both convolutional neural network (CNN)- and Transformer-based backbones on three representative action-recognition benchmark datasets, and it generally improved the performance over the single-view baseline across backbone-dataset combinations. Further empirical analyses, including training behavior, motion magnitude, temporal prediction stability, and qualitative case studies, were conducted to examine the effectiveness and behavior of the proposed two-view framework from multiple perspectives.

2026-06-30T15:00:00Z