According to an industry survey, many people miss opportunities to apply for government subsidy programs because they do not know how to apply. People also need to search manually and check whether these programs are suitable for them. To address this issue, our study develops a new generative recommender system with both users’ information and government subsidy documents. Within our recommender system framework, we modify the existing Residual Quantization Variational Auto-Encoder (RQ-VAE) model to capture deep and abstract information from subsidy documents. Using semantic IDs generated for approximately 185,610 user click-stream histories and 240,000 documents, we train our recommender system to predict the semantic IDs of the next subsidy policy documents in which a user might be interested. In 2024, we successfully deploy our generative recommender system in Wello, a Korean Gov-Tech startup. In collaboration with the Korean government, our generative recommender system could save 7.8 million dollar, that might otherwise have gone unused due to a lack of applications. Also, Wello observed a 68% improvement in Click-Through Ratio (CTR), increasing from 41.4% in the third quarter of 2024 to 69.6% in the fourth quarter of 2024. We thus anticipate that our generative recommender system will have a significant impact on both individuals and the government.

AI communicators (e.g., AI voice assistants) play an increasingly important role in how individuals receive information, and, sometimes, calling for a more comprehensive understanding of the effectiveness of messages communicated through human and non-human sources. Through a web-based experiment (N = 228), we tested how the persuasive effects of messages are influenced by their format (narrative vs. non-narrative) and the communicator (human voice vs. AI voice) in the scenario of debunking myths about COVID-19 vaccination. The findings revealed that the human communicator was perceived to be more credible and had more influence on participants’ attitude than the AI communicator. Further, the human communicator was particularly persuasive than the AI communicator in delivering a narrative persuasive message, but the effect was not mediated by perceived communicator credibility. The findings augment the literature on narrative persuasion by comparing human and non-human communicators as the delivery source. It also reveals the importance of considering non-human information communicators in research on narrative persuasive messages.

This paper proposes a sound event detection (SED) model using an EfficientNet-B2 and an attention and aggregation-based feature pyramid network (A2-FPN). In particular, the EfficientNet-B2 is first obtained from the pretrained model on the basis of the pretraining, sampling, labeling, and aggregation (PSLA) framework. Then, the A2-FPN module is applied to the outputs of the layers of the EfficientNet-B2 to deal with the different time and frequency resolutions from acoustic features. The aggregated feature map from the A2-FPN module is used as input features to two bidirectional gated recurrent unit layers. Specifically, the proposed A2-FPN-based SED model is trained by the mean-teacher approach to utilize weakly labeled and unlabeled data. Finally, the proposed A2-FPN-based SED model is applied to the detection and classification of acoustic scenes and events (DCASE) 2021 Challenge Task 4. Consequently, it is shown that the polyphonic sound event detection score (PSDS) 1 and 2 of the proposed A2-FPN-based SED model are the higher of 0.03 and 0.172, respectively, than those of the DCASE 2021 Challenge Task 4 baseline.

For readability and disambiguation of the written text, appropriate word segmentation is recommended for documentation, and it also holds for the digitized texts. If the language is agglutinative while far from scriptio continua, for instance in the Korean language, the problem becomes more significant. However, some device users these days find it challenging to communicate via key stroking, not only for handicap but also for being unskilled. In this study, we propose a real-time assistive technology that utilizes an automatic word segmentation, designed for digital minorities who are not familiar with electronic typing. We propose a data-driven system trained upon a spoken Korean language corpus with various non-canonical expressions and dialects, guaranteeing the comprehension of contextual information. Through quantitative and qualitative comparison with other text processing toolkits, we show the reliability of the proposed system and its fit with colloquial and non-normalized texts, which fulfills the aim of supportive technology.

As autonomous-vehicle technologies advance, conventional taxi and car-sharing services are being combined into a shared autonomous vehicle service, and through this, it is expected that the transition to a new paradigm of shared mobility will begin. However, before the full development of technology, it is necessary to accurately identify the needs of the service’s users and prepare customer-oriented design guidelines accordingly. This study is concerned with the following problems: (1) How should an autonomous taxi service be designed and field-tested if the self-driving technology is imperfect? (2) How can imperfect self-driving technology be supplemented by using service flexibility? This study implements an autonomous taxi service prototype through a Wizard of Oz method. Moreover, by conducting field tests with scenarios involving an actual taxi, this study examines customer pain points, and provides a user-experience-based design solutions for resolving them.

As smoking has emerged as a health-risk behavior, communication scholars and practitioners have put many efforts into finding out ways to achieve smoking cessation. In this research, participants (N = 57) were randomly assigned to a spatial augmented reality (SAR) condition (3D projection mapping) and 2D flat screen to be exposed to an anti-smoking message. This research provides insightful evidence that the effects of SAR on people’s behavioral intention to spread anti-smoking messages online could be explained by spatial presence and negative emotions. Implications for research on the potential of SAR in terms of emotions and online viral behavioral intentions are discussed.

The paper proposes an implementation of a motion sensorless control system in a very wide speed range - from zero speed operation, based on direct and quadrature inductance components of the stator flux linkage for the interior permanent magnet synchronous motor-IPMSM, without signal injection. The proposed observer is developed using the Lyapunov design, resulting a scheme with a full observer for the motor states. The observer assures stability even for zero speed. The proposed observer can be applied to other types of synchronous machines such as surface permanent magnet synchronous motor-SPMSM, dc. excitation SM, variable reluctance synchronous motor-VRSM or transverse-flux motors-TFMs. Extensive experimental results are presented to verify the principles and to demonstrate the effectiveness of the proposed sensorless control system. With the so-called by the authors-the "stator flux linkage d-q inductances components observer", the IPMSM vector controlled drive system operates from zero to the maximum speed of more then six times the rated speed-on the flux weakening region. The experiments were conducted on 2.7kW IPMSM which is vector controlled by a floating-point TMS320C6713-Digital Signal Processor. The results prove very good dynamic performances in all ranges of speed and torque operation.