Engaging learners in dialogue around controversial issues is essential for examining diverse values and perspectives in pluralistic societies. While prior research has identified productive discussion moves mainly in STEM-oriented contexts, less is known about what constitutes productive discussion in ethical and value-laden discussions. This study investigates productive discussion moves in AI ethics dilemmas using a dialogue-centric learning analytics approach. We analyzed small-group discussions among 51 undergraduate students through a hybrid method that integrates expert-informed coding with data-driven topic modeling. This process identifies 14 discussion moves across five categories, including Elaborating Ideas, Position Taking, Reasoning & Justifications, Emotional Expression, and Discussion Management. We then examined how these moves relate to discussion quality and analyzed sequential interaction patterns using Ordered Network Analysis (ONA). Results indicate that emotive/experiential arguments, as well as explicit acknowledgment of ambiguity, are strong positive predictors of discussion quality, whereas building on ideas is negatively associated. ONA further reveals that productive ethical discussions are characterized by interactional patterns that connect emotive/experiential expressions to evidence-based reasoning. These findings suggest that productive ethical discussion is grounded not only in reasoning and justification but also in the constructive integration of emotional expression.

Deaf and hard-of-hearing (DHH) students often experience delays in print literacy due to limited access to spoken language. This study developed a prototype of an AI-integrated digital textbook (AIDT) for Korean language learning, which incorporated visual learning materials and simplified versions of complex texts to enhance accessibility and support reading comprehension. Personalized instruction tailored to individual needs was also embedded, based on insights gathered through prior participatory design phases. A usability test was conducted with six DHH students and four teachers using five learning tasks. The results revealed both the strengths and limitations of the AIDT prototype. Students appreciated the interactive and visually rich features by reporting increased engagement and a greater sense of independence in their learning. However, they encountered challenges related to insufficient sign language support and a non-intuitive interface. Teachers recognized the potential of AIDT to reduce instructional workload and to promote self-directed and personalized learning. At the same time, they emphasized the need for clearer navigation features to better track student learning progress. These findings suggest that AIDTs can be a valuable tool for supporting Korean language learning among DHH students when designed with accessibility and usability in mind. We propose future research directions that include refining the interface design, enhancing sign language scaffolding, improving navigation features, and exploring classroom integration to better support the diverse learning needs of DHH students.

This study explores the student perceptions of Generative AI (GenAI) tools, such as ChatGPT, on academic writing across different educational levels in higher education. Using topic modeling, this study analyzed open-ended responses from 189 university students in Korea to identify the dominant themes related to these tools. The findings reveal that students across all education levels generally recognize the potential of ChatGPT to enhance writing efficiency, idea generation, and organizational structure in their academic work. However, concerns vary by educational level, with graduate students expressing a more cautious and critical attitude towards the use of GenAI tools. These insights provide a deeper understanding of the impact of GenAI on academic writing from a student perspective and offer valuable considerations for instructional support in higher education.

Minimax problems are notoriously challenging to optimize. However, we present that the two-timescale extragradient method can be a viable solution. By utilizing dynamical systems theory, we show that it converges to points that satisfy the second-order necessary condition of local minimax points, under mild conditions that the two-timescale gradient descent ascent fails to work. This work provably improves upon all previous results on finding local minimax points, by eliminating a crucial assumption that the Hessian with respect to the maximization variable is nondegenerate.

In nonconvex-nonconcave minimax optimization, two-timescale gradient methods have shown their potential to find local minimax (optimal) points, provided that the timescale separation between the min and the max player is sufficiently large. However, existing two-timescale variants of gradient descent ascent and extragradient methods face two shortcomings, especially when we search for non-strict local minimax points that are prevalent in modern overparameterized setting. In specific, (1) these methods can be unstable at some non-strict local minimax points even with sufficiently large timescale separation, and even (2) computing a proper amount of timescale separation is infeasible in practice. To remedy these two issues, we propose to incorporate two simple but provably effective schemes, double-step alternating update and increasing timescale separation, into the two-timescale extragradient method, respectively. Under mild conditions, we show that the proposed methods converge to non-strict local minimax points that all existing two-timescale methods fail to converge.

Can we effectively train a generative adversarial network (GAN), i.e., optimize a minimax problem, similar to classification neural networks, i.e., minimize a function, by gradient methods? Currently, the answer to this question is “No”. Despite the extensive studies, training GANs still remains challenging, and diffusion-based generative models are largely replacing GANs. When training GANs, we not only struggle with finding stationary points, but also suffer from the so-called mode-collapse phenomenon, generating samples that lack diversity compared to the training data. Due to the nature of GANs, mode-collapse is likely to occur when we find an optimal point for the maximin problem, rather than the original minimax problem.This suggests that answering to an open question of whether there exists a first-order method that only converges to (local) optimum of minimax problems can resolve these issues. None of the existing methods possess such a property, neither theoretically nor practically. This is in contrast to standard gradient descent successfully finding local minima. In nonconvex-nonconcave minimax problems, Jin et al. are the first to suggest an appropriate notion of local optimality, taking account of the order of minimization and maximization. They also presented a partial answer to the question above, showing that two-timescale gradient descent ascent converges to strict local minimax optima. However, the convergence to general local minimax optimum was left mostly unexplored, even though non-strict local minimax optima are prevalent. Our recent findings illustrate that it is possible to find some non-strict local minimax optima by a two-timescale extragradient method.This positive result brings new attention to the open question. Furthermore, we wish to revive discussion on the appropriate notion of local minimax optimum. This was initially discussed by Jin et al., but not much thereafter, which we believe is crucial in answering the open question.

One aim of detection proposal methods is to reduce the computational overhead of object detection. However, most of the existing methods have significant computational overhead for real-time detection on mobile devices. A fast and accurate proposal method of human detection called personness estimation is proposed, which facilitates real-time human detection on mobile devices and can be effectively integrated into part-based detection, achieving high detection performance at a low computational cost. Our work is based on two observations: (i) normed gradients, which are designed for generic objectness estimation, effectively generate high-quality detection proposals for the person category; (ii) fusing the normed gradients with color attributes improves the performance of proposal generation for human detection. Thus, the candidate windows generated by the personness estimation will very likely contain human subjects. The human detection is then guided by the candidate windows, offering high detection performance even when the detection task terminates prior to completion. This interruptible detection scheme, called anytime detection, enables real-time human detection on mobile devices. Furthermore, we introduce a new evaluation methodology called time-recall curves to practically evaluate our approach. The applicability of our proposed method is demonstrated in extensive experiments on a publicly available dataset and a real mobile device, facilitating acquisition and enhancement of portrait photographs (e.g. selfie) on widespread mobile platforms.

In this work, we study a real-time human detection method for mobile devices using window proposals. We find that the normed gradients, designed for generic objectness estimation, are also able to rapidly generate high quality object windows for a singlecategory object. We also notice that fusing the normed gradients with additional color feature improves the performance of objectness estimation for the single-category object. Based on these observations, we propose an efficient method, which we call personness estimation, to produce candidate windows that are highly likely to contain a person. The produced candidate windows are used to search over feature maps of an image so that a human detection method can achieve high detection performance within a short period of time. We further present how personness estimation can be efficiently combined into part-based human detection. Our experiments indicate that the proposed method is directly applicable to mobile devices, and allows real-time human detection.

This paper presents a vehicle sensor and actuator fault detection algorithm for automated vehicles. The diagnostic system is designed to monitor steering wheel angle, yaw-rate, and wheel speed sensors and steering, throttle, and brake actuators used by the lateral and longitudinal controllers of the vehicle. Different combinations of the observer estimates, the sensor measurements, and the control commands are used to construct a bank of residuals. A fault in any of the vehicle sensors and actuators leads to increase of the unique subset of residuals. The adaptive threshold is used to enable exact identification of the abnormal increase of residual. The fault detection performance and its reliability of the proposed algorithm have been investigated via computer simulation studies and real-time vehicle tests. The enhancement of the fault detection allows for realization of autonomous driving vehicle which uses actuation by embedded computer.

This paper presents a time delay compensation algorithm for environmental sensors of automated driving systems. The time delay involved with the transmission of the measurements from the sensors to the processor cannot be negligible because it is responsible for estimation and control of the system. As the automotive environmental sensors such as laser scanner or radar perform measurements at a constant frequency, the measurement time latencies can be assumed to be constant. From this aspect, the constant time delay characteristics is analyzed via vehicle tests and compensated by forward estimation based coordinate transformation. The proposed compensation algorithm has been verified via test data based open loop simulation of Automated Driving Systems (ADS). It is shown that the proposed compensation enhances environment perception performance and driver's safety.

This paper deals with lateral disturbance compensation algorithm for an application to a Motor Driven Power Steering (MDPS) based driving assistant system. The lateral disturbance such as wind force and load from bank angle reduces the driver refinement and increases the possibility of an accident. In order to reduce the maneuvering effort of the driver in the disturbing situation, the lateral disturbance compensation algorithm has been proposed. The characteristics of the compensation system including a human driver model and the steering system have been mathematically analyzed. The control strategy using the motor overlay torque as a control input which improves the human steering behavior under the lateral disturbance has been proposed. A numerical simulation of the proposed algorithm has been conducted by full vehicle model and lateral driver model which represents steering behavior of human driver. The human torque and lateral deviation under the lateral disturbance are confirmed to be reduced by the simulation results.