[CSE7001/8001] Creative Software (2026 Spring)

Date:  03/06 (offline, B06)

Speaker: Dr. Sung-Ho Bae, Associate Professor, Kyung Hee University

Title: Psychophysical Vision Models and Their Applications to Image Processing

Abstract: The human visual system (HVS) is considered a very complex nonlinear function, and active research has been carried out in psychophysics to reveal the characteristics of the HVS. This presentation first introduces our recent studies on mathematical modeling of the HVS characteristics. Next, we introduce various applications of the HVS models in computer vision/image processing problems. Specifically, we introduce our new foundation that HVS has different distortion sensitivity depending on the distortion types and texture characteristics. Based on our foundation, we developed a new image quality assessment (IQA) metric that shows significantly high correlations with perceived visual quality. Furthermore, we extended our IQA method to have desirable mathematical properties to be applied in convex optimization problems, such as valid metric properties, differentiability, and convexity which were possessed by MSE (Mean Squared Error). Finally, we applied our HVS models and IQA methods in practical image quality optimization problems, such as video coding, and super-resolution, resulting in better performance in the perspective of visual quality perception compared to existing MSE-based methods.

Date: 04/10 (offline, B06)

Speaker: Dr. Sungwon In, Assistant Professor, Kyung Hee University

Title: From XR to Immersive Data Science: New Ways of Interacting with Data

Abstract: Extended Reality (XR) is transforming how users interact with digital information. This talk first introduces the fundamental concepts of XR, including virtual reality, augmented reality, and mixed reality, and discusses how these technologies enable new forms of interaction beyond traditional desktop interfaces. Next, we will explore how XR can support complex computational workflows, particularly in the context of data analysis and programming. Specifically, the talk will present recent research on immersive data science, focusing on immersive computational notebooks that allow users to spatially organize code, data, and documentation in three-dimensional environments. By leveraging spatial interaction and embodied cognition, these systems aim to improve sensemaking, workflow organization, and the interpretation of complex analytical processes. Finally, the talk highlights future opportunities and challenges in designing immersive systems that support next-generation data science and human-AI collaboration.

Date:   05/08 (offline, B06)

Speaker: Dr. Sungha Choi, Assistant Professor, Kyung Hee university 

Title:Toward On-Device Multimodal Agentic AI with Efficient Small Language Models

Abstract: As AI shifts from the era of Generative AI to Agentic AI, new challenges arise. Agentic AI systems often rely on many repeated LLM calls for planning and reasoning, which leads to high cost and strong dependence on cloud models. This also raises privacy concerns, since personal data may need to be sent off-device. A promising direction is on-device Agentic AI powered by small language models (SLMs), which can reduce cost while keeping sensitive data local. However, replacing LLMs with SLMs often leads to performance degradation. In this talk, I discuss how SLMs can approach LLM-level capability through efficient and latent reasoning, and how streaming multimodal data generated by personal devices can be efficiently processed and understood to enable practical on-device Agentic AI.

Date:  05/22 (offline, B06)

Speaker: Dr. Kevin Nam, Assistant Professor Kyung Hee university 

Title: Privacy-Preserving AI Technologies in a Data-Driven Society: Challenges and Breakthroughs

Abstract: As much as AI services have become an essential part of our daily lives, we began facing a "Mutual Privacy Dilemma" where both user data and proprietary AI models require simultaneous protection. This talk introduces Privacy-Preserving AI (PPAI) as a technical framework to resolve this conflict. We will explore the challenges of private inferences on various AI applications such as cloud-based services and on-device AI, along with several breakthroughs that exploit the use of Privacy Enhancing Techniques (PETs) that made us possible to have a 'peek' at what we may achieve in the future. By analyzing the characteristics of PPAI with PETs, this session highlights how PETs enable us to establish a high-trust infrastructure for future AI ecosystems, and the remaining obstacles that still need to be tackled.

Date: TBD

Speaker: Dr. Gwangmu Lee,    Postdoc Researcher, EPFL  

Title: The chronicle of software vulnerability detection

Abstract: With the threats to cybersecurity ever-increasing, prematurely detecting vulnerabilities in software systems is deemed as one of the promising ways to thwart exploitation attempts from attackers at the fundamental level. In this talk, we first briefly recap software vulnerabilities and move on to the historical evolution of vulnerability detection techniques (i.e., how it has evolved up until now). Finally, we go over the current most popular vulnerability detection technique--fuzzing--and discuss its possible future development.