-
- [Student] [23.11.14] Beyond the Sight, Faculty of Global Convergence, Arts Electronica to perform art ①
- The International Media Art Festival, ARS Electronica, was held in Linz, Austria, from September 6 to 10. The festival, which was held around the core agenda of the AI era, says that humans themselves need to ask questions about truth, ownership, rights, control, and responsibility under the theme of "Who Owns the Truth?" Students majoring in culture and technology convergence at the Department of Global Convergence also participated in the Ars Electronica Festival and performed their own art to their fullest. Under the same team called Beyond the Sight, they were divided into City of Hope: Hoffnungburg [Kim Ji-seop (18), Lee Ji-min (19), Discord [Cho Yong-min (18), Park Joon-ha (Cultec/Video 19), The Universe Thinks [Hong Eun-ji (Cultec/Consumer Studies 20), Park Ji-hye (Cultec 20), and Lee Soo-min (Cultec 20). Students said they received a lot of inspiration and feedback from interacting with various artists through this experience. Let's meet Beyond the Sight, who has gained a great driving force in taking the leap to the wider world. Q. Hello. Nice to meet you. Please introduce your work. | | City of Hope: Hoffnungburg Lee Ji-min: City of Hope: Hoffnungburg is an interactive art form that requires participation from visitors. Visitors measure their heart rate through a heart rate measurement sensor. And appreciate the virtual cities that are created. The audience's heart rate data is used as a medium through which the virtual city of Hoffnungburg is created and destroyed. In fact, it was designed to create a virtual building according to the heart rate of the visitors so that each visitor could build a different virtual city. We produced 3D works using Unreal Engine 5, and Arduino and heart rate measurement sensors were used to render data in real time. | | Discord Cho Yong-min: Discord is a media art installation artwork that utilizes projection mapping, objects, and sound effects. It is produced in the form of installation art as a means of effectively delivering messages to the audience, and video media and sound effects are repeatedly played in the form of loops. Discord makes the audience feel audiovisually that an object representing the Earth is unstable due to the sounds of disgust and conflict. The sound generated by the woofer speaker shakes the earth-shaped spherical object connected by wire, and according to the sound, the image reacts as if it were covering the shadow of the spherical object. Through this, the audience can look back on our society and pursue peace by conveying the message that small sounds of hate and conflict can shake the whole planet. | | The Universe Thinks Hong Eun-ji: The Universe Thinks is an interactive piece that expresses the uniqueness of human intelligence with data visualization tools. Three data (average number of filament strands of neurons and cosmic giant structures, beta wave data, and voice frequency data) were used in the work. These three types of data overlap and change throughout the exhibition, including real-time voice data from visitors collected through the microphone. Q. What message did you want to convey through this work? | | City of Hope: Hoffnungburg Kim Ji-seop: The main exhibition theme of ARS Electronica was 'Who Owns the Truth?' Through this exhibition, our team thought about the relationship between artificial intelligence and technology. In fact, I thought that the Fourth Industrial Revolution was an endless development in technology, and that the living environment changed rapidly as it efficiently handled vast amounts of data that were difficult for only humans to handle. Our team thought that only humans could develop technology, and artificial intelligence was also looking at it from the same perspective. So I wanted to show through this work that the full owner of this technology is human. | | Discord Cho Yong-min: Our team thought about what would be the biggest issue that people all over the world could share in common in modern society. Then, I came up with the idea of conveying the message that if small hatred and discrimination spread, it could return to a big conflict, and the work "Discord" started from there. Conflicts on large and small scales have always existed historically, but in modern society, diversity and equality are especially respected, and conflicts between traditional and new perceptions are emanating with the development of the media. In response, our team defined the global spread of mutual hate and discrimination as the biggest fundamental problem in modern society and wanted to raise the audience's awareness of these problems. Through Discord, the sound of hate and discrimination is shown to the audience audiovisually to shake the Earth unstable, allowing the audience to reflect on hate and discrimination and ultimately displaying the work with the aim of peace. | | The Universe Thinks Hong Eun-ji: ARS Electronica held this year's festival based on the question, "Who owns the truth?" Our team, who was interested in digital art, came across a public offering program held by the Department of Culture and Technology Convergence in our school. Since the era of Asian Super Intelligence (ASI) is not far away, we started brainstorming with the question, "Is our intelligence really special?" Then, I learned that the human neuron neural network and the universe's giant structure not only externally resembled, but also had excellent mathematical similarities among many natural phenomena, and I wanted to convey the message that "nature itself represents the uniqueness of human intelligence." Q. I wonder if there are any differences compared to the work I've been doing so far. | | City of Hope: Hoffnungburg Kim Ji-seop: I haven't done many works to compare them, but when I've done them, I wanted to show the convergence of technology and art. In this work, I wanted to create something that develops the work by interacting with external "machines," so I made a production using "Arduino," which others may be familiar with. In the past, if we had manufactured with devices with clear usage and utilization, it seems that designing a machine to expand through Arduino was different from the work we had done so far. Lee Ji-min: I've never made a finished work before. I've had a few experiences planning games and organizing teams so far, but I've never made art like this before. The City of Hope: Hoffnungburg exhibition at the 2023 ARS Electronica is my first work. Compared to game production, I think the characteristic of this interactive art work is that it was able to implicitly express the message I wanted to convey. Games and our works (interactive art) have something in common that they only work with user manipulation. However, while games are designed and produced to deliver "fun," I believe the purpose of interactive art is to deliver a key message through the display. I think this is different from the previously planned content. | | Discord Cho Yong-min: As our team consisted of the Department of Imaging and Nuclear Communication, we had experience producing many works such as motion graphics and movie videos. I've also worked on commercial advertising videos and music production, but most of them were for the general public in Korea. This exhibition at the ARS Electronica Festival was new in that it featured works with numerous foreign artists on the global stage. The genres were really diverse, and I was able to have cross-border exchanges by talking to high-quality artists. Numerous artists and audiences in Europe did not just appreciate the work, but shared and agonized over the meaning of the work by giving constructive feedback. In the process, we were able to have a differentiated exhibition experience that was not possible in Korea. | | The Universe Thinks Park Ji-hye: Before that, I only made works with class projects and I didn't have much experience in making works, so this ARS Electronica was my first step as an artist. As it was my first time, I was disappointed with myself because I didn't know many things and was inexperienced. Honestly, I still don't know what art is. I suddenly think that not knowing what it is will be art. I think when I came back to the text and felt disappointed that I wasn't good at it and didn't know, I rather acknowledged that I didn't know and thought about what I could do and what was the next step. There were many things I've never seen before, and I stayed up all night to make it happen, but I think I could see my limitations. Hong Eun-ji: It's the first time I've ever been involved in digital art or creating exhibition works, not art majors. All they showed in front of people was pictures or videos edited for club performances or hobbies. But everything was a new challenge as I participated as an artist in a global digital art exhibition. Q. Is there a memorable episode that happened during the creative process? | | City of Hope: Hoffnungburg Kim Ji-seop: Most of all, what happened at the exhibition hall was the most memorable. During the exhibition, the role of curator was also performed at the same time, but it was difficult to explain in English and a second language, not Korean, but the audience's position was similar to ours. Even if we didn't understand each other in words, it was most impressive to understand the flow of the conversation. Lee Ji-min: I remember working all night at Hyehwa Cafe. There was a large-scale revision during the production process of the work, and at that time, there were signals saying, "Let's finish it even if we stay up all night." In the late hours, we took turns taking a nap for an hour to cheer each other up, and eventually we completed it. | | Discord Cho Yong-min: Originally, the concrete object to be used in the work was planned to be made using a 3D printer, but the professor who saw the plan asked me to make a prototype by the next day. At that time, with the mindset that the project might fail if I couldn't make a prototype, I remember using woodlock bought at a stationery store and making it by hand for nearly 20 hours. It was difficult at that time, but when I think about it, it was an essential process in the process of making the work. |The Universe Thinks Park Ji-hye: It was fun extracting EEG data measurements myself. Three brainwave data were used to show human intelligence, and beta wave data was extracted from five types of brainwaves. During the meeting, I recorded the EEG data, Eun-ji recorded the data when writing the book report, and Sumin observed the changes when memorizing the play script. It was amazing because it was my first time recording EEG data. Hong Eun-ji: When I look back on the creative process, I remember worrying about whether it would work out as the first runner of Sungkyunkwan University and working in a hurry on Deadline that suddenly appeared. A fun episode was when we first extracted brain waves with an EEG meter. It was fascinating when our EEG graph was drawn on the iPad screen after successfully connecting to the machine. When I closed my eyes and was relaxed, I saw the alpha brainwave data go up, and when I imagined or focused on something, I watched it fluctuate in the beta brainwave quite interestingly.
-
- 작성일 2023-11-20
- 조회수 321
-
- [Student] [SPARK] 2023 K-Hackerton Awards (Minister of Science and Technology Awards)
- The SPARK team, led by industry-academic professor Park Hee-sun, won the grand prize at the 2023 K-Hackerton. The 2023 K-Hackerton hosted by the Ministry of Science and ICT is It is a challenge based on digital new technology convergence solutions (SW, AI, IoT, metaverse, etc.) to solve special problems in people's lives and communities. A total of 21 teams that passed the regional preliminary round advanced to the finals and won the grand prize through the announcement of the finals. ------------------------------------ The 11th K-Hackerton Grand Prize in 2023 - Entry: CareBot (Interactive Recording Emotional Diary via Generative AI) > Topic: Developing CareBot to help ease the daily stress and depression of modern people >Function Developing a Korean LLM with GPT-based empathy to create an interactive recording diary Record the user's experience and remind them of when they experienced something similar and how they resolved it to provide customized feedback Provide diary summary and emotional analysis results focusing on the conversation content (emotion by human relationship, Emotion analysis by case) - Participants: Park Sung-wan (fourth grade in software), Song Hyun-bin (fourth grade in software), Kang Seok-hoon (department of biomecatronics/computer engineering and informatics), and Kim Jae-young (system management/ computer engineering and informatics) - Leadership: Industrial-academia Professor Park Hee-sun - Award details: Grand Prize (Minister of Science and Technology Information and Communication Award) - Prize money: 5 million won
-
- 작성일 2023-11-20
- 조회수 300
-
- [Research] DASH Lab, Three Papers are accepted for publication at CIKM 2023 International Conference and Hosting the First Internat
- DASH Lab’s three papers have been accepted for CIKM (Conference on Information and Knowledge Management) 2023, one of the top-tier international academic conferences in the artificial intelligence and information retrieval field. The papers will be presented in October. Authors are doctoral candidates in computer science and engineering, Eunju Park and Binh M. Le, along with master’s students, Beomsang Cho in computer science and engineering, Sangyoung Lee in artificial intelligence, Seungyeon Baek in artificial intelligence, Jiwon Kim in artificial intelligence. The papers are as follows: 1.Machine Unlearning Research 2.Research on Deepfakes in Collaboration with CSIRO’s Data61 in Australia 3.Research on Datasets for Online ID fraud detection Also, the 1st international workshop on anomaly and novelty detection in satellite and drones systems is hosted at CIKM 2023. The organizing committee consists of Simon S. Woo from Sungkyunkwan University, Shahroz Tariq from CSIRO’s Data61, Youjin Shin from Catholic University, Daewon Chung from Korea Aerospace Research Institute. This workshop is centered around anomaly detection in the time-series and vision data of satellite and drone systems. 1. Sanyong Lee and Simon Woo, “UNDO: Effective and Accurate Unlearning Method for Deep Neural Networks”, Proceedings of the 32nd ACM International Conference on Information & Knowledge Management. 2023. Machine learning has evolved through extensive data usage, including personal and private information. Regulations like GDPR highlight the "Right to be forgotten" for user and data privacy. Research in machine unlearning aims to remove specific data from pre-trained models. We introduce a novel two-step unlearning method, UNDO. First, we selectively disrupt the decision boundary of forgetting data at the coarse-grained level. However, this can also inadvertently affect the decision boundary of other remaining data, lowering the overall performance of the classification task. Hence, we subsequently repair and refine the decision boundary for each class at the fine-grained level by introducing a loss to maintain the overall performance while completely removing the class. Our approach is validated through experiments on two datasets, outperforming other methods in effectiveness and efficiency. 2. Beomsang Cho, Binh M. Le, Jiwon Kim, Simon S. Woo , Shahroz Tariq, Alsharif Abuadbba, and Kristen Moore , “Toward Understanding of Deepfake Videos in the Wild”, Proceedings of the 32nd ACM International Conference on Information & Knowledge Management. 2023. Deepfakes have become a growing concern in recent years, prompting researchers to develop benchmark datasets and detection algorithms to tackle the issue. However, existing datasets suffer from significant drawbacks that hamper their effectiveness. Notably, these datasets fail to encompass the latest deepfake videos produced by state-of-the-art methods that are being shared across various platforms. This limitation impedes the ability to keep pace with the rapid evolution of generative AI techniques employed in real-world deepfake production. Our contributions in this IRB-approved study are to bridge this knowledge gap from current real-world deepfakes by providing in-depth analysis. We first present the largest and most diverse and recent deepfake dataset (RWDF-23) collected from the wild to date, consisting of 2,000 deepfake videos collected from 4 platforms targeting 4 different languages span created from 21 countries: Reddit, YouTube, TikTok, and Bilibili. By expanding the dataset’s scope beyond the previous research, we capture a broader range of real-world deepfake content, reflecting the ever-evolving landscape of online platforms. Also, we conduct a comprehensive analysis encompassing various aspects of deepfakes, including creators, manipulation strategies, purposes, and real-world content production methods. This allows us to gain valuable insights into the nuances and characteristics of deepfakes in different contexts. Lastly, in addition to the video content, we also collect viewer comments and interactions, enabling us to explore the engagements of internet users with deepfake content. By considering this rich contextual information, we aim to provide a holistic understanding of the evolving deepfake phenomenon and its impact on online platforms. 3. Eun-Ju Park, Seung-Yeon Back, Jeongho Kim, and Simon S. Woo, ”KID34K: A Dataset for Online Identity Card Fraud Detection”, Proceedings of the 32nd ACM International Conference on Information & Knowledge Management. 2023. Though digital financial systems have provided users with convenient and accessible services, such as supporting banking or payment services anywhere, it is necessary to have robust security to protect against identity misuse. Thus, online digital identity (ID) verification plays a crucial role in securing financial services on mobile platforms. One of the most widely employed techniques for digital ID verification is that mobile applications request users to take and upload a picture of their own ID cards. However, this approach has vulnerabilities where someone takes pictures of the ID cards belonging to another person displayed on a screen, or printed on paper to be verified as the ID card owner. To mitigate the risks associated with fraudulent ID card verification, we present a novel dataset for classifying cases where the ID card images that users upload to the verification system are genuine or digitally represented. Our dataset is replicas designed to resemble real ID cards, making it available while avoiding privacy issues. Through extensive experiments, we demonstrate that our dataset is effective for detecting digitally represented ID card images, not only in our replica dataset but also in the dataset consisting of real ID cards. 4. The 1st International Workshop on Anomaly and Novelty detection in Satellite and Drones systems (ANSD '23) The workshop on Anomaly and Novelty Detection in Drones and Satellite data at CIKM 2023 aims to bring together researchers, practitioners, and industry experts to discuss the latest advancements and challenges in detecting anomalies and novelties in drone and satellite data. With the increasing availability of such data, the workshop seeks to explore the potential of machine learning and data mining techniques to enable the timely and accurate detection of unexpected events or changes. The workshop will include presentations of research papers, keynote talks, panel discussions, and poster sessions, with a focus on promoting interdisciplinary collaboration and fostering new ideas for tackling real-world problems. Should you have questions, please ask professor Simon S. Woo(swoo@g.skku.edu) in DASH Lab(https://dash.skku.edu).
-
- 작성일 2023-09-18
- 조회수 1121
-
- [Faculty] Kim Yoo-sung, a professor of software department, won the 2023 Spectrum Challenge Competition
- Kim Yoo-sung, a professor of software department, won the 2023 Spectrum Challenge Competition ○ The CSI Lab team has won first place for four consecutive years since 2020 ○ Development of efficient frequency co-use plan in next-generation Wi-Fi environment Software department CSI Lab. Park Tae-gun (Master's Course), Nine-ho (Undergraduate Research Institute), Heo Chan-yong (Intern), and Kim Yoo-sung's team won first place in the final round of the 2023 Spectrum Challenge hosted by the Korea Electronics and Telecommunications Research Institute (ETRI). The Spectrum Challenge has been held annually since 2019 in accordance with the National R&D Plan with the aim of developing a government-led research and development system into an open challenge competitive research and development system. This year's Spectrum Challenge was held for five weeks as a final competition between excellent teams that advanced to the winner's match from 40 teams participating in the competition. There are two types of challenges, Type 1 identifies the propagation characteristics used in a given propagation environment through AI technology, and Type 2 develops reinforcement learning algorithms for resource allocation and scheduling resistant to dynamic environmental changes. In this competition, Hanyang University, Handong University, and Korea University, including our university, were selected as excellent teams. Sungkyunkwan University's Computer Systems and Intelligence Lab (CSI Lab) team won the first place in the algorithm development type 2 to find a radio resource allocation technique that can optimize radio utilization efficiency and communicate in an unlicensed communication environment. In particular, the CSI Lab team has achieved the feat of winning first place for four consecutive years. the body of an article e-newspaper https://www.etnews.com/20230825000155 뉴스1 https://www.news1.kr/articles/?5151113
-
- 작성일 2023-09-18
- 조회수 737
-
- [Faculty] Dr. Kim Mi-soo) was appointed as an assistant professor of artificial intelligence at Chonnam National University
- Dr. Kim Mi-soo, majoring in electronic and electrical computer engineering (computer engineering major)(leading professor Lee Eun-seok) , was appointed as an assistant professor of artificial intelligence at Chonnam National University Dr. Kim Mi-soo, majoring in electronic and electrical computer engineering at the University of Information and Communication She was appointed as an assistant professor of artificial intelligence at Chonnam National University's AI Convergence University as of September 2023. Dr. Kim Mi-soo conducts research at the Software Engineering Research Laboratory (adult professor Lee Eun-seok), In August 2021, he received a Ph.D. from the Department of Electronic Electrical and Computer Engineering (majoring in Computer Engineering) on the subject of Deep learning and natural language processing-based framework for information retrieval-based bug localization. After that, he studied Deep Learning-based Automated Software Debugging and Software Debugging for Deep Learning System for two years as a postdoctoral researcher at the Software Convergence Research Institute of Software University. Dr. Kim Mi-soo has presented a total of 17 outstanding international conferences (13 outstanding conferences of BK/Information Science Society) and It has published a total of three SCI(E) papers and has a total of 10 awards. Dr. Kim Mi-soo said, "We would like to conduct convergence research suitable for the AI convergence era through research on applying AI technology to software engineering (SE4SE), software engineering to AI technology (SE4AI), and environmentally friendly software engineering, "We will make efforts to contribute to society by conducting research on improving sustainability in the field of soft air engineering, which had relatively little interest in carbon neutrality," She said.
-
- 작성일 2023-09-07
- 조회수 931
-
- [Faculty] 2023 SKKU Teaching Award Winner Selected
- 2023 SKKU Teaching Award Winner Selected The Ministry of Education selected 21 teachers as winners of the 2023 SKKU Teaching Award. The award ceremony was held at 4:50 p.m. on August 17 (Thursday) at Jo Byeong-du International Hall, the 600th Anniversary Hall of the Humanities and Social Sciences Campus. (From left in the first row, "professor" word omitted) △ David Roberts (undergraduate) △ Woo Sang-soo (undergraduate) △ Park So-jung (Confucianism, oriental studies) △ Son Hye-sook (English literature) △ Kim Young (French literature) △ Sungjae-ho(law school) △ Kim Min-woo (psychology) △ Kim Hyun-chul (business administration department), Kim Yong-seok, Oh Jong-min(business administration), △ Kim Seung-chul (Degree of Life Science) △ Kim Hyun-soo (Degree of Mathematics) △ Mitragherch (Degree of Electronic Electrical Engineering) △ Kim Young-hoon (Degree of New Material Engineering) △ Lee Sang-won (Degree of Pharmacy) △ Yang Ae-yeon (Department of Pharmacy) △ Han Sun-young (Department of Mathematics Education) △ Oh Jung-soo (Department of Convergence Biotechnology) △ Kim Tae-hee (Department of Sports Science) △ Lim Eun-hyuk (Department of Commerce) △ Lee Jong-wook (Department of Software) 21 teachers, including those, were honored with the SKKU Teaching Award. Meanwhile, the SKKU Teaching Award system is an honorable award to select and reward teachers and instructors who led the quality improvement of out university education, and since its first implementation in 2011, a total of 93 people have been selected until last year. ○ Contact Us: School Affairs Team (02-760-1053/yurimk@skku.edu )
-
- 작성일 2023-09-07
- 조회수 793
-
- [Research] Professor Eom Young-ik's laboratory (Distributed Computing Laboratory, DCLab.) is approved of publication of the SOSP 20
- The 29th ACM Symposium on Operating Systems Principles (SOSP 2023) has been approved for publication in the paper "MEMTIS: Efficient Memory Tiering with Dynamic Page Classification and Page Size Determination" by Professor Eom Young-ik of the Distributed Computing Laboratory and Dr. Lee Tae-hyung. The SOSP Society is the world's leading society for researchers, developers, and programmers in computer systems (BK21+ Recognized International Congress for Computer ScienceIF=4). This paper proposes how to effectively build large memory systems required by modern data center and cloud computing environments. Professor Eom Young-ik's research team proposed MEMTIS, a new hierarchical memory system utilizing DRAM, non-volatile memory (NVM), and CXL memory devices, which are next-generation hardware. Based on its own high-performance memory page management techniques, MEMTIS delivers up to 169% higher performance than modern tiered memory systems. This study was conducted as an international joint study between Professor Eom Young-ik's research team and Professor Min Chang-woo's research team from Virginia Tech in the United States. In addition, through the publication of this SOSP paper, the distributed computing laboratory is the first domestic laboratory to publish two or more papers (SOSP 2021 FragPicker, SOSP2023 MEMTIS). In addition, Professor Eom Young-ik's research team will publish its third top-tier conference paper this year alone, following ASPLOS and MobiCom. [SOSP 2023] The 29th ACM Symposium on Operating Systems Principles, October 23-26, 2023 https://sosp2023.mpi-sws.org/ [About the thesis] MEMTIS: Efficient Memory Tiering with Dynamic Page Classification and Page Size Determination Taehyung Lee, Sumit Kumar Monga, Changwoo Min, Young Ik Eom 29th Symposium on Operating Systems Principles (SOSP 2023) Abstract: The evergrowing memory demand fueled by datacenter workloads is the driving force behind new memory technology innovations (e.g., NVM, CXL). Tiered memory system is a promising solution which harnesses such multiple memory types with varying capacity, latency, and cost characteristics in an effort to reduce server hardware costs while fulfilling memory demand. Prior works on memory tiering make suboptimal (often pathological) page placement decisions because they rely on various heuristics and static thresholds without considering overall memory access distribution. Also, deciding the appropriate page size for an application is difficult as huge pages are not always beneficial as a result of skewed accesses within them. We present Memtis, a tiered memory system that adopts an informed decision-making for page placement and page size determination. Memtis leverages access distribution of allocated pages to optimally approximate the hot data set to the fast tier capacity. Moreover, Memtis dynamically determines the page size that allows applications to use huge pages while avoiding their drawbacks by detecting inefficient use of fast tier memory and splintering them if necessary. Our evaluation shows that Memtis outperforms state-of-the-art tiering systems by up to 169.0% and their best by up to 33.6%. Distributed Computing Lab: http://dclab.skku.ac.kr/xe/
-
- 작성일 2023-08-28
- 조회수 750
-
- [Research] System Security Laboratory (Professor Lee Ho-joon) is Approved tp Publish of Papers in ACM CCS 2023
- The paper "Capacity: Cryptographically-Enforced In-process Capabilities for Modern ARM Architectures" by Dinh Kha (Ph.D. candidate), Cho Kyu-won (Ph.D. candidate), and Noh Tae-hyun (Master's candidate), under the guidance of Professor Lee Ho-jun (https://sslab.skku.edu), has been accepted for publication at the ACM Conference on Computer and Communications Security (CCS) 2023, one of the four major security conferences. The paper will be presented in November. Today's software poses a significant challenge in eliminating vulnerabilities due to its large and complex code base, as well as continuous changes, which often lead to numerous security incidents. In particular, the monolithic nature of various software components residing in a single address space makes the entire program vulnerable even with a single security flaw. To address this issue, the research on In-Process Isolation (IPI) has been widely conducted, aiming to mitigate the risks of vulnerabilities in different domains by isolating programs into multiple domains. The proposed technology, Capacity, extends the existing access control capabilities of operating systems using ARM's new hardware features, namely Pointer Authentication and Memory Tagging Extension, to achieve capability-based access control. Capacity implements a Capability system by cryptographically signing memory pointers and file descriptors, which are reference types for process resources, using keys unique to each domain and verifying their use in all instances. By adhering to the capability philosophy, robust mechanisms are employed to maintain the security of signed references, ensuring high security levels. The practicality and performance of Capacity have been validated through its application to real-world programs such as NGINX and OpenSSH.
-
- 작성일 2023-08-08
- 조회수 1201
-
- [Research] Prof. Jee hyong Lee' Research Lab(IISLab) Publishes a paper in the ICCV 23
- [Abstract] Deep learning models need to detect out-of-distribution (OOD) data in the inference stage because they are trained to estimate the train distribution and infer the data sampled from the distribution. Many methods have been proposed, but they have some limitations, such as requiring additional data, input processing, or high computational cost. Moreover, most methods have hyperparameters to be set by users, which have a significant impact on the detection rate. We propose a simple and effective OOD detection method by combining the feature norm and the Mahalanobis distance obtained from classification models trained with the cosine- based softmax loss. Our method is practical because it does not use additional data for training, is about three times faster when inferencing than the methods using the input processing, and is easy to apply because it does not have any hyperparameters for OOD detection. We confirm that our method is superior to or at least comparable to state- of-the-art OOD detection methods through the experiments.
-
- 작성일 2023-07-25
- 조회수 1473