The continuous advances in computing are reshaping the entire education system, and eventually all disciplines and industries worldwide, under the rapid development in machine learning algorithms, AI and neuromorphic computing, quantum computing, etc.
Workshop on Future Computing invites experts and leaders from academia, industry, and the public sector to share, elaborate, and discuss critical issues, developments, breakthroughs and new perspectives in future computing, including theories, methodologies, architectures, systems, applications, social implications, etc. The annual WFC is organized by Miin Wu School of ComputingSOC, NCKU, and the 1st WFC was held last year and very well received by attendees.
Topics of the workshop this year include Computing Architecture, high-performance & memory-centric computing, AI Robotics, Computational Biomedicine, AI-Assisted Healthcare, AI and Intelligent Mobility, Splendid AI Applications, and visions on future computing.
Miin Wu School of Computing aims to cultivate interdisciplinary and innovative talents with specific domain knowledge and computing expertise. Accordingly, these talents can use advanced computing technology both to solve major social problems and to benefit our nation. Hence, in order to encourage outstanding scholars to lead interdisciplinary research teams and cooperate with SOC, we launched the 2021 NCKU Miin Wu School of Computing AI Summer Summit. The research fields of this summit 2021 focus mainly on Future Computing, AI Robotics, and Computational Biomedicine.
Launching Ceremony of National Intercollegiate Athletic Games 2021
【AI x Future｜AI趨勢論壇】AI-Powered 3D Live Tour Application for Real Estate Industry
The real estate industry has undergone significant transformations with the emergence of innovative technologies. AI-powered solutions have played a pivotal role in enhancing the property viewing experience. This talk explores the development of an AI-Powered 3D Live Tour Application tailored specifically for the real estate industry, with a particular focus on the technology of 3D layout reconstruction from panoramas. The application harnesses the power of artificial intelligence and computer vision algorithms to create immersive and interactive virtual tours of real estate properties. By leveraging panoramic images captured through specialized cameras or smartphones, the application employs advanced 3D layout reconstruction techniques to accurately recreate the physical spaces in a virtual environment. In this presentation, I will highlight three works that tackle the challenge of 3D indoor layout reconstruction from different input settings, including single-view panorama and multi-view panoramas. I will further introduce a transformer-based network in the multi-view setting, enabling simultaneous learning of pose registration and layout reconstruction. This novel approach showcases the potential for more comprehensive and accurate property layout reconstruction, benefiting both potential buyers and real estate professionals.
AI x Future AI趨勢論壇："Deepfake" - Machine Learning for Face Image Generation-COPY
Over recent years, machine learning has revolutionized the generation of photorealistic facial images, with image-to-image transformation techniques being employed in a plethora of applications, including data augmentation, entertainment, virtual reality, and even in the synthesis of deepfake images/videos. In this talk, I plan to review the general machine learning methodologies employed for face image generation before delving into our group's novel research on cross-domain heterogeneous face generation models. I will elucidate how we model facial geometry, sketches, expressions, and
lighting, and how these models enable inference about unseen test subjects. The talk will conclude with the presentation of a framework for a facial generator designed to recover faces obscured by masks, a crucial innovation for surveillance systems in the era of Covid-19.
Jan. 11 AI Master Lecture: Unsupervised, Adaptive, and Advisable Visual Learning
Professor Trevor Darrell is founding co-director of Berkeley Artificial Intelligence Research BAIR, Berkeley Deep Drive BDD, and BAIR Commons. He will give a talk in person about unsupervised, adaptive, and advisable visual learning.
讓科技發揮人性的Problem-Solver 曾煜棋教授大師講座：A Day of a Robot and AI
敏求智慧運算學院很榮幸邀請到曾煜棋教授蒞臨成大演講，講題：A Day of a Robot and AI。他將會就robotic visual task, video action recognition, human motion prediction, thermal image synthesis, digital stand-in, WiFi fingerprinting, medical image inpainting, and continual learning等與機器人相關的研究領域分享最新研究發展。
Computing Architecture大師講座：Dr. Tse-Yu Yeh - A Journey of an Architect from Software to Chip Design
n this talk, Dr. Yeh is going to share his experience from various stages of his career in big computer companies as well as start-up companies. He will talk about his school experience as a computer science undergraduate and a computer architecture graduate student. He will also talk about his professional career as an architect, a microarchitect, a verification lead, a manager of a high performance CPU team, and a VP of silicon engineering.
With the age of Big Data, AI/ML, and IoT technologies coming, people may feel more untrusted of the living environment with certain nervousness and anxiety when using smart devices in daily life. Based on years of industry experience in microcontroller applications, the speaker wants to share his personal experience with the PC/Notebook industry to IoT and future AI/ML enabled smart things, taking into account chip security and the functions required for the development of smart devices including edge and node boxes.
智慧醫療大師講座：NIH Open Data Platforms for Precision Health Research
范揚政教授目前帶領美國國家衛生研究院NIH 院內臨床資訊部門，以及神經異常暨中風研究中心NINDS IT及生物資訊部門，過去建置過多項IT系統和大型生醫資料庫，近期則協助啟動醫療AI資料庫建置計畫。
本次演講，范教授將以NIH Open Data Platforms for Precision Health Research為題與大家分享。
讓數據動起來！資料視覺化大師-馬匡六教授講座 ："Machine Learning Enhanced Data Visualization"
Visualization has become an essential tool in many areas of study using a data-driven approach to problem solving and decision making. However, it can be computationally expensive and also take both novices and experts substantial effort to derive desired visualization results from data for exploration, analysis, or storytelling. Amid active research in new visualization designs, there is a growing interest and opportunity in applying machine learning ML to perform data transformation and to assist in the generation of visualization, aiming to strike a balance between cost and quality/interactivity. In this talk, I will first give a brief overview of my visualization work, and then I will use three of our projects to show how we leverage ML for better supporting common data visualization and analytics tasks.
Image alignment and stitching have been widely applied in multimedia and computer
vision applications. The fundamental idea of image alignment and stitching is image
matching. By matching corresponding key points between two images, image
alignment can be achieved based on the transformation among key points. By aligning
two images, these two images can be stitched. To address this issue, I will introduce
invariant local descriptors for image matching.
MiinStore: the Platform, the Applications and its Gigabit Image Transmission Environment
This talk describes MiinStore, an ABIoT platform AI, Big Data and IoT for cross-domain research and development. The basic structure of MiinStore include IoTtalk IoT+big data, Quanta QOCA big data+AI and Accton ECcloud with P4-switch transport network.
作育全球科技英才 推動未來AI生活—黃正能教授講座 "Real World Image Recognition Challenges"
In this talk, Prof. Huang will start with reviewing recent advances on few-shot and
zero-shot image recognition tasks, then present our research works on long-tailed
recognition tasks, for both closed-set and open-set scenarios, which achieve SOTA
performance on most open-set long-tailed recognition OLTR benchmark datasets.
Making the Invisible Visible: Toward High-Quality Deep THz Computational Imaging
In this talk, we will introduce the characteristics of THz imaging and its applications. We will also show how to break the limitations of THz imaging with the aid of complementary information between the THz amplitude and phase images sampled at prominent frequencies i.e., the water absorption profile of THz signal for THz image restoration. To this end, we propose a novel physics-guided deep neural network model, namely Subspace-Attention-guided Restoration Network SARNet, that fuses such multi-spectral features of THz images for effective restoration.
Computing In-X for Supporting Big Data Applications
In this talk, Prof. Du will introduce several potential improvements, including non-volatile memory, active storage devices, and software-defined networks, in our current computing environment that may help us to face these challenges. In general, to solve these challenges we need to speed up both the data processing and accessing capabilities/performance as well as to increase the space capacities of memory and storage.
Coordinated Exploitation of Big Visual Data Under 5G Era
In this talk, I will first present an automated and robust human/vehicle tracking directly in 3D space through self-calibration of static and moving monocular cameras. When the cameras fail to reliably achieve these tasks due to poor lighting or adverse weather conditions, a radio object detection and tracking framework to detect and track objects purely from radio signals captured by mmWave radars based on an innovative cross-modal supervision framework.
In this presentation, Prof. Chen will first introduce the policies that have been enacted in Germany in response to the energy transition and the coexistence of conventional and renewable energy sources. Then, Prof. Chen will introduce the application of Internet of Things in renewable energy, and the important role of Internet of Things in regulating the electricity market. Then, Prof. Chen will also introduce the German electricity market. Through various trading mechanisms in the electricity market, the supply and demand of electricity are balanced and grid stability is ensured.
In this talk, we will review image compression fundamentals, transforms, and standards. We will answer some questions and address important issues which need to be fully understood and kept in mind in developing new techniques on image compression.
Seminar: Learning without Labeling for Visual Applications
Supervised training with deep Convolutional Neural Networks CNNs has achieved great success in various visual recognition tasks. However, it requires large amount of well-annotated data. Data labeling, especially for large-scale image dataset, is very expensive. How to learn an effective network without the need of training data labeling has become an important problem for many applications.
AI 趨勢跨域論壇II：Data-driven for fault detection and diagnosis
The development of information technology and process technology have been enhanced the rapid changes in high-tech products and smart manufacturing, specifications become more sophisticated. Large amount of sensors are installed to record equipment condition during the manufacturing process. In particular, the characteristics of sensor data are temporal. Most the existing approaches for time series classification are not applicable to adaptively extract the effective feature from a large number of sensor data, accurately detect the fault, and provide the assignable cause for fault diagnosis. This talk presents different methods for fault detection and diagnosis and also extends the topics related to prognostic and health management.