演講者: 王鈺強 教授
主題: Connecting Vision & Language for Image Understanding and Synthesis
簡介: With the development and applications of AI, it has been a great challenge for researchers to verify whether the AI model is able to learn supporting and explainable information from the data observed.In this talk, we focus on learning from cross-modality data, particularly text and image data. The target text-image related applications include image captioning, text-to-image manipulation, text-to-image retrieval, and so on. We will also talk about our recent CVPR 2022 work, semantics-guided image completion, in which we propose AI models to observe the input image and its semantic information. As a result, our model is able to outpaint the input image with semantic elements properly introduced.
演講者：潘天佑 副院長 (前微軟亞洲研究院副院長、香港大學SPACE客席副教授) 主題：Introducing Microsoft Research - Example of an Industrial Research Lab. 簡介：Microsoft Research (MSR) is a global community that's united by a deep curiosity about solving the world's most intractable problems. We apply our focused expertise and rigorous study to achieving novel breakthroughs that might be applied across a broad range of impactful solutions. And who we are of course feeds what we do. Our ultimate goal is to achieve positive impact for society, which we magnify by collaborating across disciplines, both within MSR and through external partnerships, and by engaging openly with the broader research community.
演講者：Johnny 技術長 (訊連/玩美) 主題：電腦視覺在業界的開發實務與應用 簡介： 機器學習透過電腦視覺可以發展出許多新的應用改變人類日常生活，此演講將介紹機器學習與電腦視覺在業界的實際應用與成果，相關應用包含人臉辨識、防偽偵測、物件導向視訊編輯、聲音樂器分離、數位試妝、數位染髮、數位珠寶穿戴、膚質分析、妝容推薦、微表情偵測等應用。
這場演講也會有一點徵才相關的內容喔！
演講者：李志清 執行長(臺灣希望創新股份有限公司) 主題：無人機群飛技術之研發與應用 簡介： 近年來無人機帶給人們最大的改變，無非就是各種空拍應用上的進步與發展，然而無人機的科技發展速度其實早已超出我們的想像。無人機群飛展演可以算是一種跳脫原有框架的無人機科技與科技藝術結合，隨著許多廠商陸續在世界各地綻放，這項應用越來越為人所知，而臺灣在近年來的無人機表演應用上也有重大突破，無人機展演幕後到底怎麼操作及能帶給國內什麼樣的新氣象呢?無人機群飛技術的架構原理為何？這項技術在未來的生活中又會有什麼樣的創新應用呢？
演講者：華碩 Miranda Wu(毋文涵) 產品處長 主題： Building a PaaS solution to enable the Interoperability of the Healthcare Technology 簡介：Hospital information systems are used by hospitals to manage the patient experience from appointment booking, to the hospital visit, and to billing. More often than not, the information is captured in Electronic Medical Records (EMR) data silos and proprietary formats that makes data exchange painstakingly challenging. ASUS AICS is on the mission to build a PaaS solution to enable the interoperability of the healthcare information technology, binging together service providers and healthcare workers to deliver the best possible healthcare quality and achieve better health outcomes.
演講者： 吳家麟 教授(台大資工)
主題：Some Personal Opinions on the Related Research of the Metaverse
簡介：Most of the predictions made by Market research agencies about the Metaverse or related applications show that it can lead to substantial business opportunities and give them highly positive comments. Such considerable business opportunities of the Metaverse have attracted Microsoft, Facebook (renamed Meta), NVIDIA, and other well-known companies or technology companies to compete for investment.However, in reality, the number of lost users in the illustrative platform "Horizon Worlds" was steady. Meta initially set a goal to have 500,000 monthly active users by the end of 2022. However, now the number is only 200,000, which is lower than the 300,000 users recorded in February at the beginning of the same year, making people question how long the metaverse platform "Horizon Worlds" could last.Comments about the Metaverse have gone from very positive to quite negative in just one year. What's wrong with this emerging field?
In this talk, I will share some personal opinions on the related research of the Metaverse from the following perspectives: