| 主持人(PI) | 研究主題(Research Topic) | 研究介紹(Introduction) | 其他資訊(Other Information) |
|---|---|---|---|
| 張原豪 Yuan-Hao Chang | 嵌入式系統與存儲系統研究 Study of embedded systems and storage systems | 研究主題包含: 嵌入式系統、作業系統、檔案系統、記憶體管理、儲存系統管理、非揮發性記憶體研究、能源採集系統 Studied topics include embedded systems, file systems, operating systems, memory management, storage system management, non-volatile memory, and energy-harvesting system. |
PI個人首頁(PI's Information) : https://homepage.iis.sinica.edu.tw/~johnson/ 實驗室網址(Research Information) : http://www.iis.sinica.edu.tw/pages/johnson/ https:// Email : johnson@iis.sinica.edu.tw |
| 吳毅成 I-Chen Wu | 深度強化式學習與其應用 Deep Reinforcement Learning and Its Applications | 本研究主題是深度強化式學習與其應用. 在2015年之前, 我們的應用主要是以電腦對局遊戲為主, 如棋牌類遊戲, 益智遊戲, 電玩遊戲. 在2015年後, 我們延伸應用到許多其他應用, 如以學習為基礎的機器手臂、模型賽車、智慧路網、工廠排程、組合最佳化問題等. 若有興趣, 歡迎加入我們. Our research interests are related to Deep Reinforcement Learning (DRL) and its applications. Before 2015, we focused more on various research topics related to computer games, such as board games, card games, puzzle games, video games, etc. After 2015, we extended our researches from computer games to various DRL applications, such as learning-based robotic manipulation and car racing (like AWS DeepRacer), intelligent traffic managements, scheduling for manufacturing, combinatorial optimization, etc. If you are interested in above, welcome to join us. We believe this lab would be a very good place where you can enjoy learning DRL. |
PI個人首頁(PI's Information) : http://www.citi.sinica.edu.tw/pages/icwu/ 實驗室網址(Research Information) : https://cgilab.nctu.edu.tw/ https:// Email : icwu@citi.sinica.edu.tw |
| 王釧茹 Chuan-Ju Wang | 表示學習演算法於人工智能之應用 Representation learning and its applications | 異質性資料涵括各式結構化（如：消費記錄、產品規格）及非結構化數據（如：網友文字評論），其各自的資料結構及特徵空間大不相同，因此如何進行彼此間的關聯、整合及推論仍屬當代人工智能技術及其相關應用的一大挑戰。然而透過機器學習的非監督式學習法則有可能將異質性資料表現於共同的特徵空間之中，倘若又能在此空間中獲得優良的資料表示法，則可作為異質性資料分析的穩固基石。因此，本研究主題從深度學習及網路表示法的框架切入，深入探究其空間轉換的特性及其保留的訊息，並將針對不同的資料型式及應用情境設計對應之演算法。除了演算法設計及理解外，本實習亦具有有以下三個特色：1) 將使用真實世界的資料進行資料分析及學習；2) 將學習如何在unix-like 環境下處理大量資料並運行實驗；3) 將學習如何使用網頁前端技術進行結果之視覺化呈現。 The research topics will be related to the processing and understanding heterogeneous data (including texts, pictures, audio signals, social relations, and user behaviors) and using the deep learning and/or network embedding techniques for various AI-enable applications. In addition to the model design, during the internship, the participant will also 1) have hands-on experience with real-world data, 2) learn how to deal with large-scale data and conduct experiments under unix-like systems, 3) learn how to visualize the learned results using front-end web programming techniques. |
PI個人首頁(PI's Information) : https://cfda.csie.org/~cjwang/ 實驗室網址(Research Information) : http://cfda.csie.org https:// Email : cjwang@citi.sinica.edu.tw |
| 柯向上 Hsiang-Shang Ko | 型別互動程式設計／量子程式的圖像推理 Interactive type-driven programming / Diagrammatic quantum programming | 本實驗室有兩個實習題目供選擇。 1. 型別互動程式設計 程式語言若有基本的型別系統 (type systems)，我們便能避免寫出某些無意義的程式（例如使用某函式時應輸入字串，我們卻傳入整數）。若有更強的型別系統，我們不僅能排除更多無意義的程式，甚至只能寫出有意義的程式。依值型別 (dependent types) 的表達能力直接對應於高階邏輯 (higher-order logic)，足以表達正確程式應滿足的性質；當依值型別與互動式開發環境 (IDE, interactive development environment) 結合，寫程式時 IDE 就能進行型別推導、回答我們程式各部分應滿足什麼性質，並依型別資訊幫我們（自動或半自動地）產生程式。 這部分實習內容將以 ‘PLFA’ 這份線上教材入門： * Philip Wadler, Wen Kokke, and Jeremy G. Siek [2020]. Programming language foundations in Agda. https://plfa.github.io. 今年此主題的實習生建議選修或旁聽「邏輯、語言、與計算」FLOLAC 暑期研習營（可參考 2020 年網站 https://flolac.iis.sinica.edu.tw/zh/2020/），即涵蓋 PLFA 之重要部分。有了 PLFA 的基礎後，我們可試著寫一些較複雜的依值型別程式／演算法，例如： * Hsiang-Shang Ko [2021]. Programming metamorphic algorithms: An experiment in type-driven algorithm design. The Art, Science, and Engineering of Programming, 5(2):7:1–34. https://josh-hs-ko.github.io/#publication-9f9adfcc * Hsiang-Shang Ko and Jeremy Gibbons [2017]. Programming with ornaments. Journal of Functional Programming, 27:e2:1–43. https://josh-hs-ko.github.io/#publication-696aedff 2. 量子程式的圖像推理 一派計算學家於本世紀發展出「範疇量子力學」(Categorical Quantum Mechanics)，以高度抽象的範疇論 (category theory) 重新省視量子理論並構築一套更高階 (higher-level) 的論述。這套抽象的論述本質其實是一套圖像化算則 (graphical calculus)，操作起來相當省力（特別相較於傳統的線性代數計算），並能清楚顯現計算上的直覺。而且因為理論高度抽象，這套算則也能用於具機率或不確定性之程式 (probabilistic or non-deterministic programs)，甚至可以在同一語言內融合地論證量子與古典／機率性質。 這部分實習內容將以研讀討論 ‘PQP’ 這本教科書為主： * Bob Coecke and Aleks Kissinger [2017]. Picturing Quantum Processes. Cambridge University Press. ISBN: 9781107104228. https://doi.org/10.1017/9781316219317 若進度夠快，我們可試著比較 PQP 和標準教科書之作法： * Michael A. Nielsen and Isaac L. Chuang [2010]. Quantum Computation and Quantum Information. Cambridge University Press, 10th anniversary edition. ISBN: 9781107002173. https://doi.org/10.1017/CBO9780511976667 以及用 PQP 的圖像化算則試著（重新）寫一些關於量子演算法的正確性或複雜性論證。 There are two possible topics. 1. Interactive type-driven programming Basic type systems help to preclude a class of non-sensical programs (for example, passing integers to functions expecting strings). Stronger type systems preclude more non-sensical programs, and even better, allow only sensical programs. Corresponding to higher-order logic, dependent types are highly expressive and capable of describing program correctness properties; when programming with dependent types in an interactive development environment (IDE), the IDE can reason about types on our behalf and let us know what properties should be satisfied by any part of a program upon request, and generate programs (automatically or semi-automatically) based on type information. We will study the ‘PLFA’ online tutorial: * Philip Wadler, Wen Kokke, and Jeremy G. Siek [2020]. Programming language foundations in Agda. https://plfa.github.io. It is recommended that interns on this topic attend this year’s Formosan Summer School on Logic, Language, and Computation (FLOLAC) (for details see the website for the 2020 edition https://flolac.iis.sinica.edu.tw/zh/2020/), which will cover the essential parts of PLFA. Afterwards we will work on some more sophisticated dependently typed programs/algorithms, for example: * Hsiang-Shang Ko [2021]. Programming metamorphic algorithms: An experiment in type-driven algorithm design. The Art, Science, and Engineering of Programming, 5(2):7:1–34. https://josh-hs-ko.github.io/#publication-9f9adfcc * Hsiang-Shang Ko and Jeremy Gibbons [2017]. Programming with ornaments. Journal of Functional Programming, 27:e2:1–43. https://josh-hs-ko.github.io/#publication-696aedff 2. Diagrammatic quantum programming The ‘Categorical Quantum Mechanics’ project of this millennium re-examines quantum theory and builds a higher-level formulation based on the highly abstract language of category theory. Despite being abstract, the essence of the new formulation is a graphical calculus, which is easy to manipulate (especially compared to the traditional linear-algebraic calculations) and readily reveals the computational intuitions. The abstract nature of the formulation makes it applicable to probabilistic or non-deterministic programs, and we can even uniformly reason about quantum and classical/probabilistic properties within the same language. We will mainly study the ‘PQP’ book: * Bob Coecke and Aleks Kissinger [2017]. Picturing Quantum Processes. Cambridge University Press. ISBN: 9781107104228. https://doi.org/10.1017/9781316219317 If time permits, we could try to compare the approaches of PQP and the standard textbook: * Michael A. Nielsen and Isaac L. Chuang [2010]. Quantum Computation and Quantum Information. Cambridge University Press, 10th anniversary edition. ISBN: 9781107002173. https://doi.org/10.1017/CBO9780511976667 Also we could try to (re-)write some correctness or complexity arguments about quantum algorithms in terms of PQP’s graphical calculus. |
PI個人首頁(PI's Information) : https://josh-hs-ko.github.io 實驗室網址(Research Information) : https://www.iis.sinica.edu.tw/pages/joshko/ https:// Email : joshko@iis.sinica.edu.tw |
| 修丕承 Pi-Cheng Hsiu | 具間歇性感知之深度學習 Intermittent-aware Deep Learning | 此計畫屬於嵌入式系統研究領域，特別關注「間歇性智慧物聯網系統」，讓物聯網裝置可靠環境中不穩定的供電，間歇性地執行深度學習模型，無須再安裝電池而永續運作。我們開發系統軟體，讓深度學習模型得以輕鬆部署且間歇性地執行於無電池的物聯網裝置，使得AI研究人員可以專注於深度網絡設計而不是模型佈署。學生將整合並佈署我們開發的「間歇性作業系統」與「間歇性深度學習推論工具」於超低功率嵌入式裝置，並學習到系統實作與開發的經驗。 This project’s scope lies in the area of embedded systems, with a special focus on enabling battery-less IoT devices to intermittently execute deep neural networks (DNN) via ambient power. We develop system software for AI researchers to easily deploy and efficiently execute their DNN models onto battery-less devices, so that AI researchers can focus on deep network design rather than model deployment. You are expected to learn rich hands-on experience in prototype implementations and hacking system kernels, by integrating and deploying our previously developed intermittent operating system and intermittent deep learning tool onto ultra-low power embedded platforms. |
PI個人首頁(PI's Information) : http://www.citi.sinica.edu.tw/pages/pchsiu/ 實驗室網址(Research Information) : https://emclab.citi.sinica.edu.tw https:// Email : pchsiu@citi.sinica.edu.tw |
| 楊奕軒 Yi-Hsuan Yang | 自動音樂生成: MIDI、聲音、與圖像 Automatic music generation: Generating MIDI, Sounds, Images | 見英文版 We are interested in both symbolic (MIDI) domain and audio-domain music generation; the former concerns with generating MIDI scores [1, 2, 3, 4] while the latter generate sounds [5, 6, 7, 8]. We are also interested in multi-modal generation models that generate not only audio but also the visual counterpart [9, 10]. We welcome intern candidates who has solid backgrounds/experiences/understandings in deep generative models such as Transformers, GANs, and flow based models [11], with strong motivation to publish papers in top AI/ML conferences as a result of the internship. Experience in music playing and/or composition is a plus but not a must. Our lab has close collaboration with the Taiwan AI Labs, Sony Japan, and research labs in other countries. Please feel free to drop me a mail to show passions and for questions. [1] CP Transformer. https://arxiv.org/abs/2101.02402 [2] Pop Music Transformer. https://arxiv.org/abs/2002.00212 [3] EMOPIA. https://arxiv.org/abs/2108.01374 [4] MuseMorphose. https://arxiv.org/abs/2105.04090 [5] KaraSinger. https://arxiv.org/abs/2110.04005 [6] UNAGAN. https://arxiv.org/abs/2005.08526 [7] Jukebox. https://openai.com/blog/jukebox/ [8] Loop Combiner. https://arxiv.org/abs/2008.02011 [9] StyleGAN. https://arxiv.org/abs/1812.04948 [10] DALL.E. https://openai.com/blog/dall-e/ [11] https://courses.cs.washington.edu/courses/cse599i/20au/ |
PI個人首頁(PI's Information) : http://www.citi.sinica.edu.tw/pages/yang/ 實驗室網址(Research Information) : https://musicai.citi.sinica.edu.tw/ https:// Email : affige@gmail.com |
| 王柏堯 Bow-Yaw Wang | 正規驗證密碼程式 Formal Verification of Cryptography Programs | 密碼程式為資訊安全之基礎，而密碼程式多採自廣為使用之密碼程式庫如OpenSSL。本研究將利用實驗室所開發之工具，將OpenSSL及其他廣為使用之程式庫中之密碼程式，進行正規驗證。若是對正規驗證工具有興趣，也可以修改工具，以提昇使用性或效率。 Cryptography programs are the foundation of computer security. Many cryptography programs are adopted from widely used cryptography libraries such as OpenSSL. We are going to use CryptoLine to formally verify cryptography programs from OpenSSL and other widely used libraries. If you are interested in formal verification tools, you can also modify CryptoLine to improve usability or efficiency. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/~bywang 實驗室網址(Research Information) : https://github.com/fmlab-iis/cryptoline https:// Email : bywang@iis.sinica.edu.tw |
| 呂及人 Chi-Jen Lu | 深度學習的原理與應用 Deep learning: foundations and applications | 研究深度學習的原理，並探索深度學習在強化學習、影像處理、自然語言等各個領域的應用。 Study the foundation of deep learning, and explore its diverse applications in various areas such as reinforcement learning, computer vision and natural language processing. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/cjlu/ 實驗室網址(Research Information) : http://www.iis.sinica.edu.tw/pages/cjlu/ https:// Email : cjlu@iis.sinica.edu.tw |
| 王志宇 Chih-Yu Wang | 邊緣智慧系統最佳化 / 圖神經網路應用 Edge Intelligence System Optimization / Graph Neural Network Application | 從事邊緣智慧(Edge Intellignece)相關技術探討與最佳化 或 圖神經網路(Graph Neural Network)相關研究應用，可參考PI網址相關研究，確切題目待面議。 Our goal is to identify, analyze, predict, and manage the strategic behaviors of humans in various of networks such as communication network, information network, social network, etc. The main theme of the internship this year are Edge Intelligence and Graph Neural Network. |
PI個人首頁(PI's Information) : http://www.citi.sinica.edu.tw/pages/cywang/ 實驗室網址(Research Information) : http://snaclab.citi.sinica.edu.tw https:// Email : cywang@citi.sinica.edu.tw |
| 曹昱 Yu Tsao | 基於AI之生理訊號處理與自動化診斷系統 AI-based Biomedical Signal Processing for Computer-assisted Diagnosis | 本實驗室開發的語音訊號處理及降噪演算法在國際及國內均處於領先地位，同時也證明了此議題為一項前瞻且創新的研究。在應用上，我們針對開發出的演算法應用於口語溝通輔助科技上，相關臨床實驗結果也發表在數篇國際頂級論文、獲得中央研究院前瞻計畫獎、國家新創獎、World Invention Innovation Contest金牌獎等數項國內外知名獎項，並且獲得數篇國內外專利，證實此方向為一項獨特、創新且具潛力的研發方向。與其他人工智慧及生物醫學領域議題相比較，國際上在本實驗室專注的這個研究題目及產業相對較少，事實上我們覺得台灣在這個需要整合的議題具有相當優勢。目前我們已經開發了數項應用程式並實際驗證效能，獲得相當正向的結果。 The Bio-ASP Lab (Biomedical Acoustic Signal Processing Lab) in CITI, Academia Sinica was founded in November, 2011. We are dedicated to developing novel acoustic signal processing and artificial intelligence (AI) algorithms and applying them to biomedical and biology-related tasks. |
PI個人首頁(PI's Information) : http://www.citi.sinica.edu.tw/pages/yu.tsao/ 實驗室網址(Research Information) : https://bio-asplab.citi.sinica.edu.tw/ https:// Email : yu.tsao@citi.sinica.edu.tw |
| 高明達 Ming-Tat Ko | 臺語語音辨識、語音合成與轉換、國臺英語音機器互譯、臺語詞典暨語料庫的製作 Taiwanese Speech Recognition, Synthesis, and Conversion, Taiwanese-Chinese-English Speech Machine Translation, and Taiwanese Dictionary and Corpus Development | 我們呼召喜愛臺語，熱愛臺語的朋友一起來為臺灣的本土語言盡一份心力！ 我們會與王新民老師實驗室密切合作，引進最新語音技術： 1）製作完整、與時並進的國臺雙向詞典，可用於語音翻譯，讓臺灣人的語言溝通沒有距離。 2）將臺語語音辨識應用於目前尚未成形的臺語新聞、臺語鄉土劇的臺文字幕生成，不僅幫助年輕的一代學習、熟悉臺文，更可以服務諳臺語之聽障人士獲取資訊，支持政府的本土化政策。 3）目前廣為使用的 Google 小姐並不會講臺語， 我們將發展一套聽得順、聽得感動的臺語口語合成系統， 免費供應一同生活在這片土地上的人們使用。 As in the Chinese version |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/mtko/ 實驗室網址(Research Information) : https://homepage.iis.sinica.edu.tw/pages/mtko/index_en.html https:// Email : mtko@iis.sinica.edu.tw |
| 徐讚昇 Tsan-sheng Hsu | 以應用為中心的資料密集計算基礎研究 Intensive data computing foundations with applications | 資料密集計算的基礎研究 Motivated by applications, such as endgame databases construction, computer board game playing and simulations of spreading of infectious diseases, we plan to investigate fundamental problems involving massive data sets using methods such as algorithms, parallelization, implementation techniques and/or deep learning. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/~tshsu/ 實驗室網址(Research Information) : http://chess.iis.sinica.edu.tw/lab/ https:// Email : carol@iis.sinica.edu.tw |
| 呂俊賢 Chun-Shien Lu | 深度學習的安全與隱私 Security and Privacy in Deep Learning: Learning to Detect Adversarial Patch/Defake/Forgery and Learning to Encrypt with Homomorphic Encryption | see the introduction below More and more security and privacy issues gradually jeopardize the trustiness of deep learning models. The maliciously embedded adversarial patches can mislead the network functionality; the perfect but fake/forged images/videos created from deep learning models will create wrong meanings; and the privacy of learning models needs to be protected. Our goal is to study the Security and Privacy issues in Deep Learning and achieve Learning to Detect Adversarial Patch/Defake/Forgery and Learning to Encrypt with Homomorphic Encryption. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/lcs/ 實驗室網址(Research Information) : http://www.iis.sinica.edu.tw/pages/lcs/ https:// Email : lcs@iis.sinica.edu.tw |
| 黃文良 Wen-Liang Hwang | 深度神經網路最佳化 Optimization on Deep (recurrence) Neural Networks | 我們將會研究深度神經網路的最佳化方法，同時也會跟國立新加坡大學合作。 We will study methods on optimizations for deep (recurrence) neural network. This is a joint project with NUS, Singapore. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/whwang/ 實驗室網址(Research Information) : https://homepage.iis.sinica.edu.tw/pages/whwang/index_zh.html https:// Email : whwang@iis.sinica.edu.tw |
| 鐘楷閔 Kai-Min Chung | 密碼學、複雜度理論或量子密碼學之獨立研究 Independent Research on Cryptography, Complexity Theory, or Quantum Cryptography | The intern is expected to perform independent research on selected topics in (Quantum) Cryptography, Complexity Theory, or general theoretical computer science (TCS) that interest him/her. This often starts by surveying research papers and presenting it to the PI. Along the way, the intern can identify research questions with the PI, perform independent study on the questions, and discuss with the PI in research meetings. Candidate topics include, but not limited to, Quantum Key Distributions (QKD), Post-quantum Cryptography, Lattice-based Cryptography, Differential Privacy, Non-malleable Codes, Device-independent Cryptography, PRAM Cryptography, Zero Knowledge, Randomness Extractors, etc. The intern is expected to perform independent research on selected topics in (Quantum) Cryptography, Complexity Theory, or general theoretical computer science (TCS) that interest him/her. This often starts by surveying research papers and presenting it to the PI. Along the way, the intern can identify research questions with the PI, perform independent study on the questions, and discuss with the PI in research meetings. Candidate topics include, but not limited to, Quantum Key Distributions (QKD), Post-quantum Cryptography, Lattice-based Cryptography, Differential Privacy, Non-malleable Codes, Device-independent Cryptography, PRAM Cryptography, Zero Knowledge, Randomness Extractors, etc. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/kmchung/ 實驗室網址(Research Information) : http://www.iis.sinica.edu.tw/~kmchung/ https:// Email : kmchung@iis.sinica.edu.tw |
| 古倫維 Lun-Wei Ku | (1) 多模態問題生成 (看圖對話) (2) 假新聞免疫 (3) 運動科技-智慧教練 (1) Multimodal Question Generation (2) Fake News Immune (3) SportTech - AI Coach | 在這些研究主題中，將學習到自然語言處理之資訊擷取、文章分類、文字生成、知識庫使用、圖像文字結合等概念，另涵蓋自然語言基礎工具的使用及機器學習、深度學習的模型建立等先進技術，可與老師討論希望選擇的研究主題。實習期間會專注於上述研究主題並參與模型開發及論文撰寫。各主題研究內容詳述如下： (1) 在多模態問題生成專案中，我們注重在圖像(照片、影片)問題生成，以人類見到圖像時腦中建構的世界為概念，接著產生自然可開啟對話的問題。相關技術可應用於社群網站提高人氣，或是關懷病人與老人。 (2) 在假新聞免疫研究中，我們著重於研究甚麼樣的新聞內容與呈現形式，讀者會傾向於相信或不相信，我們將進行內容理解，網路模擬及使用者端的研究。 (3)運動科技-智慧教練中，我們希望開發對特定運動姿態的小樣本或無樣本學習模型，並經由圖像文字結合技術，自動生成智慧教練指導語。此研究目標為真正可用的系統。 實驗室尚有其他研究主題正在進行，可到 http://www.lunweiku.com/ 參考相關論文。 實習結束後，表現優良的同學可繼續與實驗室合作研究並發表論文。 Interns will learn how to use basic natural language processing tools, extract information from texts, classify documents, and generate dialogs. Machine learning and deep learning technologies for NLP will be touched. Interns can select the topic/team they wish to join. (1) In multimodal question generation project, we are focusing on the concept that human will construct the perception of a seen image sequence. Then we aim at generating a natural question to provoke the following pleasant conversations. This research is a topic extended from our series research on VIST (visual storytelling). The developed model can be used in social media platform or conversations with people who need mental care. (2) In fake news immune project, we focus on studying why and how readers trust fake news. We will explore approaches which mitigate the impact of fake news. Moreover, the goal is to provide a mechanism which can dynamically adjust the reading environment to increase the immunity to fake news. (3) In the sport technology project, we want to try applying few-shot or zero-shot learning on gestures/elements of the specific sport. Then we aim at generating automatic coaching instructions based on the videos of gestures/elements. (4) Interns can also choose to develop demo applications for the existing technologies in our lab. The research topics include but not limited to the above. After the internship, students with good performance can continue to work with the laboratory to research and publish papers. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/lwku/ 實驗室網址(Research Information) : http://academiasinicanlplab.github.io/ https:// Email : lwku@iis.sinica.edu.tw |
| 黃瀚萱 Hen-Hsen Huang | 知識庫中的反事實因果關係之建立與推理 Counterfactual Causal Analysis in Knowledge Bases | 知識圖譜由呈現事實的三元組構成，表達實體之間的關係。在典型的知識圖譜中，所有的事實都預設是可靠、真實的，但在現實上，知識圖譜很可能包含不確定的資訊，其中錯誤的事實甚至可能和其他事實互相衝突。在這個計畫中，我們預期將反事實知識引入知識圖譜，對不確定資訊進行反事實因果分析，藉以偵測與更正知識圖譜上不可靠的內容。除了可以確保知識圖譜的一致性，還可以進一步與深度學習模型的工作憶體整合，讓線上模型自動更正不實資訊。 Typical knowledge bases are composed of factual triples, representing the relations among entities with an assumption in mind that all the facts are true and reliable. In the real world, however, a knowledge base possibly contains uncertain information, and the untrue facts may be inconsistent with others mutually. In this project, our goal is to introduce a different kind of knowledge, the counterfactual knowledge, into knowledge bases to advance the causal analysis over the uncertain information. The results are expected to be not only useful for guarding the knowledge base integrity but also having the potential for misinformation correction in the machine's working memory. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/hhhuang/ 實驗室網址(Research Information) : http://www.iis.sinica.edu.tw/pages/hhhuang/ https:// Email : hhhuang@iis.sinica.edu.tw |
| 廖純中 Churn-Jung Liau | 應用邏輯 applied logic | 歡迎對符號邏輯與其應用有興趣的同學申請。 Our research is related to symbolic logic and its applications, including knowledge representation and reasoning. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/liaucj/ 實驗室網址(Research Information) : http://www.iis.sinica.edu.tw/pages/liaucj/ https:// Email : liaucj@iis.sinica.edu.tw |
| 林仲彥 Chung-Yen Lin | 以人工智慧來解析生物醫學大數據 Harness Biomedical Big data for Genome Biology in AI | 我們的團隊主要研究模式與非模式生物之多維基因體學（OMICS），包括基因體、轉錄體、單細胞轉錄體、蛋白質交互網路、腸道微生物與疾病關連等巨量資訊數據分析，同時也定序、重組與註解了多個重要經濟生物之基因體，目前致力於個人化基因體的重新組裝，探討不同程度的序列變異與疾病之間的關係，並利用人工智慧模型，來以全新的視角，來解析原有的生物醫學問題。我們是一個跨領域的研究團隊，成員來自資料科學、生物醫學與資訊技術等各類專業領域，歡迎不同背景（資訊、統計、數學及生物相關）的人才一起合作。研究範圍以單細胞基因解析、水生經濟動物基因體育種、精準健康老化、病原智慧分型、新型抗菌/抗病毒藥物的開發篩選與合成驗證、及利用人類腸道與環境微生物來進行人工智慧疾病預測等課題為主，同時發展新的高速計算工具及雲端分析平台，以及引入深度學習等策略，來探討基因、病原與環境的三角互動關係。 Our team's main goal is to analyze big omic data, which may lead us to know more about the secrets of biological regulations hidden among massive data deluge. By combining open source tools and self-developed programs/ platforms, we have assembled, annotated, and decoded several aquatic genomes with high economic importance. We are currently developing new approaches to fill the gaps in the assembled human genome to pave the way for personalized medicine and precision medicine. New approaches like deep learning will be introduced to rediscover our studies. Several platforms/ applications we developed in AI and biological knowledge are focus on smart typing of upper respiratory pathogens and novel antibiotics identification even design. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/cylin/ 實驗室網址(Research Information) : http://eln.iis.sinica.edu.tw https://hub.docker.com/u/lsbnb Email : cylin@iis.sinica.edu.tw |
| 陳駿丞 Jun-Cheng Chen | 深偽偵測與深度生成模型之影像應用 Deepfake Detection and Deep Generative Models for Image Applications | 隨著深度生成模型的快速發展，有心人士可以利用這些技術， 大量製造深偽影片和假消息，破壞他人名譽與社會信任，如日前網紅換臉事件，尤其深偽生成演算法不斷推陳出新，開發有效的深偽偵測演算法，已經是非常重要的研究議題‧ 要設計有效的深偽偵測器，更需要知道背後深度生成模型的背景知識‧ 在這個實習中，你可以接觸到各種深度生成模型在影像處理和電腦視覺的應用(如影像超解析度、去躁、風格轉換等應用)，當你對深度生成模型有一定了解後，我們手把手帶你設計一個深偽影像偵測器‧ 表現優秀的學生，實習結束後，可以獲得續聘兼任研究助理的機會繼續研究，並可發表論文至電腦視覺與影像處理國際知名會議或參加國際競賽為國爭光‧ With the fast development of deep generative model, malicious people can leverage these technologies to massively produce deepfake videos and fake news. These greatly defame others and destroy societal trust, such as the event of the net celebrity who replaces the faces of the videos with other celebrities. As the new deepfakes are invented very fast, it has become a very important research issue to design an effective deepfake detector. To achieve this goal, it is essential to understand the underlying knowledge of deep generative models. During the internship, we will guide you through various deep generative models and their applications to image super-resolution, denoising, and style transfer, etc, and finally we will design a deepfake detector together. For those who performs well during the internship, we will offer you the opportunity to be hired as a part-time research assistant and continue the research for famous international computer vision and image processing conferences or competitions. |
PI個人首頁(PI's Information) : http://www.citi.sinica.edu.tw/pages/pullpull/ 實驗室網址(Research Information) : http://www.citi.sinica.edu.tw/pages/pullpull/ https:// Email : pullpull@citi.sinica.edu.tw |
| 陳伶志 Ling-Jyh Chen | 以參與式環境感測系統為基礎的資料分析研究 Data Analysis using Fine-grained and Spatio-temporal Participatory Sensing Data | 在過去幾年中，我們已建立一個跨國性的大型細懸浮微粒(PM2.5)網路感測系統，擁有每天散佈在 59 個國家，超過 20,000 個 PM2.5 微型感測站，每個感測站以每五分鐘一筆的頻率上傳溫濕度與 PM2.5 的即時感測資料，目前已成為全球數一數二的 PM2.5 微型感測資料中心。 在這個專案中，我們希望延伸我們的研究觸角，探討跨感測器間的環境資料融合議題，同時整合新的環境感測設備，進行兼具學理、創意與應用價值的資料混搭與進階分析。內容可以是（但並不局限於）即時污染源的溯源、微型感測器的資料品質確保分析、環境感測與社經資源成本推估、環境感測與即時生理訊號的整合分析，甚或是其他更具創新與挑戰的研究議題。 我們歡迎對本項研究主題有興趣、有想法，並且願意接受挑戰的優秀人才加入我們的團隊，一同學習、努力、並對當前重大的環境議題做出貢獻。 Over the past few years, we have built a large-scale PM2.5 sensing system with more than 20,000 PM2.5 micro-sensing stations scattered in 59 countries. Each sensor station uploads real-time sensing data of temperature, humidity and PM2.5 at a frequency of one stroke every five minutes, and has now become one of the world's leading PM2.5 micro-sensing data centers. In this summer internship project, we hope to extend our research tentacles, discuss the fusion of environmental data across different sensors, and at the same time integrate new environmental sensing devices to conduct data mashup analysis with theoretical, creative and applied value . The content can be (but not limited to) the traceability of real-time pollution sources, data quality assurance analysis of micro-sensors, environmental sensing and socio-economic resource cost estimation, integrated analysis of environmental sensing and real-time physiological signals, or even other innovative and challenging research topics. We welcome outstanding talents who are interested in this research topic, have ideas, and are willing to accept challenges to join our team to learn, work, grow, and contribute to current major environmental issues. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/cclljj/ 實驗室網址(Research Information) : https://sites.google.com/site/cclljj/research https:// Email : cclljj@iis.sinica.edu.tw |
| 林仁俊 Jen-Chun Lin | 基於注意力機制的鏡頭分類 Shot classification based on attention mechanism | 專業攝影師通常在拍攝人像時會使用合適的鏡頭類型(例如，中長鏡頭)來使拍攝到的人像更具美感。然而，民眾在拍攝照片時卻往往缺乏此類人像構圖的概念與技巧。為此，在這個計畫中，我們旨在創建鏡頭分類的深度學習技術，並希望將開發後的深度神經網路整合至手機App上，以幫助大眾能拍攝出更好的人像照片。 實習生預計從鏡頭分類或其他相關主題中選定題目進行獨立研究。實習結束後，若同學表現優良則可繼續與實驗室合作研究並發表論文。 Professional photographers usually use the appropriate shot type (e.g., medium long shot) to make the filmed photo more touching. Amateurs, however, often lack the skills and knowledge of such composition when taking photos. To this end, this project aims to create a novel deep-net model for shot classification that can be integrated into a mobile app to help people take better portrait photos. Interns are expected to conduct independent research on selected topics from shot classification or other related topics. After the internship, students with good performance can continue to work with the laboratory to research and publish papers. |
PI個人首頁(PI's Information) : https://homepage.iis.sinica.edu.tw/pages/jenchunlin/index_zh.html 實驗室網址(Research Information) : https://sites.google.com/site/jenchunlin/ https:// Email : jenchunlin@iis.sinica.edu.tw |
| 蔡懷寬 Huai-Kuang Tsai | 生物資訊 bioinformatics | 近年來，生物資訊在醫學領域受到相當大的關注，不僅是定序技術的進步，或是在數據分析的能力也持續在更新。本實驗室除了與國內許多生物試驗單位合作外，更從世界知名大型研究機構取得資料，以大數據分析解開生物中複雜的調控機制與原理。 我們的研究方向著重於真核生物的基因體，並整合多種體學資料 (multi-omics) ，以不同的生物觀點，更具系统性的探討基因體層級的交互關係，與其在演化上的重要性。而我們最新的研究主題著重在人類重大疾病與大數據資料庫整併，藉由不同的序列資料尋找目前生物醫學上未解出的困境。我們的研究方法透過資料探勘和機器學習來建立分析模型，用來預測生物的基因調控，同時達到個人化精準醫療的開發及癌症預測應用。 本實驗室想要找對生物資料運用感興趣的大專生,你可以來自資工或生物背景，但應該熟悉至少一種程式語言及對生物學感興趣。我們會提供生物資訊學相關領域的知識訓練，因此只要您對於跨領域研究感興趣，也想解決目前生物領域面臨的瓶頸，歡迎您加入我們的研究團隊。 The Tsai lab studies big data from biological systems using bioinformatic techniques and statistical methods. We work with biologists to seek insights into the genomics of eukaryotic organisms. By integrating multi-omics data, we study genome-wide regulatory systems on gene expressions and their significance in evolution. In addition, we are currently expanding into the area of biomedical informatics, aiming at integrating disease information with sequencing data for development of applications in precision medicine. We use methods such as data mining and machine learning in our studies on regulatory mechanisms in genomics, with the aim of building predictive models with potentials for applications. We are seeking interns with a background in either computer science or biological science. The applicant should have experience in using at least one programming language and have a strong interest in biology. We will provide training in bioinformatics-related domain knowledge, and we expect our interns to be able to learn from team members from different backgrounds. If you are passionate in taking up the challenge of solving biological problems with techniques in informatics, we welcome you to join our team! |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/hktsai/ 實驗室網址(Research Information) : https://bits.iis.sinica.edu.tw/?id=1 https:// Email : hktsai@iis.sinica.edu.tw |
| 蘇黎 Li Su | 以人為核心的音樂人工智慧 Human-centered music AI | 音樂與文化科技實驗室致力於探討最前沿的數位訊號處理與深度學習技術，應用於各種音樂人工智慧的熱門議題。我們特別關注以人為核心的音樂人工智慧研究，包括但不限於： 1. 音樂內容辨識與理解、自動採譜 2. 音樂與多媒體內容生成與互動 3. 計算音樂學 4. 深度學習、認知科學與生物音樂學 5. 音樂人工智慧技術之應用與評估方法 我們歡迎資訊/電機等理工相關科系或音樂相關科系背景，或有志於跨領域研究之同學應徵。 熟悉深度學習、音訊處理、影像處理、電腦圖學、人機互動、UI/UX、認知科學、音樂學等任一領域者優先考慮。 The mission of the Music and Culture Technology Lab is to solve cutting-edge research topics on music AI using novel deep learning and signal processing research technologies. We specifically concentrate on of human-centered music AI, including (but not limited to) the following research directions: 1. Music content understanding and automatic music transcription 2. Music and multimedia content generation/interaction 3. Computational musicology 4. Deep learning, cognitive science and biomusicology 5. Application and evaluation method of music AI technology We welcome students with background in EE/CS, musicology and related fields and who are interested in inter-disciplinary research to join our intern program. The students who are familiar to deep learning, signal processing, image processing, computer graphics, HCI, UI/UX, cognitive science, or musicology will be considered with first priority. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/lisu/ 實驗室網址(Research Information) : https://mctlab.iis.sinica.edu.tw/mctl/ https://github.com/Music-and-Culture-Technology-Lab Email : lisu@iis.sinica.edu.tw |
| 劉庭祿 Tyng-Luh Liu | 新興電腦視覺與深度學習技術 Emerging Computer Vision and Deep Learning Techniques | Current research topics: 1. Vision transformer related research 2. Computer vision techniques for point clouds 3. Action detection and recognition 4. Deepfake detection, anomaly detection 5. Unsupervised, semi-supervised representation learning 6. Few-shot, zero-shot, meta learning 7. Federated learning Current research topics: 1. Vision transformer related research 2. Computer vision techniques for point clouds 3. Action detection and recognition 4. Deepfake detection, anomaly detection 5. Unsupervised, semi-supervised representation learning 6. Few-shot, zero-shot, meta learning 7. Federated learning |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/liutyng/ 實驗室網址(Research Information) : https://homepage.iis.sinica.edu.tw/~liutyng/ https:// Email : liutyng@iis.sinica.edu.tw |
| 洪鼎詠 Ding-Yong Hong | 深度學習軟體與硬體協同優化研究 Deep Learning Software/Hardware Co-optimization | 我們將研究深度學習軟體與硬體協同優化方法。(1) 研究如何利用編譯器技術, 優化深度學習程式, 使其在AI加速器上達到最佳的運算效能。(2) 根據AI加速器提供的特別運算與加速功能, 重新設計深度學習模型architecture/pruning/quantization/parallelization的優化方案。 We aim to study hardware/software co-optimization for deep learning workloads. (1) Exploiting compiler techniques to accelerate deep learning applications on AI accelerators. (2) Exploiting AI accelerator's special acceleration instructions to re-design AI models for better architecture/pruning/quantization/parallelization/resource utilization. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/dyhong/ 實驗室網址(Research Information) : http://www.iis.sinica.edu.tw/pages/dyhong/ https:// Email : dyhong@iis.sinica.edu.tw |
| 王大為 Da-Wei Wang | 醫療資料分析與機器學習技術應用 Medical data analysis using machine learning technology | 研究主題為資料分析與機器學習技術在醫療領域的應用，包含：自動語音辨識 (Automatic Speech Recognition, ASR)、結構化資料分析等。 The research area is using statistical data analysis methods and machine learning technology in the medical field. The study topic can include automatic speech recognition (ASR) and structured data analysis. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/wdw/ 實驗室網址(Research Information) : http://chess.iis.sinica.edu.tw/lab/ https:// Email : |
| 穆信成 Shin-Cheng Mu | 函數編程與程式正確性推理之相關問題 Functional Programming and Reasoning about Programs | 我的研究興趣是程式語言與函數編程，近年來也包括 concurrent 程式的型別系統與 Hoare logic. 它們的共同點是使用符號推理的方式確保程式的正確性。不論在哪個典範中，我們都希望把寫程式視作一個可用數學與邏輯方式推理的行為。程式的正確性可用型別系統或邏輯推演保證，甚至可用規格與需求開始，經由數學方法一步步推導出程式。 本領域可做的大方向包括 * 設計幫助推理用的符號、程式語言、型別系統等。 * 挑選一些演算法問題，嘗試以數學方法實際證明演算法之正確性，或將演算法推導出來。 * 研究 concurrent 程式以及其型別系統 (session type) 與邏輯之關係。 * 以函數編程語言為工具，開發 Hoare logic 與指令式語言編程課程使用的教學系統。 如對以上題目有興趣，在三個月的實習期間，我們可用一到一個半月的時間學習相關理論（函數編程、型別、邏輯等），用剩下的時間研究新東西或開發系統。 My research interest concerns programming language and functional programming, and extends to Hoare logic and type systems for concurrent programs. The common theme is that programming is seen as a formal, mathematical activity. Correctness of a program can be guaranteed by logical reasoning or type system. Or, a program can even be derived stepwise from its specification. Possible topics include: * design symbols, languages, or type systems that aids the programmers in reasoning about programs; * pick an algorithm, and apply our approaches to prove its correctness or even to derive an algorithm; * study the type system (session type) for concurrent programs and its relationship with logic; * develop tools for teaching Hoare logic and reasoning of imperative programs, using a functional programming language. More details can be discussed. If you are interested, we can spend the first 1 to 1.5 months of the internship studying the background knowledge, before diving into developing something new. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/scm/ 實驗室網址(Research Information) : https://www.iis.sinica.edu.tw/zh/page/ResearchGroup/ProgrammingLanguagesandFormalMethods.html https://scm.iis.sinica.edu.tw/ncs/ Email : scm@iis.sinica.edu.tw |
| 黃彥男 Yennun Huang | AIoT/入侵偵測/隱私保護/智慧農業系統開發及AI資料科學分析 AIoT/Intrusion Detection/Data Privacy/Smart Agriculture system development and AI scientific data analysis | 1.進階持續性滲透攻擊偵測研發 - 偵測系統開發與測試 - 系統稽核紀錄和網路封包分析 - 人工智慧框架使用 - 圖論與圖資料庫應用 - 滲透測試工具學習 2.智慧物聯網 a.物聯網系統 -嵌入式系統開發 -藍芽手環開發 -全端系統開發 b.微機器學習 -智能移動偵測 -邊緣運算開發 3.資料安全系統程式開發 - 將從事資料安全、隱私保護之Deep Learning, Federated Learning等相關領域研究工作 - Python、Linux、Matlab、Deep Learning Toolbox基礎程式撰寫能力 - 聯邦式學習、資料去識別化保護技術相關研究 4. 智慧農業計畫 - 工作內容為資料蒐集、彙整、分析 - 使用機器學習方法訓練模型 - 資料科學相關 1. Intrusion Detection Development - detection system development and testing - system audit logs and network packets analysis - AI framework usage - graph theory and graph database application - penetration testing tools learning 2. AIoT a. IoT system -Embedded System development -Bluetooth Bracelet Development -Full Stack Development b. tinyML -Smart Motion Detection -Edge Computing Development 3. Data science system program development - Over the research work for data privacy with deep learning and Federated learning - Basic programming ability for Python, Linux, Matlab, Deep Learning Toolbox - Research on data privacy technology 4. Smart Agriculture Project - Work includes data collection, compilation, and analysis - Use machine learning methods to train models - Data science related |
PI個人首頁(PI's Information) : http://www.citi.sinica.edu.tw/pages/yennunhuang/ 實驗室網址(Research Information) : http://www.citi.sinica.edu.tw/pages/yennunhuang/ https:// Email : joannamary@citi.sinica.edu.tw |
| 陳孟彰 Meng Chang Chen | 以深度學習與系統稽查紀錄來分析惡意軟體行為 deep learning for malware dynamic analysis with audit logs | 本計畫將使用深度學習方法以及使用系統稽查黨來分析惡意軟體的行為與方法。歡迎對深度學習與(或)系統與網路安全的同學來申請。我們最近的工作請參考: Yi-Ting Huang, Chi Yu Lin, Ying-Ren Guo, Kai-Chieh Lo, Yeali S. Sun, and Meng Chang Chen, "Open Source Intelligence for Malicious Behavior Discovery and Interpretation," to appear in IEEE Transactions on Dependable and Secure Computing (Special Issue on Explainable AI for Cyber Threat Intelligence Applications), SI-Explainable Artificial Intelligence for Cyber Threat Intelligence (XAI-CTI) Application 以及 Shun-Wen Hsiao, Yeali Sun, Meng Chang Chen, "Hardware-Assisted MMU Redirection for In-guest Monitoring and API Profiling," IEEE Transactions on Information Forensics & Security, volume 15, pages 2402-2416, January 2020. This summer intern project will investigate malware behaviors and attacks by using the deep learning method with system audit logs. Students interested in deep learning and/or system/network security are cordially welcome to apply. For our recent work, please refer to Yi-Ting Huang, Chi Yu Lin, Ying-Ren Guo, Kai-Chieh Lo, Yeali S. Sun, and Meng Chang Chen, "Open Source Intelligence for Malicious Behavior Discovery and Interpretation," to appear in IEEE Transactions on Dependable and Secure Computing (Special Issue on Explainable AI for Cyber Threat Intelligence Applications), SI-Explainable Artificial Intelligence for Cyber Threat Intelligence (XAI-CTI) Application and Shun-Wen Hsiao, Yeali Sun, Meng Chang Chen, "Hardware-Assisted MMU Redirection for In-guest Monitoring and API Profiling," IEEE Transactions on Information Forensics & Security, volume 15, pages 2402-2416, January 2020. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/mcc/ 實驗室網址(Research Information) : http://www.iis.sinica.edu.tw/pages/mcc/ https:// Email : mcc@iis.sinica.edu.tw |
| 楊得年 De-Nian Yang | 元宇宙之多媒體社群深度學習 Multimedia, Social Networks, and Deep Learning in the Metaverse | (一)社群資料探勘、機器學習、與演算法設計： • 基於虛擬、擴增實境(VR/AR)或元宇宙(metaverse)的推薦系統：如規劃避免3D暈眩或撞到障礙物的虛擬和現實路徑、團體消費中最大化共同和個人的喜好與虛擬世界社群網路中NFT交易推薦系統。 • 社群影響力分析與優化：如選定優惠券最佳兌換率投放目標、建立利益衝突迴避的審查人員團體與圖(graph)中子結構的資訊融合。 • 其他應用領域的推薦系統：如社群直播組合推薦、為團體活動安排與活動潛在的參與者推薦。 (二)多媒體網路演算法設計與分析： 藉由分析問題NP困難度及不可近似性的方法，以及高階演算法設計技巧，來解決多媒體網路中的各類應用問題 • 虛擬實境和元宇宙網路：如規劃有線及無線網路資源配置和排程方式、選定3D多視角影片所需傳輸及合成之場景和最佳化多媒體網路資源效率及確保使用者的沉浸體驗。 • 軟體定義網路中的各類效能優化問題：如在具有網際網路工程任務之動態群組，將最小化總頻寬消耗和更新規則的總數，並確保線路／節點容量。 • 行動邊緣運算情境：如藉設計高階演算法建置高效、可靠的社群物聯網(Social IoT)。 元宇宙是一個整合了多重虛擬世界的生態系，人們能透過其化身(Avatar)在元宇宙中社交、購物和創作，現實世界的物品也能透過數位分身(Digital Twin)的形式存在其中。長期以來，我們關心未來元宇宙中的各種社群網路問題，包括虛擬與擴增實境(VR/AR)的朋友或NFT推薦系統、即時串流平台推薦系統、社群影響力分析及社群資料探勘；此外，我們也關心下個世代的網路優化問題，包括多媒體與軟體定義網路效能優化、有線及無線網路資源配置、單播/群播排程方式和建置可靠的社群物聯網(Social IoT)。在這裡，你可以學到的技術包括圖神經網路(Graph Neural Network)、機器學習、張量分解技術、分析問題NP困難度及不可近似性的方法、整數/線性/半正定規劃、動態規劃、隨機湊整、對偶理論、抽樣方法等高階演算法設計技巧。歡迎想出國留學、增強實作能力、有元宇宙創業憧憬的同學，於今年夏天加入我們，一起探索未來元宇宙與社群物聯網的無限可能。 A. Social network data mining, machine learning, and algorithm design: Research tensor decomposition, neural network, machine learning, and other technical solutions for: • Virtual, augmented reality (VR/AR) or metaverse recommendation system (e.g., user-item configuration recommendation, planning a path avoiding 3D motion sickness and obstacles, and virtual world social network in NFT markets). • Social influence analysis and optimization (e.g., coupon allocation for redemption maximization, reviewer recommendation, and fusing graph substructures information into node features). • Recommendation systems for other applications (e.g., social live streams recommendation system, group activities arrangement, and potential customers recommendation). B. Algorithm design and analysis for multimedia and software-defined network: We analyze NP-hardness, design approximation algorithms, and use advanced algorithm techniques to solve problems in various multimedia and software-defined networks. • Virtual reality (VR) and metaverse applications (e.g., design resource allocation methods for wireless/wired networks and scheduling algorithms, select the required scene to synthesize in 3D multi-view videos, and optimize the resource use efficiency and users' immersive experience in multimedia networks). • Various performance optimization problems in software-defined networks (e.g., minimize the total bandwidth consumption and the total number of update rules and ensure line/node capacity in dynamic groups with Internet engineering tasks). • Mobile edge computing scenarios (e.g., build high-performance and reliable social IoT applications). Welcome those who plan to study abroad and enhance their implementation skills or are interested in the metaverse. Please join us this summer. Let's explore the opportunities of the future metaverse and social IoT. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/dnyang/ 實驗室網址(Research Information) : http://www.iis.sinica.edu.tw/pages/dnyang/ https:// Email : dnyang@iis.sinica.edu.tw |
| 張佑榕 Ronald Y. Chang | 5G/6G無線通訊與應用 5G/6G Wireless Communications and Applications | 見英文介紹 Several paid internships are available at CITI, Academia Sinica. The intern will participate in one of the following projects: 1) intelligent reflecting surfaces (IRSs) aided sixth generation (6G) wireless communications (reference: https://arxiv.org/pdf/2109.09267 ). 2) sensor-based satellite Internet of Things (IoT) for wildfire detection (reference: https://arxiv.org/pdf/2109.10505 ). 3) machine learning for wireless communications and IoT applications. 4) space and cellular network integration. Intern responsibilities include attending group meetings, conducting research at a similar level as full-time RAs, and preparing research reports, slides, presentations, and/or research papers. Extensions (with pay) after the official two-month internship are possible upon demonstration of satisfactory performance. Preferred qualifications: 1) EE/CS/Communication Engineering major or a related area; 2) Strong knowledge of wireless communication, signal processing, and/or machine learning; 3) Good programming skills; 4) Plans to pursue advanced study domestically or abroad. |
PI個人首頁(PI's Information) : https://www.citi.sinica.edu.tw/pages/rchang/index_en.html 實驗室網址(Research Information) : https://www.citi.sinica.edu.tw/~rchang/ https:// Email : rchang@citi.sinica.edu.tw |
| 王建民 Chien-Min Wang | 雲端運算與人智運算 Cloud Computing and Human-Centered Computing | (1) 整合記憶體內資料儲存的雲端計算平台：MapReduce是目前利用雲端計算來處理巨量資料方面，最常用的平行計算模型。然而我們發現有一類雲端應用，雖然非常適合MapReduce模型，但是其執行效能卻非常低落，而且計算規模也有很大的限制。這類應用包括用於基因定序的後綴陣列排序和近來很受重視的演化式計算。我們將對現有的Hadoop平台進行擴充和改進，融合記憶體內資料儲存，提出一個泛用的加強型雲端計算平台，以提升執行效能和規模擴充性。我們也將實作後綴陣列排序以及演化式計算，以驗證我們所提出的架構，對於這兩種應用的執行效能和應用規模有多大的提升。我們相信這樣的雲端計算平台不但對於學術研究有很大的貢獻，還能大幅拓展雲端計算平台的應用。 (2) 使用遺傳式編程探究監督式機器學習：本研究計畫透過嘗試解決兩個不同需求的應用問題，來探討監督式機器學習的兩個不同階段。不同於時下熱門的深度學習方法使用類神經網路模型和倒傳遞式訓練，本研究計畫探索機器學習的另一種可能性與方向，也就是遺傳式編程(Genetic programming)。其使用數學表達式模型和演化式搜尋學習，有益於機器學習結果的理解、推導與運用，符合Explainable AI所提倡之概念。要完成本計畫的目標，預期需探討的主要研究議題包括：適應(目標)函數設計、學習(演化)運算子新增與修改、觀察樣本資料處理、平行化或GPU加速、以及效能測試與方法驗證等。 (3) 人智運算的穿戴運算系統：研究穿戴式電腦及裝置在人智計算中的應用，特別是在社交網路方面的應用。我們計劃中的人智運算系統應具備的三種能力：具有瞭解周遭環境與人們情況的能力，可提供使生活更美好的服務，和透過感官與人類自然地互動。為了實現這三個能力，我們計劃中將從三個研究學科來發展：情境識別，雲服務，以及擴增實境。藉由研究相關的穿戴式電腦及裝置，開發更佳的人機整合功能，並透過社交網路系統之系統分析，研究並開發穿戴式社交網路系統，以提升使用者經驗為目標，並且提供更適合的情境感知技術與實境服務的增強實境功能。 (1) A MapRedice framework with an In-memory Data Store: MapReduce is a powerful programming model for processing large data sets with a parallel, distributed algorithm on clouds. The Hadoop framework is the most popular implementation of MapReduce and widely adopted in the processing of large datasets. However, our previous experience on suffix array construction with Hadoop shows that it might result in excessive disk usage and access. Therefore, the performance is degraded and the scale of the application is limited. In this project, we aim at efficient and scalable processing of expansive MapReduce (EMR) applications with in-memory data stores. EMR applications, including suffix array construction and evolutionary computation, are a group of applications that have performance and scalability issues with Hadoop. We shall integrate an in-memory data store with Hadoop and propose a MapReduce framework for EMR applications to enhance their performance and scalability. To validate the benefit of the proposed framework, we shall use suffix array construction and evolutionary computation as our testbed. (2) Exploring supervised machine learning with genetic programming: Instead of adopting the widely used deep learning techniques, this project aims at another possibility and research direction, i.e., genetic programming, which employs evolutionary searching/learning strategy and mathematical expression-based models that are helpful for the understanding and use of the outcomes of machine learning process. To reach our goal, the research issues that need to be carefully addressed include the design of fitness function, the invention and modification of evolutionary operators, data processing for observation samples, acceleration with parallel or GPU computing technologies, and performance testing and validation. (3) Wearable Computing Systems and Applications in Human-Centered Computing: The goal of this project is to investigate the application of wearable computers and devices in Human-Centered Computing, especially those applications on social networks. A human centered computing system should have three abilities: understanding the context of the surrounding area and humans, providing the service that makes the lives better, and interacting with human naturally through perception. To realize these three abilities, we plan to adopt three corresponding research disciplines: context recognition, cloud service, and augmented reality. Wearable computers and social network services will be integrated to build the proposed wearable social network system. The proposed system will provide more convenient and user-friendly human-computer interaction. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/cmwang/ 實驗室網址(Research Information) : http://www.iis.sinica.edu.tw/page/research/ComputerSystem.html?lang=zh https:// Email : cmwang@iis.sinica.edu.tw |
| 吳真貞 Jan-Jan Wu | 深度學習在異質系統架構中之效能優化 Optimizing Performance of Deep Learning on Heterogeneous System Architecture | 近年來,將多種neural network模型結合起來以提高深度學習能力的趨勢日益增加，此稱為複合式神經網路模型(hybrid neural network model)。例如，許多應用程序將 CNN 和 RNN 結合起來進行視頻字幕，視頻問題解答，自動醫療報告生成，股票交易分析，電影評論分析和污染物預測。隨著越來越多的AI應用程式採用複合式模型，優化複合式模型的執行以縮短推理時間已成為一項及時而關鍵的研究課題。此外, CPU+GPU 異構系統架構是現代計算機中的常見架構。目前常見的運算方式是在GPU上同時運行CNN和RNN, 此GPU-only運算方式未能充分利用CPU + GPU異構系統架構所提供的計算能力而導致較長的推理時間。 此外, 許多新型AI應用, 例如推薦系統, 知識圖譜等, 使用GNN/GCN作為深度學習訓練與推理的網路模型. GNN/GCN包含較複雜的不規則計算行為以及大量的稀疏矩陣(sparse matrix)計算. 這些計算在傳統GPU不易獲得良好執行效能. 然而近年CPUs提供強大的向量指令(例如Intel AVX512 向量指令可同時計算8個64-bit資料) , 其gather/scatter指令可快速存取非連續記憶體位址資料, 為不規則計算與sparse matrix計算開啟新契機. Tensor core GPU 也為sparse matrix計算做特殊硬體優化, 在稀疏度為50%時可達到兩倍加速. 如何運用AI compiler技術以及優化演算法設計使GNN/GCN等複雜模型充分利用向量指令或 Tensor core GPU的硬體優勢以達最佳運算效能亦是極具挑戰性的研究議題. 本實驗室研究方向為：(1) 透過AI編譯器(例如TVM 和MLIR)的優化技術,並配合資源配置和排程演算法設計, 研究如何利用異質運算平台(heterogeneous platform)上多CPUs、多GPUs、以及CPU+GPU+AI加速器等運算環境，提高深度學習模型(特別是複合式模型)的執行效能。(2) 使用MLIR AI compiler framework發展一系列GNN/GCN 優化技術 並實作於AVX512 + GPU + Tensor core之異質系統架構. The future computing environment would contain several levels of computing units: master CPUs, secondary CPUs, heterogeneous AI accelerators, GPUs, and/or FPGAs. We believe that the key to a successful deep learning implementation relies on the smart interaction between and effective utilization of the computing resources in the environment. To solve this issue, we plan to model it as a graph problem, since the artificial neural network is typically a computational graph. So, the problem becomes: given an objective such as maximizing throughput or lowering power consumption, (1) how to partition the computational graph, and (2) how to schedule the computation to the computing resources. For the graph of complex networks (e.g., hybrid models and multi-models), two data dependencies can exist: one is between the operations within the same iteration (i.e., at the same timestamp), and one is between the operations across different iterations (i.e., through time). For the dependency within the same iteration, we will investigate different graph partitioning granularities, including coarse-grained (i.e., model-level), fine-grained (i.e., operation-level), and specific-grained (i.e., according to the compute engine types). For the dependency between different iterations, we will investigate different parallelization techniques, such as pipelining and loop skewing, to schedule the computations from different iterations to run in parallel. By mapping the computations to their most suitable acceleration devices and overlap the computations to hide execution latency, we believe significant performance can be achieved. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/wuj/ 實驗室網址(Research Information) : https://www.iis.sinica.edu.tw/zh/page/ResearchGroup/ComputerSystem.html https://www.iis.sinica.edu.tw/zh/page/ResearchGroup/ComputerSystem.html Email : wuj@iis.sinica.edu.tw |
| 蔡孟宗 Meng-Tsung Tsai | 串流式圖論演算法 Graph Streaming Algorithms | 我的研究興趣在探討如何使用 O(n) 的記憶體空間處理各式的圖論計算問題，這裡 n 是指輸入圖的節點個數。 我們假設圖的邊是按照某個最糟的順序一條一條給演算法，而且只給一次。一張 n 個節點的圖，最多會有 Ω(n^2) 條邊，因為限制只能使用 O(n) 的記憶體空間，勢必得強迫演算法 "忘記" 大部分曾經讀進來的邊。在這個前提下，如何設計演算法完成各式的圖論計算問題？ 在這嚴格的限制下，或許心裡的第一問題是："是否大部分的圖論問題都不能在使用 O(n) 記憶體的狀況下完成計算？" 目前的研究文獻已經證實，許多圖論計算問題可以，但也有許多圖論計算問題，保證無法在這限制下計算出來。後者的情況，常常能找到方法在使用少量的空間下，找到 (1) 不錯的近似解、(2) 具有隨機成分的最佳解 (在很高的成功機率下)、或 (3) 具有隨機成分的不錯的近似解 (在很高的成功機率下)。這邊的機率與輸入的圖無關，只和演算法使用的隨機成分有關。 近期實驗室的研究成果有： 1. 存在一般圖上的 NP-complete 圖論計算問題，可以使用 O(n) 空間回答！ 2. 對於將輸入圖拆分成盡可能少的無環子圖這個圖論計算問題，任何演算法都需要 Ω(n^2) 的記憶體空間才能找到最佳的拆分法！但存在演算法，只要 O(n) 的記憶體空間就能找到近似於最佳解的拆分法。 在這個專題，我們預期可以學習到如何使用數學工具回答："在侷限的記憶體空間下，有哪些圖論計算問題可以被解決？有哪些圖論計算問題保證無法被解決？以及你喜歡的圖論計論問題是屬於哪一類？" We are interested in whether a graph problem can be computed using O(n) space, where n denotes the number of vertices in the input graph. We assume that the edges of the input graph are given to algorithms one by one, in an arbitrary order, and only once. Note that an n-vertex graph may have Ω(n^2) edges. If an algorithm uses O(n) space, then it has to "forget" much information of the input. Given the restriction, can we design algorithms to solve graph problems? One may wonder whether there are many problems that can be solved using little space. It has been shown in the literature that: dozens of graph problems can be solved using little space, while dozens of graph problems cannot. In the latter case, the community usually can come up with a solution that approximates the best possible to within some factor, a solution that matches an optimal one with high probability, or a solution that approximates the best possible to within some factor with high probability. The probabilities here depend only on the randomness used in algorithms, and do not depend on the input graph. The recent results obtained by our lab include: 1. There exists some NP-complete graph problem on general graphs that can be computed using O(n) space! 2. For any streaming algorithm, decomposing a graph into the least number of acyclic subgraphs requires Ω(n^2) space. However, this problem can be well approximated using O(n) space. In this independent study, we expect to learn how to apply mathematical methods to answer the questions: whether a graph problem can be solved using little space, and which category your favorite graph problem belongs to? |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/mttsai/ 實驗室網址(Research Information) : http://www.iis.sinica.edu.tw/pages/mttsai/ https:// Email : mttsai@iis.sinica.edu.tw |
| 鄭湘筠 Hsiang-Yun Cheng | 適用於資料密集程式之新世代記憶體系統設計 Energy-efficient future memory systems for data-intensive applications | 近年來資料密集程式，像是深度學習、圖論分析、基因序列分析等，越來越盛行，這些資料密集程式在運算時往往需要大量的記憶體存儲空間與高效的資料存取，然而目前主流的運算系統無法滿足這些需求, 使得我們必須重新思考如何設計未來的電腦系統。 一個亟具潛力的設計方向是從傳統運算單元為主的系統切換到記憶單元為主的系統，在記憶體內直接做部分的運算減少資料傳輸造成的額外耗時與耗能。許多新興的記憶體技術，像是 Intel 的Optane Memory、電阻式記憶體(ReRAM) 等，具備高密度低漏電之特性且兼具存儲與運算功能，為實現以記憶體為中心的運算系統帶來新的曙光，但由於硬體技術尚不成熟以及和傳統截然不同的運算模式，在系統設計上有許多尚待克服之挑戰。 本實習計劃的目標為針對資料密集程式之應用情境，探討不同層面上之設計挑戰，包括電路與元件階層、計算結構階層、及演算法階層，並以軟硬體協同設計的方式， 設計高效能低耗電之新世代記憶體系統。實習生可選擇參與下列研究主題，或其他相關研究議題。 1. 設計基於記憶體內運算的深度學習加速器，並從軟體與硬體層面優化提升效能。 2. 設計適用於加速圖論分析演算法之新興記憶體系統。 3. 基於記憶體內運算，以軟硬體協同設計提升基因序列分析演算法之效能。 In recent years, data analytics applications that must process increasingly large volumes of data, such as deep learning, graph analytics, genome data analytics, etc., have become more and more popular. These big data applications demand large memory capacity and efficient data access. Unfortunately, mainstream computing systems with DRAM-based main memory are not designed to meet their needs. This forces us to fundamentally rethink how to design future computing platforms. One promising solution is to shift from contemporary processor-centric design towards revolutionary memory-centric design. Emerging memory technologies, such as Intel's Optane Memory, resistive RAM, etc., offer superior density, non-volatile property, and computing-in-memory capability. These promising features enable them to open up new opportunities to build future in-memory computing systems for big data applications. Despite being promising, bringing such a system into practice remains challenging due to the hardware constraints and the distinct computing behavior. Our goal is to study the design challenges at different system layers, including device/circuit level, architecture level, and algorithm level, and propose cross-layer designs to fully exploit the potential of in-memory computing systems. Candidate topics include, but are not limited to, the following: 1. Cross-layer co-design to improve the reliability and energy efficiency of deep learning algorithms via in-memory computing. 2. Cross-layer co-design to accelerate graph analytic algorithms via in-memory computing. 3. Cross-layer co-design to accelerate genome data analytics via in-memory computing. |
PI個人首頁(PI's Information) : http://www.citi.sinica.edu.tw/pages/hycheng/ 實驗室網址(Research Information) : http://www.citi.sinica.edu.tw/pages/hycheng/ https:// Email : hycheng@citi.sinica.edu.tw |
| 楊柏因 Bo-Yin Yang | 後量子密碼學 postquantum cryptography | 我們探討後(抗)量子密碼學，即在中大型量子電腦來到後仍可使用的公鑰密碼系統。你有機會一起來改變世界！ We investigate postquantum (quantum-resistant) cryptography, which is public-key cryptosystems that remains secure after the arrival of medium to large scale quantum computers. Come, and you have the chance to change the world! |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/byyang/ 實驗室網址(Research Information) : http://troll.iis.sinica.edu.tw/by-publ https:// Email : byyang@iis.sinica.edu.tw |
| 莊庭瑞 Tyng-Ruey Chuang | 1. 相互豐富的研究資料與維基資料 2. 創新以及可永續的研究資料管理和協作 1. Mutual Enrichment between Research Data and Wikidata 2. Innovative and Sustainable Research Data Management and Collaboration | 1. 開放研究資料已不再是新的口號。從 Open Data 到 FAIR Data 有各種倡議與原則，但面臨研究實務上的議題時，針對不同學科領域的研究資料，存在著不同程度的想像空間及挑戰。好的研究，始於好的資料。除了使用 Wikidata 做為「研究資料寄存所」 (網址： https://data.depositar.io/about ) 的資料集關鍵字來源，以加強資料集之間的語意連結之外，本實驗室也陸續嘗試與不同學科領域的研究夥伴，進行各種資料的爬梳及結構化的處理。 2. 我們將從社群 (community)、技術 (technology)、協作 (collaboration)、以及研究 (research) 四面向， 協力發展台灣本地在研究資料管理的實踐社群。此實踐社群將以我們已開發的「研究資料寄存所」（網址： https://data.depositar.io/about ）為實踐的場域之一。本計畫的預期成果包括：培養研究資料管理人才、參與開放資料軟體系統的國際協作專案、提昇研究資料管理實踐社群在台灣的規模與內涵、以及參與並貢獻所能到全球研究社群。 更多資訊可參閱以下： 1. Improving data discovery through Wikidata - WikidataCon 2019 https://commons.wikimedia.org/wiki/File:Improving_data_discovery_through_Wikidata_-_WikidataCon_2019.pdf 2. Open Repositories for Scholarly Communication and Participatory Research https://m.odw.tw/u/trc/m/rda-p18-panel/ 1. We will study WiIkidata, and use Wikidata to enrich research datatsets, and vice versa. We will study and use Graph DB, for example TerminusDB, to build and maintain knowledge store and to connect it to Wikidata. We will further enhance our research data repository (called depositar, website: https://data.depositar.io/about ) with Wikidata, and vice versa. 2. We will work on the community, technology, collaboration, and research aspects of research data management. We will help develop a community of practice for research data management in Taiwan. A research data repository we have developed (called depositar, website: https://data.depositar.io/about ) can function as a starting place where the communities practice research data management. The expected outcome of such a effort includes: cultivating research data management talents, participating in international collaborative projects for open data software systems, elevating the scale and capacity of the research data management community in Taiwan, and participating in and contributing to the global research community. Please refer to the following for more information: 1. Improving data discovery through Wikidata - WikidataCon 2019 https://commons.wikimedia.org/wiki/File:Improving_data_discovery_through_Wikidata_-_WikidataCon_2019.pdf 2. Open Repositories for Scholarly Communication and Participatory Research https://m.odw.tw/u/trc/m/rda-p18-panel/ |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/~trc/ 實驗室網址(Research Information) : http://data.depositar.io/about/ http://rdm.depositar.io/ Email : trc@iis.sinica.edu.tw |
| 葉彌妍 Mi-Yen Yeh | 深度學習與巨量資料探勘技術於相關人工智慧應用 Deep Learning and Big data mining for AI applications | 本實驗室將會探討各式深度學習模型，特別是Graph-based DNN models, 以及巨量探勘技術如何發展人工智慧應用，例如知識圖譜應用、推薦系統應用、運動科技應用等。 We study various deep machine learning models, especially the graph-based DNN models, as well as Big Data Mining techniques to develop various AI applications such as Knowledge graph applications and recommender systems. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/miyen/ 實驗室網址(Research Information) : http://www.iis.sinica.edu.tw/pages/miyen/ https:// Email : miyen@iis.sinica.edu.tw |
| 周彤 Tung Chou | 後量子密碼系統設計、實作與攻擊 Designing, implementing, and attacking post-quantum cryptosystems | 由於量子電腦的發展將危脅到現有的公鑰密碼系統，後量子密碼逐漸受到越來越多重視。在本次實習中，同學們將學習到後量子密碼系統的實作、設計與攻擊等議題。建議有基礎程式設計(C, Python)能力或有代數相關知識(群環體論)的同學參加。 Post-quantum cryptography is considered important as large-scale quantum computers are expected to break all the contemporary public-key cryptosystems. In this internship, participants are going to learn about how post-quantum cryptosystems can be designed implemented, and attacked. Recommended for those who has basic ability in programming (C, Python) or knowledge in abstract algebra (groups, rings, fields). |
PI個人首頁(PI's Information) : https://tungchou.github.io/ 實驗室網址(Research Information) : https://tungchou.github.io/ https:// Email : blueprint@citi.sinica.edu.tw |
| 王新民 Hsin-Min Wang | 語音辨識、語者暨語言辨識、語音合成與轉換、語者分離與分段、語音機器翻譯與問答系統 speech recognition, speaker/language recognition, speech synthesis and conversion, speaker separation and segmentation, speech translation, spoken question answering system | 研究主題中文介紹：語音處理是有高度發展前（錢）景的領域。在目前 AI 領域中，相對於影像、電腦視覺、自然語言處理，可說是一片藍海。我們致力於契合台灣語境（國語、臺語、客語、英語）的語音研究，在學術上既能與國際最高殿堂接軌，在系統上也不失本土化的應用意涵。 1）在語音辨識方面，我們的辨識器要聽得懂年輕人的國語、老人家的臺語，對於喜歡繞英文的人來說，也要難不倒它；加上語者過濾的技術，更要能專一辨識出特定語者的聲音，達到客製化、不受環境干擾的效果。 2）在語音合成方面，我們的合成系統不僅要會說國語、臺語，也要能夠使用語音轉換的技術，客製出使用者指定的人聲，不僅有娛樂性，更是語音保存、有聲書製作不可或缺的技術。 3）目前的「純」機器翻譯及「純」文字問答並不稀奇，真正能便利地應用於智慧家庭、音箱的使用環境還得藉助語音。因此，不僅我們的翻譯系統要能處理國人所需的國臺、國英互譯，我們的問答系統也要達到快速聆聽、高效反應的效果。 註：我們與高明達老師在臺語語音處理有密切合作。對臺語語音處理研究工作有高度熱忱及興趣的同學亦可申請加入高老師實驗室，若獲錄取將與我們一起合作。 Speech processing is a highly promising (lucrative) field. In the current AI field, compared to imaging processing, computer vision, and natural language processing, speech processing can be said to be a blue ocean. We are committed to speech processing research that fits the Taiwanese context (Mandarin, Taiwanese, Hakka, and English), and can be academically connected with the highest international halls, and the system does not lose the meaning of localized applications. 1) In terms of speech recognition, our recognizer must be able to recognize Mandarin speech of young people and Taiwanese speech of the elderly. For those who like to mix English in their speech, it should not be troublesome. Coupled with the technology of adaptation, it is necessary to be able to identify the voice of a specific speaker specifically to achieve the effect of customization and robustness to environmental interference. 2) In terms of speech synthesis, our synthesis system must not only speak Mandarin and Taiwanese, but also be able to use voice conversion technology to customize the user-specified human voice, which is not only entertaining, but also an indispensable technology for spoken language preservation and audiobook production. 3) The current text machine translation and text Q&A are not uncommon. It is really convenient to use in smart homes and speaker environments with the help of voice. Therefore, not only our machine translation system must be able to handle the translation between Mandarin and Taiwanese and between Mandarin and English that local people need, but also our Q&A system must achieve the effect of fast listening and efficient response. Note: We have close cooperation with Dr. Ming-Tat Ko in Taiwanese speech processing. Students with high enthusiasm and interest in Taiwanese speech processing research can also apply to join Dr. Gao's laboratory. Those who are admitted will work with us. |
PI個人首頁(PI's Information) : http://www.iis.sinica.edu.tw/pages/whm/ 實驗室網址(Research Information) : http://slam.iis.sinica.edu.tw/ https:// Email : whm@iis.sinica.edu.tw |
| 馬偉雲 Wei-Yun Ma | 廣告文案或新聞報導的自動生成技術 Natural Language Generation for Advertisement and News Report | 實習生主要會利用深度學習(Deep Learning)從事以下研究： 1. 廣告文案或新聞的自動生成 當輸入是一款手機的規格表，系統能自動生成出一篇具有說服力的廣告文案。或是當輸入是一場NBA的比賽數據表，系統能自動生成出一篇緊張刺激的播報新聞。我們希望透過深度學習當中的增強式學習(Reinforcement Learning)以及語言模型，打造一個這樣的文字生成系統，能夠一方面忠於輸入的表格內容，另一方面能發揮創造力，寫出多變化又文情並茂的文章。 2. 不限主題的閒聊機器人 即所謂 Chitchat Chatbot，也就是沒有特定目的的聊天. 目前這類型的bot大多數的作法是利用深度學習當中的seq-to-seq model來建構，但是，這樣的作法通常無法產生有意義或是較為深入的回應，多數會流於插瞌打渾或是賣萌。其中的關鍵，在於bot缺少了對於聊天主題相關的基本常識，就像是user要跟bot討論劉德華，bot應該對劉德華的各個fact(身份，作品...)有足夠認識，回的response才會豐富有意義，不然巧婦難為無米之炊，沒有知識就不容易產生有意義的回應，我們希望將grounded knowledge以及更豐富的語義訊息encode在model之中。我們透過深度學習當中的增強式學習(Reinforcement Learning)，已經訓練出一個不限主題的LINE閒聊機器人-詞庫小妍(LINE官方帳號：@359mcmgs)。 3. 自動知識學習系統\ 我們知道新的知識會夜以繼日的不斷產生，一個具有AI能力的系統最重要的功能之一就是能夠從大量的資料當中，分析資料，加以理解，組織成結構化知識。我們實驗室過去已經開發了人類的知識網(E-HowNet)，打下堅實基礎，此專案的目標是進一步加以擴張，利用深度學習技術將關鍵的關係三元組合從閱讀的文章中自動抽取出來，如 (”哈登” ，MemberOf，”火箭隊”) 或是 (“麥特載蒙”，PlayerOf，”心靈捕手”)等等。 4. 事實推論或事件預測系統 對於一個新事物，人們往往會根據基本常識、已知的事實、經驗的法則等等進行新事物的推測，包含事實或是事件的推論，例如以下的事實推論：已知A說中文，A又是B的哥哥，那麼很高的機率B也會說中文。又例如以下的事件推論：“買麵包”後會有很高的機率會在近期“吃麵包”。在一個龐大的文本或是複雜的知識圖譜當中，推論的關係往往數量龐大，有時甚至複雜到超越人力所能規範與理解，我們希望藉由深度學習技術能自動化的在文本或是知識圖譜當中進行新事物的推測。 As the Chinese version. |
PI個人首頁(PI's Information) : https://homepage.iis.sinica.edu.tw/pages/ma/index_zh.html 實驗室網址(Research Information) : https://ckip.iis.sinica.edu.tw/ https:// Email : ma@iis.sinica.edu.tw |