【演講】2019/11/19 (二) @工四816 (智易空間),邀請到Prof. Geoffrey Li(Georgia Tech, USA)與Prof. Li-Chun Wang(NCTU, Taiwan) 演講「Deep Learning based Wireless Resource Allocation/Deep Learning in Physical Layer Communications/Machine Learning Interference Management」
IBM中心特別邀請到Prof. Geoffrey Li(Georgia Tech, USA)與Prof. Li-Chun Wang(NCTU, Taiwan)前來為我們演講,歡迎有興趣的老師與同學報名參加!
演講標題:Deep Learning based Wireless Resource Allocation/Deep Learning in Physical Layer Communications/Machine Learning Interference Management
演 講 者:Prof. Geoffrey Li與Prof. Li-Chun Wang
時 間:2019/11/19(二) 9:00 ~ 12:00
地 點:交大工程四館816 (智易空間)
活動報名網址:https://forms.gle/vUr3kYBDB2vvKtca6
報名方式:
費用:(費用含講義、午餐及茶水)
1.費用:(1) 校內學生免費,校外學生300元/人 (2) 業界人士與老師1500/人
2.人數:60人,依完成報名順序錄取(完成繳費者始完成報名程序)
※報名及繳費方式:
1.報名:請至報名網址填寫資料
2.繳費:
(1)親至交大工程四館813室完成繳費(前來繳費者請先致電)
(2)匯款資訊如下:
戶名: 曾紫玲(國泰世華銀行 竹科分行013)
帳號: 075506235774 (國泰世華銀行 竹科分行013)
匯款後請提供姓名、匯款時間以及匯款帳號後五碼以便對帳
※將於上課日發放課程繳費領據
聯絡方式:曾紫玲 Tel:03-5712121分機54599 Email:tzuling@nctu.edu.tw
Abstract:
1.Deep Learning based Wireless Resource Allocation
【Abstract】
Judicious resource allocation is critical to mitigating interference, improving network efficiency, and ultimately optimizing wireless network performance. The traditional wisdom is to explicitly formulate resource allocation as an optimization problem and then exploit mathematical programming to solve it to a certain level of optimality. However, as wireless networks become increasingly diverse and complex, such as high-mobility vehicular networks, the current design methodologies face significant challenges and thus call for rethinking of the traditional design philosophy. Meanwhile, deep learning represents a promising alternative due to its remarkable power to leverage data for problem solving. In this talk, I will present our research progress in deep learning based wireless resource allocation. Deep learning can help solve optimization problems for resource allocation or can be directly used for resource allocation. We will first present our research results in using deep learning to solve linear sum assignment problems (LSAP) and reduce the complexity of mixed integer non-linear programming (MINLP), and introduce graph embedding for wireless link scheduling. We will then discuss how to use deep reinforcement learning directly for wireless resource allocation with application in vehicular networks.
2.Deep Learning in Physical Layer Communications
【Abstract】
It has been demonstrated recently that deep learning (DL) has great potentials to break the bottleneck of the conventional communication systems. In this talk, we present our recent work in DL in physical layer communications. DL can improve the performance of each individual (traditional) block in the conventional communication systems or jointly optimize the whole transmitter or receiver. Therefore, we can categorize the applications of DL in physical layer communications into with and without block processing structures. For DL based communication systems with block structures, we present joint channel estimation and signal detection based on a fully connected deep neural network, model-drive DL for signal detection, and some experimental results. For those without block structures, we provide our recent endeavors in developing end-to-end learning communication systems with the help of deep reinforcement learning (DRL) and generative adversarial net (GAN). At the end of the talk, we provide some potential research topics in the area.
3.Machine Learning Interference Management
【Abstract】
In this talk, we discuss how machine learning algorithms can address the performance issues of high-capacity ultra-dense small cells in an environment with dynamical traffic patterns and time-varying channel conditions. We introduce a bi adaptive self-organizing network (Bi-SON) to exploit the power of data-driven resource management in ultra-dense small cells (UDSC). On top of the Bi-SON framework, we further develop an affinity propagation unsupervised learning algorithm to improve energy efficiency and reduce interference of the operator deployed and the plug-and-play small cells, respectively. Finally, we discuss the opportunities and challenges of reinforcement learning and deep reinforcement learning (DRL) in more decentralized, ad-hoc, and autonomous modern networks, such as Internet of things (IoT), vehicle -to-vehicle networks, and unmanned aerial vehicle (UAV) networks.
Bio:
Dr. Geoffrey Li is a Professor with the School of Electrical and Computer Engineering at Georgia Institute of Technology. He was with AT&T Labs – Research for five years before joining Georgia Tech in 2000. His general research interests include statistical signal processing and machine learning for wireless communications. In these areas, he has published around 500 referred journal and conference papers in addition to over 40 granted patents. His publications have cited by 37,000 times and he has been listed as the World’s Most Influential Scientific Mind, also known as a Highly-Cited Researcher, by Thomson Reuters almost every year since 2001. He has been an IEEE Fellow since 2006. He received 2010 IEEE ComSoc Stephen O. Rice Prize Paper Award, 2013 IEEE VTS James Evans Avant Garde Award, 2014 IEEE VTS Jack Neubauer Memorial Award, 2017 IEEE ComSoc Award for Advances in Communication, and 2017 IEEE SPS Donald G. Fink Overview Paper Award. He also won the 2015 Distinguished Faculty Achievement Award from the School of Electrical and Computer Engineering, Georgia Tech.
Li-Chun Wang (M'96 -- SM'06 -- F'11) received Ph. D. degree from the Georgia Institute of Technology, Atlanta, in 1996. From 1996 to 2000, he was with AT&T Laboratories, where he was a Senior Technical Staff Member in the Wireless Communications Research Department. Currently, he is the Chair Professor of the Department of Electrical and Computer Engineering and the Director of Big Data Research Center of of National Chiao Tung University in Taiwan. Dr. Wang was elected to the IEEE Fellow in 2011 for his contributions to cellular architectures and radio resource management in wireless networks. He was the co-recipients of IEEE Communications Society Asia-Pacific Board Best Award (2015), Y. Z. Hsu Scientific Paper Award (2013), and IEEE Jack Neubauer Best Paper Award (1997). He won the Distinguished Research Award of Ministry of Science and Technology in Taiwan twice (2012 and 2016). He is currently the associate editor of IEEE Transaction on Cognitive Communications and Networks. His current research interests are in the areas of software-defined mobile networks, heterogeneous networks, and data-driven intelligent wireless communications. He holds 23 US patents, and have published over 300 journal and conference papers, and co-edited a book, “Key Technologies for 5G Wireless Systems,” (Cambridge University Press 2017).
同時也有1部Youtube影片,追蹤數超過80萬的網紅Science Experiments with Physics Engine,也在其Youtube影片中提到,強化学習で人に二足歩行を覚えさせました。「proximal policy optimization (PPO)」というアルゴリズムを使っています。 Proximal Policy Optimization Algorithms https://arxiv.org/abs/1707.06347 T...
algorithms optimization 在 AOPEN Taiwan Facebook 八卦
人工智慧=未來趨勢?! AOPEN在浪潮跟你一起IoT🤣🤣
#李開復先生怎麼說
#各位看官意下如何
#人工智慧這樣行👊👊
#偵測到你無所遁形
#人潮就是錢潮🤑🤑🤑🤑🤑
#AOPEN #建碁
【李開復Quartz專文:「人形機器人」將進入千家萬戶是無稽】人工智能時代的到來已經指日可待了。但是目前情感機器人、家庭機器人都離我們很遠。下面是外媒Quartz跟我的約稿:
▲ 機器人能幫著賺錢、省錢、提高生產力,也助人們回歸人性
人工智能時代的到來已經指日可待了。但是目前情感機器人的發展方向有點南轅北轍。
首先,讓我們明確一下人工智能的幾個要點:
人工智能擅長對目標明確的工作進行優化(但是不能創造,沒有感情)。
機械控制的發展速度較人工智能軟件的發展要緩慢得多。
傳感器雖然得到迅猛發展,但價格昂貴、體積偏大且太耗電。
鑒於以上原因,人形機器人將馬上進入千家萬戶的說法,簡直是無稽之談。當機器人在言談舉止各方面都與人類極其相似時,普通家庭用戶對機器人的「人類素質」的期望也會變得高不可攀。僅僅這種期望所帶來的失望就足以讓很多公司的「未來十年讓科幻小說成為現實」的展望受挫,更別提消費市場對價格的苛刻要求了。
機器人的開發要牢記實用性這一原則:機器人或能創造效益,或能節省成本,或能提高生產,或可以提供娛樂。依託現有技術製造的工業機器人將高效製造出其它機器人;商用機器人將會帶來更多經濟收益(例如替代保安、前台和司機等職位);家用機器人將能發揮家用電器和玩具的功能——它們簡單易用且不具備任何「人性素質」。
這樣的機器人未必具備人類外形。工業機器人就是在黑暗廠房(例如富士康最先進的廠房)或者配備了智能升降機倉庫里(例如我們投資的開源機器人Dorabot)從事勞務的機器;商用機器人的形式和用途就更多樣了:它們也許就是一排攝像頭(例如曠視科技的產品),或者是一家自動商店(例如F5未來商店)。自動駕駛車將有車的外形——除了那種低速貨運、功能固定的運輸工具,例如機場鋪設的自動車道,或者從停車場到商店、主題公園的運輸設備(例如UISEE馭勢科技);消費機器人也許會像一個揚聲器(例如亞馬遜的Echo)、一台電視機、一台吸塵器(例如Roomba)、一個教學玩具(例如奇幻工房的Dash Bot)或者一台用於家庭聯繫的平板電腦(例如小魚在家)。
人工智能也會與時俱進嗎?這一點毋庸置疑。聲音識別技術將更精准,電腦視覺技術也會提高,SLAM技術將讓機器人的動作更加流暢,機器人將會翻譯,還會針對限定領域進行對話。機器人也可能會瞭解我們的情緒並能模仿人類的情緒。這種情緒模仿將從搞笑的、娛樂性的發展為一定程度上能產生共鳴的模仿。誠然,這種模仿也都不是自發性的。在未來數十年,機器人還不能獨立進行常識性的推理、創造及規劃工作,它們也不會擁有自我意識、情感及人類的慾望。那種「全知全能人工智能」尚不存在,而且現在已知的開發技術也無法開發出此類機器人。這種技術在未來數十年都不會出現,也許永遠都不會出現。
人形機器人的研發對人工智能科學家充滿了誘惑力,而對人形機器人的預測也順理成章地激發著科幻小說家們的創作靈感。但是我們和人工智能有著本質區別:我們會創造,AI只會在創造的基礎上優化;我們多愁善感,AI冷酷無情;我們具備常識判斷能力,而AI只會從特定領域的大數據獲得信息。一言以蔽之,人類所長正是AI所短,而AI所長也是人類所短。
展望未來,人類最前沿的領域將是創造及社交領域。因此,我們應該推動機器人向它們所擅長的領域發展,例如進行重復性工作、優化工作或者創造財富的實用性工作。而我們人類也應該做一些我們擅長的工作:創新、創造、社交溝通或者娛樂。
我一直倡導要開發一些實用性機器人,鼓勵人們進入服務行業。但我不支持製造「類人」機器人。這種機器人開發難度大,而且永遠無法滿足人們的期望,因此,這種機器人的勝算不大。我分析的正確與否暫且不論,但是有一點我們需要有清晰的認識,那就是,未來十年,AI將大規模地取代那些依靠人力的、重復性的、分析性的崗位。因此,我們要肩負起創造更多社會服務性崗位的職責,而不是空想或謀劃一個充斥著「不適用於人類」職位的社會。
◀英文原文▶
Robots should make money, save money, increase productivity, or deliver entertainment—and let humans be human
Robots should make money, save money, increase productivity, or deliver entertainment—and let humans be human
The age of artificial intelligence (AI) and robotics is upon us, but the current fad of emotional humanoid robots is not headed in the right direction.
First, let’s understand what robotics are based on:
AI algorithms which are very good at optimization of explicitly defined goals (but cannot create, and have no feelings)
Mechanical control which advances much slower than AI software algorithms
Sensors which are rapidly improving but are often still too expensive, too large, or too power-hungry
Given the above, it is ludicrous to think that human-like robots will roam our homes any time soon. When a robot looks like a person, talks like a person, and has features like a person, home users will have unattainable human-capability expectations. The disappointment alone will doom any company hoping to bring science fiction to the living room in the next decade, not to mention the price-sensitivity for consumer markets.
Robotics must begin with utilitarianism in mind—robots should make money, save money, increase productivity, or deliver entertainment. There will be industrial robots that build other robots in high-volume, manufactured with today’s technologies. There will be commercial robots that deliver economic value (such as replacing security, receptionists, and drivers). There will be consumer robots that mimic today’s appliances and toys, requiring no consumer education, and causing no human-capability expectation.
These robots won't look like a person. The industrial robot is a giant factory run in the dark by machines (like at Foxconn’s most advanced factories), or a warehouse with smart forklifts (like our investment Dorabot). The commercial robot comes in various forms and applications. It might look like an array of cameras (like our investment Megvii) or an automated store (like our investment F5 Future Store). The autonomous vehicle will look like a car, except will be first deployed in low-speed, freight, or fixed-function transport—such as in airport autonomous car-only lanes, or in transport from parking garages to shopping malls/theme parks (like our investment UISee). And the consumer robot may look like a speaker (like the Amazon Echo), a TV, a vacuum cleaner (like Roomba), an educational toy (like our investment Wonder Workshop Dash Bot), or a pad-on-steroids for family communications (like our investment Ainemo).
Will AI capabilities increase over time? Of course. Speech recognition will get better, computer vision will improve, SLAM will be improved to help the robot move around fluidly, and the robot will be able to translate languages, or have a dialog within limited domains. The robot may be able to read some of our emotions, or mimic certain human emotions. But this mimicking will go from laughable and entertaining to occasionally acceptable—and generally not genuine. For decades to come, robots by themselves will be unable to learn common sense reasoning, creativity, or planning. They also won't possess the self-awareness, feelings, and desires that humans do. This type of “general AI” does not exists, and there are no known engineering algorithms for it. I don’t expect to see those algorithms for decades, if ever.
Trying to make robots human-like is a natural temptation for robotics and AI scientists, and predicting humanoid robots comes naturally to science fiction writers. But we humans simply think differently from AI. We create and AI optimizes. We love and AI is stoic. We have common sense and AI learns patterns from big data in a singular domain. Simply stated, we are good at what AI is not, and AI is good at what we are not.
In the future, the human edge will be in creativity and social interaction. Therefore, we need to focus robotics development toward what they’re good at: repetitive tasks, optimization, and utilitarian value creation. We should also let people do what they’re good at: innovation, creation, human-to-human interaction, and performing services.
I am an advocate of making utilitarian robots, and encouraging people to go into service jobs. I am not an advocate of making humanoid service robots—it is too hard today, and will not meet people’s expectations; therefore they will likely fail. Whether or not my analysis is correct, we need to be reminded that in the next decade AI will replace a massive number of manual-labor, repetitive, and analytical jobs. We have a human responsibility to help create societal service jobs—not dream or plan a society in which all jobs come with a sign “humans need not apply.”
algorithms optimization 在 國立陽明交通大學電子工程學系及電子研究所 Facebook 八卦
【深度學習硬體加速器設計】微學分課程來囉!
To teach students theory, algorithms, Python programming, ASIC optimization of contemporary neural networks in terms of performance, accuracy, model size, energy efficiency.
課程名稱:深度學習硬體加速器設計
課程資訊:http://ict.nctu.edu.tw/?p=2559
選課時間:109/8/11(二)~8/26(三)
請至NCTU-ICT課程選課系統報名──
http://140.113.159.211/ict/course/
【選課流程】
1、初次登入請務必完整填寫「基本資料」。
#基本資料有誤則選課無效
2、選課請點選「選課報名」,選課期間可自行取消報名。選課截止後如需退選,請於三日內回覆中選通知信辦理。
#無故未到課或未依規定退選將禁止選課兩個月
algorithms optimization 在 Science Experiments with Physics Engine Youtube 的評價
強化学習で人に二足歩行を覚えさせました。「proximal policy optimization (PPO)」というアルゴリズムを使っています。
Proximal Policy Optimization Algorithms
https://arxiv.org/abs/1707.06347
Twitter:https://twitter.com/physics_engine0
BGM:
「Trick or treat」written by GT-K
「Halloween Monsters」written by ISAo.
#物理エンジンくん