人工智能與機(jī)器學(xué)習(xí)綜合研究【大學(xué)組】

計(jì)算機(jī)科學(xué)/數(shù)據(jù)科學(xué)/人工智能/深度學(xué)習(xí)/機(jī)器學(xué)習(xí)

項(xiàng)目背景

機(jī)器學(xué)習(xí)(ML)和人工智能(AI)兩大主流科技將成為我們社會(huì)未來(lái)所依賴的基礎(chǔ)設(shè)施。然而，隨著這些領(lǐng)域的巨大快速增長(zhǎng)，越來(lái)越多的程序員與科學(xué)家似乎依賴于ML和AI的“黑盒”使用方法。也就是在不了解理論性質(zhì)的情況下,直接應(yīng)用計(jì)算技術(shù)和對(duì)應(yīng)框架。這一令人擔(dān)憂的現(xiàn)象對(duì)人工智能領(lǐng)域的發(fā)展毫無(wú)益處，本項(xiàng)目旨在填補(bǔ)這一空白，圍繞著機(jī)器學(xué)習(xí)算法的理論基礎(chǔ)展開(kāi)并佐以代碼實(shí)現(xiàn)，讓學(xué)生在實(shí)踐的同時(shí)也能深度理解其背后的理論支持。

項(xiàng)目介紹

本項(xiàng)目將帶領(lǐng)學(xué)生詳細(xì)了解機(jī)器學(xué)習(xí)的主要方法和當(dāng)前的研究方向，涵蓋機(jī)器學(xué)習(xí)中的不同算法的分析與對(duì)比。項(xiàng)目在討論至今仍有效的如決策樹(shù)的經(jīng)典算法外，還將討論以深度學(xué)習(xí)為例的改變了機(jī)器學(xué)習(xí)領(lǐng)域的新技術(shù)。學(xué)生將在導(dǎo)師的指導(dǎo)下挑選研究問(wèn)題，在項(xiàng)目結(jié)束時(shí)完成項(xiàng)目報(bào)告，進(jìn)行成果展示。

This course is designed to give students a detailed overview of the key approaches and current research directions in machine learning. Using a fast-paced, but still detailed set of lectures, the course will cover the different challenges in machine learning, different algorithms to address these challenges, and a comparative analysis of different algorithms to allow one to pick the right algorithm for a given machine learning task. The discussion will include both “classic” algorithms that are still effective today (e.g., decision trees) as well as new techniques that have transformed the area of machine learning (e.g., deep learning). The lectures will be accompanied with a set of group-based project assignments that will provide students hands-on experience with different machine learning algorithms to solve important tasks. The project assignments will be replaced by reading or research assignments for students who do not have programming background. Finally, students will prepare a research report and presentation to be shared with the class.

個(gè)性化研究課題參考 Suggested Research Fields

欺騙性、重復(fù)性的廣告檢測(cè)算法研究 Research on Deceptive and Duplicate Advertisement Detection Algorithms

針對(duì)用戶搜索記錄的酒店推薦算法 Recommendation System for Hotel Reservations Based on the User’s search History

根據(jù)網(wǎng)約車(chē)當(dāng)前運(yùn)行軌跡，預(yù)測(cè)本次行程時(shí)間的算法開(kāi)發(fā) Predict the total travel time of taxi trips based on their initial partial trajectories

預(yù)測(cè)土壤的物理化學(xué)成分 Predict physical and chemical properties of soil using spectral measurements

適合人群

大學(xué)生

計(jì)算機(jī)科學(xué)、數(shù)據(jù)科學(xué)、人工智能、機(jī)器學(xué)習(xí)專(zhuān)業(yè)或?qū)σ陨蠈?zhuān)業(yè)感興趣的學(xué)生； 學(xué)生需要具備線性代數(shù)及概率論與數(shù)理統(tǒng)計(jì)基礎(chǔ)，至少會(huì)使用一門(mén)編程語(yǔ)言并修讀過(guò)算法與數(shù)據(jù)結(jié)構(gòu)，有機(jī)器學(xué)習(xí)項(xiàng)目開(kāi)發(fā)經(jīng)驗(yàn)的申請(qǐng)者優(yōu)先

導(dǎo)師介紹

伊利諾伊大學(xué)香檳分校終身正教授

Dr. Rakesh is a Professor in the Electrical and Computer Engineering Department at the University of Illinois at Urbana Champaign with research and teaching interests in computer engineering, including recent interest in hardware for machine learning. His research has been recognized through one ISCA Influential Paper Award, one 10 Year Retrospective Most Influential Paper (MIP) Award (ASPDAC), several best paper awards and best paper award nominations (IEEE MICRO Top Picks, ASPLOS, HPCA, CASES, SELSE, IEEE CAL), ARO Young Investigator Award, and UCSD CSE Best Dissertation Award. His teaching and advising have been recognized through Stanley H Pierce Faculty Award and Ronald W Pratt Faculty Outstanding Teaching Award. He previously served as a Co-Founder and Chief Architect at Hyperion Core, Inc, a microprocessor chip startup. Rakesh has a BS from IIT Kharagpur and a PhD from University of California at San Diego.

導(dǎo)師現(xiàn)任伊利諾伊大學(xué)厄巴納香檳分校電子和計(jì)算機(jī)工程系的終身正教授，導(dǎo)師的研究興趣在于計(jì)算機(jī)工程與用于機(jī)器學(xué)習(xí)的計(jì)算機(jī)硬件。導(dǎo)師的研究成果曾獲ISCA最具影響力論文獎(jiǎng)、十年最具影響力論文獎(jiǎng)、及其他最佳論文獎(jiǎng)與提名(IEEE MICRO Top Picks、ASPLOS、HPCA、CASES、SELSE、IEEE CAL)。他的教學(xué)成果受廣泛認(rèn)可，并獲得了多份杰出教學(xué)獎(jiǎng)。除學(xué)術(shù)研究外，導(dǎo)師行業(yè)經(jīng)驗(yàn)也極為豐富，曾擔(dān)任Hyperion Core, Inc(微處理器芯片初創(chuàng)公司)的聯(lián)合創(chuàng)始人和首席架構(gòu)師。

任職學(xué)校

伊利諾伊大學(xué)香檳分校（UIUC）創(chuàng)建于1867年，是一所著名公立研究型大學(xué)，美國(guó)大學(xué)協(xié)會(huì)（AAU）成員，被譽(yù)為“公立常春藤”。學(xué)校的工程類(lèi)專(zhuān)業(yè)更是在世界大學(xué)中享有盛譽(yù)。2020年U.S.News世界大學(xué)排名顯示，UIUC計(jì)算機(jī)類(lèi)項(xiàng)目全部位列全美Top 10，計(jì)算機(jī)系統(tǒng)位列美國(guó)Top3。

項(xiàng)目大綱

機(jī)器學(xué)習(xí)分類(lèi)算法

Classification techniques – decision trees, k-NNs, perceptrons, and na?ve bayes

機(jī)器學(xué)習(xí)擬合算法

Regression – Linear and Logistic Regression, Clustering – k-means and EM clustering

神經(jīng)網(wǎng)絡(luò)與深度學(xué)習(xí)

Neural Networks and deep learning

強(qiáng)化學(xué)習(xí)與聯(lián)邦學(xué)習(xí)

Reinforcement learning, Ensemble methods

項(xiàng)目回顧與成果展示

Program review and presentation

論文輔導(dǎo)

Project deliverable tutoring

時(shí)間安排與收獲

7周在線小組科研學(xué)習(xí)+5周論文輔導(dǎo)學(xué)習(xí) 共125課時(shí)

學(xué)術(shù)報(bào)告

優(yōu)秀學(xué)員獲主導(dǎo)師Reference Letter

EI/CPCI/Scopus/ProQuest/Crossref/EBSCO或同等級(jí)別索引國(guó)際會(huì)議全文投遞與發(fā)表（可用于申請(qǐng)）

結(jié)業(yè)證書(shū)

成績(jī)單