Outline Advanced AI

Overview advanced ai l.jpg
1 / 17
1179 days ago, 478 views
PowerPoint PPT Presentation
General Thoughts and Teaching Philosophy I . Center of the course is giving more indepth information in the ranges specified on the past slight and to figure out how to peruse, abridge, show, and assess exploratory papers.Interactive dialog of papers and research topicsNo duping! No duping! No cheating!Teaching Philosophy: You will need to confront some feedback; else, you won't lea

Presentation Transcript

Slide 1

Outline Advanced AI History of AI thr. 1995 AI and the Web Traditional Clustering Graph Mining Spatial and Spatio-Temporal Data Mining Reinforcement Learning and Learning to Lean Shape-based Image Retrieval Hyla-Tree Frog

Slide 2

General Thoughts and Teaching Philosophy I Focus of the course is giving more indepth information in the regions said on the past slight and to figure out how to peruse, outline, exhibit, and assess logical papers. Intelligent talk of papers and research subjects No swindling! No tricking! No duping! Showing Philosophy: You should confront some feedback; else, you won't learn anything. Learning without presenting yourself to blunders is outlandish!! In the event that you don't realize what you foul up, it is difficult to make strides! Regardless of in the event that you like it are not, you should ramble in this course. Amid the course you will make a considerable measure of casual, unstructured introductions. Learn You accomplish Something Feedback

Slide 3

General Thoughts and Teaching Philosophy II No tasks and just 2 tests first week of March and 3 rd week of April. Learning by doing!! I am mindful that a large portion of you are not very experienced in these matters; thusly, my desires are at first very low. Not perusing papers that will be talked about on a specific day is not worthy! One goal if this course is to depict what other do in your own words – thusly, no replicating from any sources (web or schoolmates). On the off chance that you confront specific or bizarre issues when taking this course: converse with me amid my available time or send me an email. Amid the course you will likewise make 2 more formal introductions.

Slide 4

General Thoughts and Teaching Philosophy III You will likewise get some presentation concerning composing abstracts, outlines, presentations, white paper, and conclusions. Figuring out how to compose included to recognize what individuals expect concerning what you compose and how what you compose will be judged . In this course, we will attempt it a few showing techniques, some of which will be amended or even relinquished as the course advances. Is a by item you will ideally show signs of improvement comprehension on the most proficient method to direct a logical venture and on the best way to outline and present its outcomes. Incidentally, Dr. Eick will exhibit 10-40 minute addresses that give foundation learning concerning the papers that will be talked about next.

Slide 5

Topics Covered History of AI thr. 1995 – the 1994 Turing Award Lectures AI and the Web – how to survey the pertinence of site pages? Conventional Clustering Shape-based Image Retrieval Spatial and Spatio-Temporal Data Mining Reinforcement Learning and Learning to Lean Graph Mining Papers assignment to zones: I: 2; II: 2-4; III:3-4; IV:1; V:3-5 VI:2-3; VII:0-1.

Slide 6

Paper Reading List COSC 7363 AI through 1995: Edward Feigenbaum's and Raj Reddys 1994 Turing Award Lectures in CACM, May 1996, pages 97-112 . Likely, just Feigenbaums address "How the "What" turns into the 'How'" will be talked about. Page & Brin (Google Publication), The PageRank Citation Ranking: Bringing Order to the Web, 1998; walkthrough paper; maybe this is a superior wellspring of the work: Page & Brin (Google Publication), The Anatomy of a Large-Scale Hypertextual Web Search Engine ,1999. CV of Sergey Brin & potcast of 2005 Interview with Sergey Brin ( http://www.itconversations.com/appears/detail795.html ) Hanghang Tong, Christos Faloutsos, and Jia-Yu Pan , Fast Random Walk with Restart and Its Applications , Proc. ICDM Conference, Hong Kong, China, Dec. 2006; best research paper grant; 2 understudy administered walkthrough Langville & Meyer, Deeper Inside Page Rank, Internet Mathematics, Vol. 1, No. 3, 335-380, 2004; potential understudy introduction paper Original DBSCAN Paper; 2-understudy administered walkthough Rousseaux Original Silhouette Paper; walkthrough; figure out how to compose a presentation and a dynamic Likely, Original EM Paper; McLachlan, G. furthermore, Peel, D. (2000). Limited Mixture Models. J. Wiley, New York. groups of 2 read the paper, figure out how to compose a conclusion Clustering with Bregman Divergences by A. Banerjee, S. Merugu, I. S. Dhillon, and J. Ghosh , in Journal of Machine Learning Research , vol. 6, pages 1705-1749, October 2005

Slide 7

Paper Reading List COSC 7363 Cyrus Shahabi, Maytham Safar, A test investigation of option shape-based picture recovery procedures , Multimedia Tools and Applications , Springer Netherlands, November 2006 . S. Shekhar, P. Zhang, Y. Huang, R. Vatsavai, Trend s in Spat ial Data Mining, book part in Data Mining: Next Generation Challenges and Future Directions , H. Kargupta, A. Joshi, K. Sivakumar, and Y. Yesha(eds.), AAAI/MIT Press, 2003 . Some Spatial Statistics Paper Co-area Mining with Rare Spatial Features by Yan Huang, Jian Pei, and Hui Xiong distributed in Journal of GeoInformatica, vol. 10, issue 3 , 2006 . Mirco Nanni, Dino Pedreschi. Time-centered thickness based grouping of directions of moving articles . i n Journal of Intelligent Information Systems (JIIS), 27(3):267-289 , Special Issue on "Mining Spatio-Temporal Data", 2006. H. Cao, N. Mamoulis, and D. W. Cheung,  " Discovery of Periodic Patterns in Spatiotemporal Sequences ," IEEE Transactions on Knowledge and Data Engineering ( TKDE ) , to show up. Fortification Learning: A Survey by Kaelbling, L. P. furthermore, Littman, M. L. in JAIR, 1996 ( http://www.cs.cmu.edu/afs/cs/extend/jair/bar/volume4/kaelbling96a-html/rl-survey.html ) Very Likely Something from the Thrun book " Learning to Learn ", 1998. In the event that enough time left: Yan, X. also, Han, J. 2002. gSpan: Graph-Based Substructure Pattern Mining , i n Proceedings of the 2002 IEEE International Conference on Data Mining (ICDM '02) Washington, DC.

Slide 8

Teaching Plan Next 5 Weeks Janurary 16: Course Overview (see likewise http://www2.cs.uh.edu/~ceick/7363/7363.html ) January 18: (Read) Feigenbaum Paper  How to peruse a paper January 23: Prepare Slow PaperWalkThrough Page/Brin Paper January 25: Mostly Lecture January 30: Prepare Slow PaperWalk Through ICDM06 Best Research Paper Award Feb. 1: TBDL - Brin/Page Background?!? Feb. 6: Likely Reddy Turing Award Paper Feb. 8: Student Guided Discussion DBSCAN paper Feb. 13: Student Guided Discussion Silhouette Paper Feb. 15: Student Presentation Langville/Mayer Paper (discretionary) Feb. 20: On Writing Abstracts, Introductions, Conclusions Feb. 22: Bregman Divergences Paper

Slide 9

Course Activities A considerable measure of casual introductions and exchanges 1 formal introduction about a paper canvassed in the course Writing abstracts, presentations, conclusions and paper audits - learning by doing 2 Quizes that make inquiries about papers we have perused Discussions Learning how to peruse, outline, present, and survey papers. Foundation learning 'on the most proficient method to play out an examination extend' and 'on the most proficient method to be effective in your exploration/vocation'. Talking about numerous other, engaging things -, for example, the 'Mammoth Squid', life of Sergey Brin - a large portion of which are still identified with one of the above exercises.

Slide 10

Forms of Covering Papers in the Course Papers will be talked about, displayed in a wide range of structures in this course: Slow Walk Through (I just arrangement to have 3 of those!!) Guided Walk Through Other Walk Throughs ( I didn't consider yet!) By noting a given arrangement of quesitons. Just Discussion 1-Page (5-page) Summary of a Paper Professional Powerpoint Presentation Profession Paper Review (  April 2007) …

Slide 11

Slow Walk Throughs Used for the initial two " AI and the web " papers Paper will be talked about passage by section Very moderate!! In this way, there will be just 2-4 of those… Course members are in charge of segments of the paper. Duties include: Lead examination Present short synopses for exhausting segments to accelerate things Ask inquiries concerning things they don't comprehend Prepare survey inquiries for alternate understudies that will be talked about either promptly or after a deferral. Everyone ought to peruse the paper deliberately including the segments you are not in charge of. It may be a smart thought to make brief synopses for the read segments and to catch, what you don't comprehend, in type of inquiries. On the off chance that you wrapped up the paper attempt to concoct your own assessment of the paper We won't just talk about the substance of the paper, additionally address the question "why a writer composes a paper especially" and "how the introduction of the examined paper could be enhanced/made all the more persuading. Furthermore, issues on the best way to compose a paper will be examined amid moderate walk throughs - these matters are Dr. Eick's duty..

Slide 12

Sergey Brin CV (Сергей Михайлович Брин) (see additionally http://en.wikipedia.org/wiki/Sergey_Brin ) Sergey Brin is prime supporter and President, Technology at Google. Initially a local of Moscow, he got a four year certification in scientific studies degree with distinction in arithmetic and software engineering from the University of Maryland at College Park. He is presently on leave from the Ph.D. program in software engineering at Stanford University, where he got his graduate degree. Sergey is a beneficiary of a National Science Foundation Graduate Fellowship and additionally a privileged MBA from Instituto de Empresa . It was at Stanford that he met Larry Page and chipped away at the venture that got to be Google. Together they established Google Inc. in 1998, and Sergey keeps on sharing duty regarding everyday operations with Larry Page and Eric Schmidt. Sergey's exploration advantages incorporate web crawlers, data extraction from unstructured sources, and information mining of substantial content accumulations and logical information. He has distributed more than twelve scholarly papers and has been a highlighted