Search and Agent - PowerPoint PPT Presentation

albert
search and agent l.
Skip this Video
Loading SlideShow in 5 Seconds..
Search and Agent PowerPoint Presentation
Download Presentation
Search and Agent

play fullscreen
1 / 68
Download Presentation
Search and Agent
381 Views
Download Presentation

Search and Agent

- - - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript

  1. Search and Agent October 8, 2002 Jae Kyu Lee Graduate School of Management Korea Advanced Institute of Science and Technology

  2. Consumer Mercantile Activities Product/service search and discovery in the information space Comparison shopping and product selection based on various attributes Prepurchase interaction Nogotiation of terms, e.g., price, delivery times Placement of order Purchase consummation Authorization of payment Receipt of product Postpurchase interaction Customer service and support (if not satisfied in X days, return product)

  3. Fundamental Stages of Buying Process • Stages (Maes’ View, CACM 1999) 1. Need identification 2. Product brokering 3. Merchant brokering (price, warranty, availability, delivery time, and reputation) 4. Negotiation (fixed price introduced only 100 hundred years ago) 5. Purchase and delivery 6. Product service and evaluation • Buying Behavior studied by Nicosia, Howard-Sheth, Engel-Blackwell, Bettman, and Andreasen • Role of Software Agent[mediate consumer behaviors] Information filtering and retrieval, personalized evaluations, complex coordination, and time-based interactions.

  4. Search and Comparison Shopping • Key Word Search and Directory • Meta-Key-Word Search and Meta-Directory • B-Cart and Meta-Malls Architecture • Agents based Comparison

  5. Search Engine • Gathering: • Manually or automatically • Search Agent vs. Mobile Agent • Indexing • Storing • Retrieval

  6. Intelligent Search on Documents • Key Word Search: String matching • Advanced Search : Conditional, AND/OR relationships, and filtering • Demonstrate the advanced features in google.com • Proximity Search: Synonyms • Languages and Translations • Concept Based Search: • Concept Object Hierarchy • Directory and Linkages • Similarity Based Search: Case Based Retrieval • Case Based Reasoning; Associations • Image Search: Content-Based Image Retrieval • Image Processing vs. Text-based Index about images

  7. Concept Object Hierarchy Home Requirement Level of Detail Balancing Buyer :e-Procurement Directory of KAIST Supplier: MarketSite.net Home + Packing machi General Office Supplies + Procurement - + Labeling mach Computer Supplies Logic Expression + + Sorting machin Computer Accessories Generation of Directory + + Peripheral Data Storage Typing machin Consistency of Directory - + Printer Supplies Binding and La - Management of Rule Base + Toner Office machina - + Use for Hewlett-Packard Fusers and acc Extract + + LaserJet 4V Printer, fax an Q & A + Manual LaserJet 5 + Printer, fax Help - + LaserJet 6P Printer, fax E-Mail + • HP Model C3903A Image Cartri Toner + • Nu-Kote Model LT105R Image Transfer roll

  8. Popular Search Engines • Google. Com and Google.co.kr • Yahoo • Empas • Naver • Daum.net • Simmani • Hanmir • Dreamwiz • Lycos • Altavista

  9. Meta-Key-Word Search and Meta-Directory Relevant Pages Directoies KeyWord Inquiry Customer • How to Acquire Knowledge about Directory? • Manual Registration • Automatic Learning Meta Directory Merged Response

  10. Meta Search Engines: Experimental • All-in-One Search Page • http://www.albany.net/allinone • CUSI • http://web.nexor.co.uk/cusi/cusi.html • Fun City Web Search • http://www.funcity.com/search.html • HyperNews • http://union.ncsa.uiuc.edu/HyperNews/get/www/searching.html • Info Market • http://www.infomkt.ibm/com/ • Meta Crawler • http://www.cs.whashington.edu:8080/ • Savvy Search • http://www.cs.colostate.edu/~drelling/smartform.html • Searchers • http://gagms.www.com/~boba/search.html • W3 Search Engine • http://cuiwww.unige.cd/meta-index.html • Web-Search • http://www.bidderford.com:80/~soaring/

  11. Mobile Agents • Mobile Agents = Mobile Robots • Why Robot Exclusion Standard • 1993-1994: Too frequent visits of mobile agents • Disturbed the visiting servers • Robot Exclusion Standard was necessary • Format • “<field>: <optionalspace><value> <optionalspace> • Values in Field: User-agent, Disallow • Examples 1: User-agent: * Disallow: /cyberworld/map/ Disallow: /tmp/ • Any robots are not allowed to visit the URLs in “/cyberworld/map/”, “/tmp/”

  12. Examples: Robot Exclusion Standard • Examples 2: User-agent: * Disallow: /cyberworld/map/ User-agent: cybermapper Disallow: • All robots except the cybermapper are not allowed to visit “/cyberworld/map/” • Examples 3: User-agent: * Disallow: / • No robot is allowed to visit.

  13. Five Types of Comparison Support(TL2.11, p. 58-63) 1. Search on Hypertext files using Agents • Bargain Finder 2. Search in a Web-based Database: human and S/W agents sharing information • Mysimon, Junglee, Jango 3. Comparable Item Retrieval and Tabular Comparison • Compare.net (Side by Side comparison possible) • Meta-Malls Comparison Architecture 4. Comparison of Multiple Items from Multiple eMarketplaces - Personalized B-Cart 5. Comparison as a Multiple Criteria Decision Making • Personalogic

  14. Characteristics of Key Comparison Sites ILA* : Internet Learning Agent engine, 50% Performance Enhancement to BF VDB* : Virtual DB ILA*: Internet Learning Agent Engine (50% Performance Enhancement to Bargain Finder) VDB*: Virtual DB

  15. Compare Shopping Sites • Google.com: Search “Comparison Shopping” • Get List of Sites • PricingCentral.com • List of search engines for item categories • Try with Notebook: Review the site and user’s rating • Dealtime.com • Specificational conditions • Display stores with rating, product specificaitions, price and link • Compare.Net • Tabular Comparison • Dell.com • By Application Category, Customized Configuration and offer price • CNET.com • Post Popular: Comparative Review • Show by Models • Show Stores with rateing, price, shipping cost, stock status, take orders, related products, Price Drop Alert • MySimon • Specification based Search

  16. Korean Comparison Sites • Omi.co.kr • Bestbuyer.co.kr • Danawa.co.kr • Enuri.com • Shop.Lycos.co.kr • Price.naver.com • Clickprice.co.kr • Yavis.com • Compare.co.kr • Shopbinder.com • Mymargin.com • Okprice.com • Shoppal.net

  17. 분야별 비교구매 서비스

  18. 분야별 비교구매 서비스

  19. 분야별 비교구매 서비스

  20. 분야별 비교구매 서비스

  21. Current Status of Comparison Shopping • Retrieval of Standard Models • Graphic Displays • Preference Based Scoring • Tabular Comparison • Configuration Support • Intelligence: Rules, Constraints, Preference, Similarity  Multiple Criteria Decision Support

  22. Opportunities in Comparison Shopping Research Opportunities • Configuration with Options: - CBR: Find the Best Standard Model - CSP: Search toward the best options • Multi-vendor based Configuration and Order Processing • Pick and Delivery necessary • Buyer Agents • Buyer’s Bahavior?

  23. Architecture of B-Cart Buyer Sellers e - Procurement e - Catlaog MyB - Cart System Collect / Order MyS - Cart . . . . . . Update Intermediaries Buyer Visit / Order e - Catlaog Interface MyI - Cart

  24. Meta-Malls Architecture • Goal • Shopping Over Multiple Independent Cyber Shopping Malls • One Stop Shopping and Payment over Multiple Malls • Product Level Comparison Support As A Multiple Critera Decision Marking

  25. Characteristics of The Meta-Malls Architecture • Meta-Malls Coordinator Keeps: • Summary Products and URL • Relationship Indices between Products and Malls • Communicate with the Mall Operators • Mall Operator • Independent Entities • But has optional linkage with Meta-Mall Coordinator • Merchant: Compatibility with the Coordinator necessary

  26. Individual Buyer’s Cart: b-Cart • Tentative Picking and Decision Support • Allow Budget Consideration • Supporting Record Keeping Possible • Individual Buyer Assistant in Buyer’s Personal Computer • Coupling with the ERP can be supported • One-Stop Payment • IC-Card Based User Access for Certification and Electronic Wallet can be supported

  27. Illustrative Screen of B-Cart

  28. Agent Based Commerce • Agent • Software Agent • Intelligent Agent • Agent Communication Languages • Robot • Softbot

  29. Definitionof Agent Technologies (U.S. Lee, ICEC 98 Tutorial) Internal definition A software module which acts with its own knowledge, belief, interest and intention against informations given from environment External definition A functional unit which acts like a human being or transacts tasks on behalf of a human with human features visually and aurally - - A general term to indicate a functional module which solves a problem autonomously toward a specific goal by recognizing situation and cooperating other systems based on interactions with environment

  30. Features of Agent [Wooldrige and Jennings (1994)] Weak Notions - • Autonomy or semiautomatic • Social ability • Reactivity • Pro-activeness - - - Strong Notions - Mentalistic notions Knowledge Belief Intention Obligation Emotion - - - - -

  31. Features of Agent Other Attributes - • Continuous running: monitoring • Mobility : Mobile Agent • Veracity (성실성) • Benevolence (선행성, 순응성) • Rationality - - - -

  32. Core technologies for Agents in EC • Autonomous / Intelligence • Recognition of environment • Problem solving / Learning mechanism - - • Communications • Task / Result / Knowledge sharing • Agreement / Negotiation / Competition / Balancing - - Anthropomorphism

  33. A Prototype of Agent Development AGENT ENVIRONMENT Processing Modu. Problem Solver Cooperation Controller Behavior Controller Communication Modu. • Human Clients • Other Agents • / Systems Message/event handler Protocols for collaboration negotiation Knowledge Modu. Knowledges for problem solving control cooperation knowledge management See UNIK-AGENT

  34. Seller R1 Seller R4 Seller R2 Seller R3 Agent-based EC: Contract Type and Protocol Requirement of Human Customer Requirement of Human Customer Buyer C1 Buyer C0 Final Approver1 Final Approver2 Final Approver3 Final Approver4

  35. Directory Architecture of UNIK-AGENT AGENT Communication Controller Problem Solver Outgoing Msg Problem Solving Manager Incoming Msg Message Base Message Manager Directory Consulting Other Agents Solution Engines Individual Messages Message Queue Mgt. Data Base Knowledge Base Message Gate

  36. KQML • KQML • Performatives ask-if, evaluate, tell, etc. • Performative Parameters sender, receiver, reply-with, in-reply-to, content, ontology, and language. AGENT A (ask-if :language KIF :ontology electronic :reply-with q1 :content (> (size chip1) (size chip2))) AGENT B (tell :language KIF :ontology electronic :in-reply-to q1 :content true)

  37. KQML Message Examples evaluate : content <expression> : language <word> : ontology <word> : reply-with <expression> : sender <word> : receiver <word> reply : content <expression> : language <word> : ontology <word> : in-reply-to <expression> : forte <word> : sender <word> : receiver <word>

  38. KQML Examples ask-if : content <expression> : language <word> : ontology <word> : reply-with <expression> : sender <word> : receiver <word> tell : content <expression> : language <word> : ontology <word> : in-reply-to <expression> : force <word> : sender <word> : receiver <word>

  39. Three Layers of EC Messages • Agent Communication Language Layer • Domain independent communication language among agents (e.g. Knowledge Query and Manipulation Language: KQML) • Electronic Commerce Layer • Message types and items for agent based commerce • Product Specification Layer • Representation the specification of products

  40. Message Standard in Three Layers (evaluate :sender C1 :receiver R1 R2 R3 R4 R5 :reply-with msg_960924_1 :ontology Agent Based Commerce :language UNIK-OBJECT :content ((title RFP) (contract_ID contract_960924) (contract_type (number_of_bid_round 1) (competitor_price_referable not_referable) (announced_estimated_price_limit enforced) (number_of_proposal_for_each_bidder 1) (bid_price_change not_allowed) (bid_price_open_time at_predetermined_time) (buyer_bidder_prenomination prenominated_by_customer) (human_involvement enforced) (bidding_price_type total_amount) (buyer_rule_for_selection_of_successful_bidder min_price) (buyer_nego_between_price_&_spec not_allowed)) Agent Ccommunication Language Layer (bid_time Electronic Commerce Layer (end_time 96/10/01/11/00)) (requirement (payment_method credit_card) (delivery_method postal) (delivery_date Oct 14, 1996) (products (item_name PC) (quantity 1) (amount (<= 2500)) Product Specifications layer (specifications (main_memory (>= 16MB)) (hard_disk (>= 840MB)) (processor (>= Pentium90)))))))

  41. Application of Agents • Filter information • Match people with similar interests • Automate repetitive behavior • Mediator in e-commerce: Comparison • Perishables: travel, theater, and concert tickets and network bandwidth availability • Surplus Inventories: gas, electricity, pencils, music, and books • Buyer agent: Monitor consumption quantity and usage patterns, and collect, interpret information on merchants and products, making decision about merchants and products, ultimately entering purchase and payment

  42. Types and Applications of Agents • Comparison Shopping • Multi-agent System • E-Commerce: Buyer and Seller • Scheduling • Mobile Agent • Assistant Agent: Personal Agent “Sylvie”, Portico • Voice Synthesis : (www.genmagic.com/portico[Closed]) • Natural Language Processing: • www.extempo.com[Closed], • More focused Intelligence : www.neuromedia.com • User Interface Agent • AdEater(www.cs.ucd.ie/staff/nick/research/ae[Closed]) • Intelligent Agents • Portals of robots: www.botsport.com[Closed], • References:www.csee.umbc.edu/agentsecommerce.media.mit.edu [Closed] • E-Mail Filtering: Maxims

  43. Applications of Agents • Music Recommendation/ Learning 1.  Firefly: Customer taste learning [Q : How to learn? => Data Mining] 2.  Similarity Engine (WWW.ari.net/ se) Compare with five recommended musics