Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Telecommunication Technology and Management Chapter 10 and Chapter 11 Thank you for all information and pictures referred in this lecture. www.att.com and www22.verizon.com Agenda Old Technology Modern Technologies Digital TV Cell Phone: Calling Features Hi-speed Internet Network Management Expert System Knowledge Discovery and Data Mining Other Intelligent Applications 2 Old Technologies http://www.supermegatrolled.com/just-nokia-3310-destroysthe-dinosaurs/ http://www.prattonline.com/ResponsePoint.htm 3 Trend of Communication 4 Trend of Communication 5 Modern Technologies Digital TV TV Wireless Receiver Requirements: Wireless service from the Wireless Access Point to the Wireless Receiver Power outlet and connection of Wireless Receiver to TV 6 7 Modern Technologies Home DVR Ability to schedule recordings and pause live TV Record up to four shows at once on a single DVR and play them back in any room Pause, fast-forward and rewind live or recorded shows on any TV or pause your recorded show in one room and pick it up in another Play the same recorded shows on different TVs at one time and control them separately 8 Modern Technologies Interactive Applications (between TV and smart phone) Browse available TV content for current or future viewing, search for content, and remotely manage recordings of shows and movies on DVR. With qualifying TV plans, download select popular TV episodes/series from the Mobile Library to select smartphones for viewing on the go. Once a show is downloaded, you can watch it anytime! 9 Modern Technologies TV Multiview It allows you to choose your own Multiview channels you want to watch. 10 Modern Technologies Phone Services on TV We can see who's calling without leaving the comfort of their couch! Caller ID notifications and Message Waiting Indicators (MWI) will display on their TV 11 Modern Technologies Calling Features Call Screening – Only accept calls from a list of phone numbers you select. Call Transfer – Send a call that’s already in progress to a different phone number. Locate Me - Provides simultaneous ringing on up to four wireless/landline numbers when someone calls your home phone. Call History – View a list of your recent calls, by date and time, either online or on your TV screen. Click to Call – Return a phone call using your Uverse TV screen and remote. Caller ID on TV – See Caller ID and Voice Mail notifications on your TV screen. 12 Modern Technologies Online Photos and Music Customers can now view the photos they've uploaded at www.flickr.com on channel 91. Share music and photos from networked Windows PCs to TV 13 Modern Technologies TV for Xbox 360 Access your DVR recordings from your existing Xbox 360. Switch seamlessly from playing games to watching TV—without switching video inputs on your TV. Access the U-verse TV Menu, Guide, and a robust library of On Demand programming. Chat: Know instantly when your friends are on Xbox LIVE®. Use Xbox IM and Chat while watching TV. 14 Modern Technologies: Customer Support 15 Modern Technologies Mobile Phone 16 Modern Technologies Digital Voice: Calling Features 17 Modern Technologies Digital Voice: Calling Features 18 Modern Technologies Hi-Speed Internet 19 Network Solution for Business 20 Fiber Optic Technology!! 21 Network Management Telecommunication networks are extremely complex systems requiring High reliability High availability The effective management of networks is a critical, but complex, task. Telecommunications industry has heavily invested in intelligent technologies. Telecommunication industry has relied on intelligent solutions to help manage telecommunication networks. http://www.qbase.gr/en/node/124 22 Network Management Building intelligent applications involved acquiring valuable telecommunication knowledge Human experts Applying this knowledge: an expert system. This knowledge acquisition process is so timeconsuming that it is referred to as the “knowledge acquisition bottleneck”. Data mining techniques are now being applied to industrial applications to break this bottleneck, Replacing the manual knowledge acquisition process with automated knowledge discovery. 23 Data Mining http://www.laits.utexas.edu/~anorman/BUS.FOR/course.mat/Alex/ 24 Expert Systems Expert systems are programs which represent and apply factual knowledge of specific areas of expertise to solve problems Require a knowledge engineer to acquire knowledge from the domain experts Encode knowledge in a rule-based expert system These rules were very “ad-hoc” and as the number of rules increased Expert system became more difficult to understand and modify 25 Expert Systems The design of telecommunication expert systems needs to recognize all telecommunication equipment incorporates self-diagnostic capabilities http://www.nextnine.com 26 Knowledge Discovery and Data Mining Knowledge discovery is a field which has emerged from various disciplines, Artificial intelligence, Machine learning, Statistics, Databases. Its process involves identifying valid, novel, potentially useful and ultimately understandable patterns in data Data mining, the most researched topic in this process, Finding interesting patterns in the data via data analysis and discovery algorithms. 27 Knowledge Discovery and Data Mining The knowledge discovery process Data preparation: selecting, cleaning and preprocessing the data (e.g., filling in missing values) and transforming it so that it is suitable for data mining Data mining: finding patterns in the data Interpretation and evaluation: interpreting and evaluating the patterns produced by data mining 28 Knowledge Discovery and Data Mining A key motivation for knowledge discovery Replace or minimize the need for the time-consuming process of manually acquiring knowledge from a domain expert. Knowledge discovery is especially attractive to the telecommunications industry since: Telecommunication networks are typically too complex to build complete simulation models Huge quantities of data are routinely available Domain experts often are not aware of subtle patterns in data and hence automated knowledge discovery can acquire new, previously unknown, knowledge 29 Network Management Applications Max & Opti-Max: Locating Problems in the Local Loop The Max (Maintenance administrator expert) system diagnoses customer reported telephone problems in the local loop, the final segment of the telephone network that connects the customer to a central office Max is a rule-based expert system Diagnoses problems based on results of an electrical test on the customer’s phone line, Specific knowledge of the customer’s phone line and general equipment knowledge. Max determines where the trouble lies and selects the type of technician to solve the problem. 30 Network Management Applications Max & Opti-Max: Locating Problems in the Local Loop Problem of Max its performance is affected by the local characteristics of each site and thus numerous rule parameters must be tuned to optimize its performance. This tuning process is time consuming and for this reason a system called Opti-Max was created to automatically tune these parameters to appropriate values. Opti-Max takes as input a set of training examples, Problem description and a diagnosis assigned by an expert, Uses a hill-climbing search to find a set of parameter values which perform well on these examples. Opti-Max performs a type of automated knowledge discovery. 31 Network Management Applications Trouble Locator: Locating Cable Network Troubles It determines the location of troubles in a local telephone cable network Data generated by a nightly automated test to help narrow down potential cables or network equipment which may be faulty; The Trouble Locator uses a Bayesian network and Bayesian inference to solve this problem. Test results are not sufficient to determine the exact cause. The system begins by generating a local plant topology graph and then from this generates a Bayesian network, where each node in the network contains state information (belief of failure) of a plant component. This system is used by preventative maintenance analysts as a decision support system. 32 Network Management Applications TASA: Finding Frequently Occurring Alarm Episodes The Telecommunication Network Alarm Sequence Analyzer (TASA) System for extracting knowledge about the behavior of the network from a database of telecommunication network alarms. The goal of this system To locate regularities in the alarm sequences in order to filter redundant alarms Locate problems in the network Predict future faults 33 Network Management Applications TASA operates in two phases First phase: specialized algorithms are used to find rules that describe frequently occurring alarm episodes from the sequential alarm data An example rule describing an alarm episode is: if alarms of types A and B occur within 5 seconds, then an alarm of type C occurs within 60 seconds with probability 0.7. 34 Network Management Applications Second phase: collections of episodes are interactively manipulated by the user Interesting episodes from the original set can be found TASA supports this process by providing operations to prune uninteresting episodes Order the set of episodes Group similar episodes 35 Network Management Applications Scout: Identifying Network Faults via Data Mining It operates by mining historical telecommunication data Machine learning Correlation techniques. Scout identifies patterns of chronic problems directly from the data by examining the network behavior over periods of days and weeks. 36 Other Intelligent Applications APRI: Predicting Uncollectible Debt The Advanced Pattern Recognition and Identification (APRI) system The output of APRI is fed into a decision support system which can take a variety of actions To predict the probability of uncollectible debt based on historical data, including data of past uncollectibles Blocking a call from being completed. APRI automatically constructs Bayesian network models for classification problems using extremely large databases. 37 ANSWER: A Hybrid Approach to Network Management Automatic Network Surveillance with Expert Rules (ANSWER) ANSWER utilizes both rule-based and objectoriented technologies Employing a rule-based extension to the C++ object-oriented programming language. 38 Forecasting Telecommunication Equipment Failures from Time Series Data Errors may occur during the transmission of data over the network, These errors can be detected and the data rerouted through alternate paths. The effect of the failure of a single component is limited due to the redundancy in modern large-scale telecommunications networks. Modern telecommunication equipment contains self-diagnostic testing capabilities. When any of these tests fail, an alarm message is sent to a centralized site, where it may be handled by a human or by an expert system 39 Challenging Works Existing Researches: http://www.research.att.com/evergreen/what_we_do /research.html?fbid=A_Kn38ajPF9#Computing and Communications Foundations Existing Software's: http://www.research.att.com/export/sites/att_labs/sof tware_tools/index.html?fbid=A_Kn38ajPF9 40 Telecommunication Union ITU (International Telecommunication Union) is the United Nations specialized agency for information and communication technologies – ICTs. สหภาพโทรคมนาคมระหว่างประเทศ http://www.itu.int/en/Pages/default.aspx National Broadcasting and Telecommunications Commission (NBTC) คณะกรรมการกิจการกระจายเสี ยง กิจการโทรทัศน์และกิจการโทรคมนาคมแห่งชาติ http://www.nbtc.go.th/wps/portal/NTC/eng 41 References http://www.att.com/ Gary Weiss, John Eddy, Sholom Weiss, “INTELLIGENT TELECOMMUNICATION TECHNOLOGIES”, AT&T Labs, AT&T Corporation, United States http://www22.verizon.com/home/aboutfios/ http://www.research.att.com/evergreen/what_we_do/research.html?f bid=A_Kn38ajPF9#Computing and Communications Foundations http://www.research.att.com/export/sites/att_labs/software_tools/ind ex.html?fbid=A_Kn38ajPF9 42