Evolving Future Information Systems: Challenges, Perspectives and Applications Ajith Abraham Machine Intelligence Research Labs (MIR Labs), USA

14. Service Level Agreements Negotiation Client Provider Can you do X for me for Y in return? No SLA SLA Can you do Z for me for Y in return? Negotiation Phase (Single or Multi-Round) SLA-Offer SLA-CounterOffer SLA-Offer

15. SLA Variations Client Providers SLA SLA Multi-provider SLA Single SLA is divided across multiple providers SLA dependencies For an SLA to be valid, Another SLA has to be agreed Client Providers SLA

16. Future Information Systems System has to be autonomous and able to continuously adapt, providing the required quality of service levels according to different service level agreements, without requiring the need of much human intervention. The challenge is to design intelligent machines and networks that could communicate and adapt according to critic or error information, self organize and resilient in case of a system, service or component failure due to natural cause or a malicious attack.

17. Problem Complexity

18. Parking a Car Generally, a car can be parked rather easily. If it were specified to within, say, a fraction of a millimeter, it would take hours of maneuvering and precise measurements of distance and angular position to solve the problem.  High precision carries a high cost  The challenge is to exploit the tolerance for imprecision by devising methods of computation which lead to an acceptable solution at low cost. This, in essence, is the guiding principle of modern intelligent computing.

20. In nature it works… Why not for our digital ecosystem?

21. Artificial Neural Network Artificial neuron Mammalian neuron

23. Rough Set – Zdzisław Pawlak The rough set concept overlaps—to some extent—with many other mathematical tools developed to deal with vagueness and uncertainty, in particular with the Dempster-Shafer theory of evidence. Rough set does not compete with fuzzy set theory, with which it is frequently contrasted, but rather complements it. One of the main advantages of rough set theory is that it does not need any preliminary or additional information about data, such as probability distribution in statistics, basic probability assignment in the Dempster-Shafer theory, or grade of membership or the value of possibility in fuzzy set theory.

27. Hybrid Approaches

28. Ant Colony Optimization

29. ACO in Real Life TSP Scheduling Clustering

30. Particle Swarm Optimization x p g p i v P best G best

32. Turbulent PSO (TPSO)

33. FATPSO – Griewank Function - 100 D

34. FATPSO – Levy Function - 100 D

35. FATPSO – Schwefel Function - 100 D

36. Smart System Application Example Improving the delivery of health care in geriatric residences

37. The Problem.. Over the past 30 years, the number of Europeans over 60 years of age has risen by about 50 percent, and now represents more than 25 percent of the population. Within 20 years, experts estimate that this percentage will rise to one-third of the population. Creating secure, unobtrusive, and adaptable environments for monitoring and optimizing healthcare will become vital in the near future. Dynamic: New patients arrive and others pass away! While the staff rotation is also relatively high and they normally work in shifts of eight hours.

39. The Environment Alzheimer Santísima Trinidad Residence of Salamanca, Spain The Residence is for Alzheimer’s patients over 65 years old. Its services and facilities include medical service, including occupational therapy and technical assistance Comprises of : a terrace and a garden; laundry and tailoring services; a hairdressing salon; a chapel and religious services; a cafeteria and various rooms, including a geriatric bathroom, a multipurpose room, and separate rooms for reading, socializing, visiting with guests, and watching TV.

40. Technologies Used Multi-agent system, which is a dynamic system for the management of different aspects of the geriatric center. Radio Frequency Identification (RFID) technology for ascertaining patients’ location. Mobile devices and Wi-Fi technology to provide the personnel of the residence with updated information about the center and the patients, to provide the working plan, information about alarms or potential problems and to keep track of their movements and actions within the center. From the user’s point of view the complexity of the solution has been reduced with the help of friendly user interfaces and a robust and easy to use multi-agent system.

41. Technologies Used System uses microchips mounted on bracelets worn on the patient ’ s wrist or ankle, and sensors installed over protected zones, with an adjustable capture range up to 2 meters. The microchips or transponders use a 125 kHz signal to locate the patients, which can be identified by consulting the software agents installed in PDA’s.

42. Software Architecture Patient : monitoring, location, daily tasks, and anomalies Doctor: treats patients Nurse: schedules the nurse ’ s working day obtaining dynamic plans depending on the tasks needed for each assigned patient Security: controls the patients ’ location and manages locks and alarms Manager: manages the medical record database and the doctor-patient and nurse-patient assignment

43. Software Architecture

44. Patient Agent The beliefs that were seen to define a general patient state: weight, temperature, blood pressure, feeding (diet characteristics and next time to eat), medication, posture change, toileting, personal hygiene, and exercise. The beliefs and goals for every patient depend on the plan or plans corresponding to the treatments or medicine that the doctors prescribe. The patient agent must have periodic communication with the doctor and nurse agent. Must ensure that all the actions indicated in the treatment are fullfiled.

45. Patient Agent Manager and Patient agents run in a central computer, but GerAg agents run on mobile devices, so a robust wireless network has been installed as an extension to the existing wired LAN. With respect to the question of failure recovery, a continuous monitoring of the system is carried out. Every agent saves its memory (personal data) onto a data base. The most sensitive agents are patient agents, so these agents save their state every hour. When an agent fails, another instance can be easily created from the latest backup

46. Agent System Manager and Patient agents run in a central computer, but other agents run on mobile devices Every agent saves its memory (personal data) onto a data base. The most sensitive agents are patient agents, so these agents save their state every hour. When an agent fails, another instance can be easily created from the latest backup

47. System Interfaces Manager Nurse

48. How Effective?

49. Automatic Design of Fuzzy Systems As a way to overcome the curse-of-dimensionality, it was suggested to arrange several low-dimensional rule base in a hierarchical structure, i.e., a tree, causing the number of possible rules to grow in a linear way according to the number of inputs. Building a hierarchical fuzzy system is a difficult task. This is because we need to define the architecture of the system (the modules, the input variables of each module, and the interactions between modules), as well as the rules of each modules.

51. Automatic Design of Hierarchical Takagi-Sugeno Type Fuzzy Systems

53. Encoding Assume that the used instruction set is I={+2, +3, x1, x2, x3, x4, where +2 and +3 denote non-leaf nodes' instructions taking 2 and 3 arguments, respectively. x1, x2, x3, x4 are leaf nodes' instructions taking zero arguments each.

54. Comparison of the incremental type multilevel FRS (IFRS), aggregated type mutilevel FRS (AFRS), and the hierarchical TS-FS for Mackey-Glass time-series prediction Model layer No. of rules No. of para. RMSE(train) RMSE(Test) IFRS 4 25 58 0.0240 0.0253 AFRS 5 36 78 0.0267 0.0256 HTS-FS 3 24 33 0.0179 0.0167

55. The structure of the evolved hierarchical TS-FS model for predicting of Mackey-Glass time-series The importance degree of each input variables for Mackey-Glass time-series xi x 0 x 1 x 2 x 3 x 4 x 5 Impo ( xi ) 0.247 0.332 0.072 0.113 0.056 0.180

57. The developed optimal H-TS-FS architectures (Irisdata)

59. The developed optimal H-TS-FS architectures (Wine data)

60. Flexible Neural Trees

61. Flexible Neural Trees

62. Flexible Neural Trees

63. Flexible Neural Trees

64. Flexible Neural Trees

65. Intrusion Detection

66. Flexible Neural Trees - IDS

67. FNT– Colon Cancer / Leukemia

68. FNT– Colon Cancer / Leukemia

69. MIMO - FNT

70. Flexible Radial Basis Function Trees

71. Flexible Radial Basis Function Trees Breast Cancer Detection

75. Establish the Context Identify Analyze Evaluate Treatment Monitoring and Review Communication and Consultation Risk Management in Practice

76. Risk Components Threat Vulnerability Asset Value

77. What is Threat? Threat is anything that is capable of acting against an asset in a manner that can result in harm. A tornado is a threat, as is a hacker. The key consideration is that threats apply the force (eg: exploit code) against an asset that can cause a loss event to occur. Threat level depends on many factors: (1) frequency of attacks (2) probability that attack being successful (3) Type and severity of attack.

78. What is Vulnerability? Weakness that may be exploited! A condition in which threat capability (force) is greater than the ability to resist that force. Vulnerability is always dependent upon the type and level of force being applied Vulnerability depends upon: (1) threat capability and (2) system threat resistance

79. What is Asset? Asset can be any data, device, or other component of the environment that supports information-related activities, which can be illicitly accessed, used, disclosed, altered, destroyed, and/or stolen, resulting in loss. Even ‘’reputation’ is an asset Asset value/loss depends upon: (1) cost (2) criticality (3) sensitivity and (4) recovery.

80. Risk Assessment – A Soft Approach There is no such thing as an “exact” value of risk. The advantage of the fuzzy approach is that it enables processing of vaguely defined variables, and variables whose relationships cannot be defined by mathematical relationships. Fuzzy logic can incorporate expert human judgement to define those variable and their relationships. The model can be closer to reality and network specific than that by some of the other methods.

82. Fuzzy modeling of Risk

83. Takagi-Sugeno Neuro-Fuzzy System

84. Hierarchical Takagi-Sugeno Type Fuzzy Systems

86. Developed Fuzzy Risk Assessment Model

88. Program obtained by the best individual is as follows: (cos(exp(1 / (log10( x[5] )) + x[0] ))) * (tan(1 / (exp(cos((tan( x[5] )) * x[0] ))))) It is to be noted that, not all variables are required for building the risk assessment models. For example, the best individual used only 3 input variables, while the worst individual required just 2 input variables. What Genetic Programming Can Do?

89. How Can We Teach Things to Computers? In order for a program to be capable of learning something, it must first be capable of being told it. John McCarthy Easy: If dogs are mammals and mammals are animals, are dogs mammals? Difficult: If most Canadians have brown eyes, and most brown eyed people have good eyesight, then do most Canadians have good eyesight? We Divide Things Into Concepts

93. Q A & Thank You [email_address]