Page 310 - Artificial Intelligence for the Internet of Everything
P. 310

288   Index


          Local infeasibility, 249          Mini-batch SGD, 26–28, 27b
          Logistic regression, 25–26        Minimum free-energy formalism, 84
                                            Mission-based cybersecurity, 92–93
                                            ML. See Machine learning (ML)
          M                                 Modeling probability distributions, 57
          Machine learning (ML), 19, 279–280  Models that are authorized to act., 126
            algorithms, 56                  Monitoring subsystem, 87–88
            optimization in, 22–31
                                            Motivation, 171–173
              generalized linear models, 29–30
                                            MTurk. See Mechanical Turk (MTurk)
              inference, 30–31              Multi-agent networks, 277
              logistic regression, 25–26    Multiagent teams, application of energy
              NAGD, 28–29                        formalism to
              problem, 23–24                  adapting team structure, 79–80
              SGD variants, 26–28             distributed collaborative search via free
              Stochastic Gradient Descent algorithm,  energy minimization, 76–79
                24–25                         motivation, 73–74
            systems, 5                        problem definition, 74–76
            uncertainty quantification in, 32–33  validation experiments
              deep neural network, 35–37       agent network structure, impact of,
              GPR, 33–34                         81–82
              neural network, 34–35            decision decomposition, impact of,
          Machine partners absent agency, 103    82–83
          Managing learned behavior, 128       discrete decision making vs. free
          Manually entered data, 126–127         energy, 81
          MAP estimation. See Maximum a posteriori  experiment setup, 80–81
                (MAP) estimation            Multiple affective argumentation patterns,
          Market incentives, 97–98               188–189
          Maximum a posteriori (MAP) estimation,  Multiple methods, 38–39
                77–78                       Multiple vendors, 225
          Maximum entropy production (MEP),  Multiplicity of perspectives, 245
                2–3                         Mutual interdependence, 104
          Max-product algorithm, 77–79
                                            MX3D bridge, 141f
          Max-product belief propagation algorithm,
                                            “myfitnesspal.com,”, 121
                77–78
          Mechanical Turk (MTurk), 106
          Media richness, 105               N
          Mental health, 122                NAGD. See Nesterov’s accelerated gradient
          MEP. See Maximum entropy production    descent (NAGD)
                (MEP)                       National Energy Regulatory Commissions’
          Merkle tree, 221–222                   Critical Infrastructure Protection
          Message passing in belief propagation, 78f  (NERC-CIP) standard, 221
          Message validation, 212–213       National Highway Safety Traffic System
          Meta agent-based emergent configuration  (NHTSA), 97–98
                systems, 273                National Institute of Standards, 88–89
          Meta-analysis, 102                National Institute of Standards and
          Metropolis-Hastings-based acceptance   Technology (NIST), 97–98
                scheme, 35–36               Negative dataset, 210
          Military-focused programs, 60     Neologism, 180
   305   306   307   308   309   310   311   312   313   314   315