Abstract
This is a review of the works of the research school of V.A. Yakubovich in the field of artificial intelligence, machine learning, adaptive systems, and robotics. The method of recurrent objective inequalities is considered in detail. The significance of the presented results for the further development of cybernetics and artificial intelligence is discussed. Special emphasis is put on Yakubovich’s seminal works on machine learning and development of the concept of finitely convergent algorithms for solving recurrent objective inequalities. Typical results of their convergence are discussed in detail as a specific illustration. The study distinguishes the contribution of the school to the formation and development of modern theories of adaptive control and mathematical robotics, particularly, the theory of adaptive robots. A special section is devoted to adaptive suboptimal control.
REFERENCES
A. L. Fradkov, “V. A. Yakubovich’s scientific school of theoretical cybernetics at the St. Petersburg (Leningrad) University,” in The History of Computer Science and Cybernetics in St. Petersburg (Leningrad), Ed. by R. M. Yusu-pov (Nauka, St. Petersburg, 2008), Vol. 1, pp. 79–83 [in Russian].
A. L. Fradkov and A. I. Shepeljavyi, “The history of cybernetics and artificial intelligence. a view from Saint Petersburg,” Cybern. Phys. 11, 253–263 (2022).
S. V. Gusev and V. A. Bondarko, “Notes on Yakubovich’s method of recursive objective inequalities and its application in adaptive control and robotics,” in Proc. IFAC World Congr. 2020, Berlin, Germany, July 12–17, 2020 (International Federation of Automatic Control, New York, 2020).
A. L. Fradkov, “Early history of machine learning,” in Proc. IFAC World Congr. 2020, Berlin, Germany, July 12–17, 2020 (International Federation of Automatic Control, New York, 2020).
A. M. Annaswamy and A. L. Fradkov, “Historical perspective of adaptive control and learning,” Annu. Rev. Control 52, 18–41 (2021). https://doi.org/10.48550/arXiv.2108.11336
A. L. Fradkov and B. T. Polyak, “Adaptive and robust control in the USSR,” IFAC-PapersOnLine 53 (2), 1373–1378 (2020). https://doi.org/10.1016/j.ifacol.2020.12.1882
V. A. Yakubovich, “Machines learning pattern recognition.” in Methods of Computation (Leningr. Gos. Univ., Leningrad, 1963), Vol. 2, pp. 95–131 [in Russian].
V. A. Yakubovich, “Machines learning pattern recognition,” Vestn. St. Petersburg Univ.: Math. 54, 384–394 (2021); Vestn. St. Petersburg Univ.: Math. 55, 71–86 (2022). https://doi.org/10.1134/S106345412201015010.1134/S1063454122010150https://doi.org/10.1134/S106345412104021X
V. A. Yakubovich, “Some general theoretical principles of the construction of trainable identification systems. I.” in Computing Machines and Problems of Programming (Leningr. Gos. Univ., Leningrad, 1965), pp. 3–71 [in Russian].
V. A. Yakubovich, “Finitely-convergent recurrent algorithms for solving objective inequalities,” Dokl. Akad. Nauk SSSR 166, 1308–1312 (1966).
V. A. Yakubovich, “Theory of adaptive systems,” Sov. Phys. Dokl. 13, 852–856 (1968).
V. A. Yakubovich, “Adaptive systems with multistep goal conditions,” Dokl. Akad. Nauk SSSR 183, 303–306 (1968).
A. Kh. Gelig and V. A. Yakubovich, “Application of a trainable recognition system to isolate a signal from noise,” in Computing Machines and Problems of Cybernetics (Leningr. Gos. Univ., Leningrad, 1968), Vol. 5, pp. 95–100 [in Russian].
V. A. Yakubovich, “On certain problem of self-learning expedient behaviour,” Autom. Remote Control 30, 1292–1310 (1969).
V. A. Yakubovich, “On organization of the "brain” of a certain class of systems that develop appropriate behavior (solved and unsolved problems),” in Proc. 4th All-Union Conf. on Neurocybernetics, Rostov, 1970 (Rostov. Gos. Univ., 1970), p. 152.
G. D. Penev and V. A. Yakubovich, “On some tasks of adaptive behavior,” Dokl. Akad. Nauk SSSR 198, 787–790 (1971).
V. A. Yakubovich and A. V. Timofeev, “On a class of self-learning systems possessing appropriate behavior,” in Control and Informational Process in Nature (Nauka, Moscow, 1971), pp. 111–113 [in Russian].
A. V. Timofeev, V. V. Kharichev, A. A. Shmidt, and V. A. Yakubovich, “One task of image recognition and description,” in Biological, Medical Cybernetics and Bionics (Nauchn. Sov. Kibern. i Inst. Kibern., Kiev, 1971) [in Russian].
S. V. Gusev, A. V. Timofeev, and V. A. Yakubovich, “Adaptation in robotic systems with artificial intelligence,” in Proc. 7th All-Union Meeting on Control Problems, Minsk, 1977, pp. 279–282.
V. N. Vapnik, “Machines learning pattern recognition,” in Pattern Recognition Learning Algorithms (Sov. Radio, Moscow, 1973), pp. 5–24 [in Russian].
B. N. Kozinets, R. M. Lantsman, and V. A. Yakubovich, “Forensic examination of close handwriting using electronic computers,” Dokl. Akad. Nauk SSSR 167, 1008–1011 (1966).
B. N. Kozinets, “About a linear perceptron learning algorithm,” in Computing Machines and Problems of Programming (Leningr. Gos. Univ., Leningrad, 1964), Vol. 3, pp. 80–83.
B. F. Mitchell, V. F. Dem’yanov, and V. N. Malozemov, “Finding the point closest to the origin of the polyhedron,” Vestn. Leningr. Gos. Univ., Ser. 1: Mat., Mekh., Astron., No. 19, 38–45 (1971).
B. F. Mitchel, V. V. Dem’yanov, and V. N. Malozemov, “Finding the point of a polyhedron closest to the origin,” SIAM J. Control 12, 19–26 (1974).
V. N. Malozemov, “On the fortieth anniversary of MDM-method,” Vestn. Syktyvkarskogo Univ., No. 15, 51–62 (2012).
V. N. Vapnik and A. Ya. Chervonenkis, “On a certain class of perceptrons,” Avtom. Telemekh. 25, 112–120 (1964).
V. A. Yakubovich, “ Finitely convergent algorithms for the solution of countable systems of inequalities, and their application in problems of the synthesis of adaptive systems,” Sov. Phys. Dokl. 14, 1051–1054 (1970).
V. N. Fomin, “Stochastic analogs of finitely convergent learning algorithms for recognition systems,” in Computing Machines and Problems of Programming (Leningr. Gos. Univ., Leningrad, 1971), Vol. 6, pp. 68–87 [in Russian].
A. L. Fradkov, “Some finitely converging solution algorithms for infinite systems of inequalities and their application in the theory of adaptive systems,” Vest. Leningr. Gos. Univ., Ser. 1: Mat., Mekh. Astron., No. 19, 70–75 (1972).
V. N. Fomin, Mathematical Theory of Trainable Recognition Systems (Leningr. Gos. Univ., Leningrad, 1976) [in Russian].
D. P. Derevitskii and A. L. Fradkov, Applied Theory of Discrete Adaptive Control Systems (Nauka, Moscow, 1981) [in Russian].
V. A. Bondarko and V. A. Yakubovich, “The method of recursive aim inequalities in adaptive control theory,” Int. J. Adaptive Control Signal Process. 6, 141–160 (1992).
A. L. Fradkov, Adaptive Control of Complex Systems (Nauka, Moscow, 1990) [in Russian].
S. V. Gusev, “A finite convergent algorithm for restoring the regression function and its use in adaptive control problems,” Autom. Remote Control 50, 367–374 (1989).
V. A. Bondarko, “Adaptive suboptimal systems with a variable dimension of the vector of adjustable parameters,” Autom. Remote Control 67, 1732–1751 (2006).
S. V. Gusev, A. V. Timofeev, and V. A. Yakubovich, “On a hierarchical system of integral robot control,” in Proc. 6th Joint Conf. on Artificial Intelligence, Moscow, 1975 (Inst. Probl. Upr. Akad. Nauk SSSR, Moscow, 1975), pp. 76–85.
G. G. Grigor’ev, S. V. Gusev, V. V. Nesterov, and V. A. Yakubovich, “Mobile robot-manipulator adaptive control,” in Adaptive Robots-82: Proc. All-Union Sci. and Tech. Conf., Nalchik, Sept. 14–16, 1982 (TsNPTO Priborostr. Prom-sti., Nalchik, 1982), pp. 89–91.
A. V. Timofeev, Robots and Artificial Intelligence (Nauka, Moscow, 1978) [in Russian].
V. N. Fomin, A. L. Fradkov, and V. A. Yakubovich, Adaptive Control of Dynamic Systems (Nauka, Moscow, 1981) [in Russian].
A. Kh. Gelig, Dynamics of Impulse Systems and Neural Networks (Leningr. Gos. Univ., Leningrad, 1982) [in Russian].
R. M. Granovskaya and I. Ya. Bereznaya, Intuition and Artificial Intelligence (Leningr. Gos. Univ., Leningrad, 1991) [in Russian].
A. V. Savkin, T. M. Cheng, Z. Xi, F. Javed, A. S. Matveev, and H. Nguyen, Decentralized Coverage Control Problems for Mobile Robotic Sensor and Actuator Networks (IEEE/Wiley, New York, 2015).
A. S. Matveev, A. V. Savkin, M. C. Hoy, and C. Wang, Safe Robot Navigation among Moving and Steady Obstacles (Elsevier/Butterworth Heinemann, Oxford, UK, 2016).
M. Hoy, A. S. Matveev, and A. V. Savkin, “Algorithms for collision-free navigation of mobile robots in complex cluttered environments: A survey,” Robotica 33, 463–497 (2015).
V. A. Yakubovich, “Adaptive suboptimal control of a linear dynamic object in the presence of a delay in control,” Kibernetika 1, 26–43 (1976).
V. A. Bondarko and V. A. Yakubovich, “Synthesis of a suboptimal adaptive system with a reference model for controlling a discrete linear dynamic object,” in Adaptation and Training in Control and Decision-Making Systems (Nauka, Novosibirsk, 1982), pp. 10–27 [in Russian].
F. Lewis and D. Vrabie, “Reinforcement learning and adaptive dynamic programming for feedback control,” IEEE Circ. Syst. Mag. 9, 32–50 (2009).
F. Lewis, D. Vrabie, and K. G. Vamvoudakis, “Reinforcement learning and feedback control: Using natural decision methods to design optimal adaptive controllers,” IEEE Circ. Syst. Mag. 32, 30 (2012).
R. S. Sutton, A. G. Barto, and R. J. Williams, “Reinforcement learning is direct adaptive optimal control,” IEEE Control Syst. Mag. 12, 19–22 (1992).
X. Yang, D. Liu, and D. Wang, “Reinforcement learning for adaptive optimal control of unknown continuous-time nonlinear systems with input constraints,” Int. J. Control 87, 553– 566 (2014).
B. Recht, “A tour of reinforcement learning: The view from continuous control,” Annu. Rev. Control Robot. Auton. Syst. 2, 253–279 (2019).
B. T. Bian and Z.-P. Jiang, “Value iteration, adaptive dynamic programming, and optimal control of nonlinear systems,” in Proc. IEEE 55th Conf. on Decision and Control, Las Vegas, Nev., Dec. 12–14, 2016 (IEEE, Piscataway, N.J., 2016).
M. Lipkovich, “Yakubovich’s method of recursive objective inequalities in machine learning,” IFAC-PapersOnLine 55 (12), 138–143 (2022).
Funding
This work was supported by ongoing institutional funding. No additional grants to carry out or direct this particular research were obtained.
Author information
Authors and Affiliations
Corresponding authors
Ethics declarations
The authors declare that they have no conflict of interest.
Additional information
Translated by S. Kuznetsov
Publisher’s Note.
Pleiades Publishing remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Matveev, A.S., Fradkov, A.L. & Shepeljavyi, A.I. Review of Works by V. A. Yakubovich’s Scientific School on Artificial Intelligence and Robotics. Vestnik St.Petersb. Univ.Math. 56, 478–492 (2023). https://doi.org/10.1134/S106345412304009X
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1134/S106345412304009X