Abstract
We present a new method for deriving functions that model the relationship between multiple signals in a physical system. The method, which we call dimensional function synthesis, applies to data streams where the dimensions of the signals are known. The method comprises two phases: a compile-time synthesis phase and a subsequent calibration using sensor data.
We implement dimensional function synthesis and use the implementation to demonstrate efficiently summarizing multi-modal sensor data for two physical systems using 90 laboratory experiments and 10 000 synthetic idealized measurements. We evaluate the performance of the compile-time phase of dimensional function synthesis as well as the calibration phase overhead, inference latency, and accuracy of the models our method generates.
The results show that our technique can generate models in less than 300 ms on average across all the physical systems we evaluated. When calibrated with sensor data, our models outperform traditional regression and neural network models in inference accuracy in all the cases we evaluated. In addition, our models perform better in training latency (over 8660× improvement) and required arithmetic operations in inference (over 34× improvement). These significant gains are largely the result of exploiting information on the physics of signals that has hitherto been ignored.
- Eric Allen, David Chase, Victor Luchangco, Jan-Willem Maessen, and Guy L. Steele, Jr. 2004. Object-oriented units of measurement. In Proceedings of the 19th Annual ACM SIGPLAN Conference on Object-oriented Programming, Systems, Languages, and Applications (OOPSLA’04). ACM, New York, NY, USA, 384--403.Google Scholar
- Tudor Antoniu, Paul A. Steckler, Shriram Krishnamurthi, Erich Neuwirth, and Matthias Felleisen. 2004. Validating the unit correctness of spreadsheet programs. In Proceedings of the 26th International Conference on Software Engineering (ICSE’04). IEEE Computer Society, Washington, DC, USA, 439--448.Google Scholar
Digital Library
- Michel Babout, Hafid Sidhoum, and Louis Frecon. 1990. AMPERE: A programming language for physics. European Journal of Physics 11, 3 (1990), 163.Google Scholar
Cross Ref
- David Barber. 2012. Bayesian Reasoning and Machine Learning. Cambridge University Press.Google Scholar
Digital Library
- Geoffrey Biggs and Bruce A. Macdonald. 2008. A pragmatic approach to dimensional analysis for mobile robotic programming. Auton. Robots 25, 4 (Nov. 2008), 405--419.Google Scholar
Digital Library
- Edgar Buckingham. 1914. On physically similar systems; Illustrations of the use of dimensional equations. Physical Review 4, 4 (1914), 345--376.Google Scholar
Cross Ref
- Donald E. Carlson. 1979. A mathematical theory of physical units, dimensions, and measures. Archive for Rational Mechanics and Analysis 70, 4 (1979), 289--305.Google Scholar
Cross Ref
- Feng Chen, Grigore Roşu, and Ram Prasad Venkatesan. 2003. Rule-based analysis of dimensional safety. In Proceedings of the 14th International Conference on Rewriting Techniques and Applications (RTA’03). Springer-Verlag, Berlin, Heidelberg, 197--207.Google Scholar
Digital Library
- Robert F. Cmelik and Narain H. Gehani. 1988. Dimensional analysis with C++. IEEE Softw. 5, 3 (May 1988), 21--27.Google Scholar
Digital Library
- Donald E. Carlson. 1978. On some new results in dimensional analysis. (1978).Google Scholar
- Harald Hanche-Olsen. 2004. Buckingham’s pi-theorem. NTNU: http://www.math.ntnu.no/˜hanche/notes/buckingham/buckingham-a4.pdf (2004).Google Scholar
- Paul N. Hilfinger. 1988. An ada package for dimensional analysis. ACM Trans. Program. Lang. Syst. 10, 2 (April 1988), 189--203.Google Scholar
Digital Library
- Daniel J. A. Hills, Adrian M. Grütter, and Jonathan J. Hudson. 2015. An algorithm for discovering Lagrangians automatically from data. PeerJ Computer Science 1 (nov 2015), e31.Google Scholar
- Mark Hills, Feng Chen, and Grigore Roşu. 2012. A rewriting logic approach to static checking of units of measurement in C. Electron. Notes Theor. Comput. Sci. 290 (Dec. 2012), 51--67.Google Scholar
- R. T. House. 1983. A proposal for an extended form of type checking of expressions. Comput. J. 26, 4 (Nov. 1983), 366--374.Google Scholar
Digital Library
- Dan Jonsson. 2014. Dimensional analysis: A centenary update. 1 (2014), 1--16.Google Scholar
- Andrew Kennedy. 1994. Dimension types. In Proceedings of the 5th European Symposium on Programming: Programming Languages and Systems (ESOP’94). Springer-Verlag, London, UK, UK, 348--362.Google Scholar
Digital Library
- Andrew J. Kennedy. 1997. Relational parametricity and units of measure. In Proceedings of the 24th ACM SIGPLAN-SIGACT Symposium on Principles of Programming Languages (POPL’97). ACM, New York, NY, USA, 442--455.Google Scholar
Digital Library
- Patrick W. Langley. 1977. BACON: A production system that discovers empirical laws. In Proceedings of the 5th International Joint Conference on Artificial Intelligence - Volume 1 (IJCAI’77). San Francisco, CA, USA, 344.Google Scholar
- Jonathan Lim and Phillip Stanley-Marbell. 2018. Newton: A language for describing physics. CoRR abs/1811.04626 (2018).Google Scholar
- Lord Rayleigh. 1915. The principle of similitude. Nature 95 (dec 1915), 66--68.Google Scholar
- William H. Press, Saul a. Teukolsky, William T. Vetterling, and Brian P. Flannery. 1996. Numerical Recipes in Fortran 77: The Art of Scientific Computing. Second Edition. Vol. 1.Google Scholar
- Mikael Rittri. 1995. Dimension inference under polymorphic recursion. In Proceedings of the Seventh International Conference on Functional Programming Languages and Computer Architecture (FPCA’95). ACM, New York, NY, USA, 147--159.Google Scholar
Digital Library
- Samuel H. Rudy, Steven L. Brunton, Joshua L. Proctor, and J. Nathan Kutz. 2017. Data-driven discovery of partial differential equations. Science Advances 3, 4 (2017), e1602614.Google Scholar
Cross Ref
- Michael Schmidt and Hod Lipson. 2009. Distilling free-form natural laws from experimental data. Science 324, 5923 (2009), 81--85. DOI:https://doi.org/10.1126/science.1165893Google Scholar
- Volker Simon, Bernhard Weigand, and Hassan Gomaa. 2017. Dimensional Analysis for Engineers. Springer.Google Scholar
- Ain A. Sonin. 2004. A generalization of the Π-theorem and dimensional analysis. Proceedings of the National Academy of Sciences 101, 23 (2004), 8525--8526.Google Scholar
Cross Ref
- Gilbert Strang. 2016. Introduction to Linear Algebra (fifth ed.). Wellesley-Cambridge Press, Wellesley, MA.Google Scholar
- G. Taylor. 1950. The formation of a blast wave by a very intense explosion. II. The atomic explosion of 1945. Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences 201, 1065 (1950), 175--186.Google Scholar
Cross Ref
- Zerksis D. Umrigar. 1994. Fully static dimensional analysis with C++. SIGPLAN Not. 29, 9 (Sept. 1994), 135--139.Google Scholar
Digital Library
- Yihua Wu. 1989. Discovering natural laws by reduction. Journal of Computer Science and Technology 4, 1 (1989), 35--51.Google Scholar
Digital Library
Index Terms
Deriving Equations from Sensor Data Using Dimensional Function Synthesis
Recommendations
Deriving equations from sensor data using dimensional function synthesis
We present a new method for deriving functions that model the relationship between multiple signals in a physical system. The method, which we call dimensional function synthesis, applies to data streams where the dimensions of the signals (e.g., length,...
Low-Dimensional Data Representation in Data Analysis
Proceedings of the 6th IAPR TC 3 International Workshop on Artificial Neural Networks in Pattern Recognition - Volume 8774Many Data Analysis tasks deal with data which are presented in high-dimensional spaces, and the 'curse of dimensionality' phenomena is often an obstacle to the use of many methods, including Neural Network methods, for solving these tasks. To avoid ...
Enhanced algorithm for high-dimensional data classification
Graphical abstractIllustration of the decision hyperplanes generated by TSSVM, MCVSVM, and LMLP on an artificial dataset. Display Omitted HighlightsIn the case of the singularity of the within-class scatter matrix, the drawbacks of both MCVSVM and LMLP ...






Comments