-
Physical Symbolic Optimization
Authors:
Wassim Tenachi,
Rodrigo Ibata,
Foivos I. Diakogiannis
Abstract:
We present a framework for constraining the automatic sequential generation of equations to obey the rules of dimensional analysis by construction. Combining this approach with reinforcement learning, we built $Φ$-SO, a Physical Symbolic Optimization method for recovering analytical functions from physical data leveraging units constraints. Our symbolic regression algorithm achieves state-of-the-a…
▽ More
We present a framework for constraining the automatic sequential generation of equations to obey the rules of dimensional analysis by construction. Combining this approach with reinforcement learning, we built $Φ$-SO, a Physical Symbolic Optimization method for recovering analytical functions from physical data leveraging units constraints. Our symbolic regression algorithm achieves state-of-the-art results in contexts in which variables and constants have known physical units, outperforming all other methods on SRBench's Feynman benchmark in the presence of noise (exceeding 0.1%) and showing resilience even in the presence of significant (10%) levels of noise.
△ Less
Submitted 6 December, 2023;
originally announced December 2023.
-
Class Symbolic Regression: Gotta Fit 'Em All
Authors:
Wassim Tenachi,
Rodrigo Ibata,
Thibaut L. François,
Foivos I. Diakogiannis
Abstract:
We introduce 'Class Symbolic Regression' (Class SR) a first framework for automatically finding a single analytical functional form that accurately fits multiple datasets - each realization being governed by its own (possibly) unique set of fitting parameters. This hierarchical framework leverages the common constraint that all the members of a single class of physical phenomena follow a common go…
▽ More
We introduce 'Class Symbolic Regression' (Class SR) a first framework for automatically finding a single analytical functional form that accurately fits multiple datasets - each realization being governed by its own (possibly) unique set of fitting parameters. This hierarchical framework leverages the common constraint that all the members of a single class of physical phenomena follow a common governing law. Our approach extends the capabilities of our earlier Physical Symbolic Optimization ($Φ$-SO) framework for Symbolic Regression, which integrates dimensional analysis constraints and deep reinforcement learning for unsupervised symbolic analytical function discovery from data. Additionally, we introduce the first Class SR benchmark, comprising a series of synthetic physical challenges specifically designed to evaluate such algorithms. We demonstrate the efficacy of our novel approach by applying it to these benchmark challenges and showcase its practical utility for astrophysics by successfully extracting an analytic galaxy potential from a set of simulated orbits approximating stellar streams.
△ Less
Submitted 17 June, 2024; v1 submitted 4 December, 2023;
originally announced December 2023.
-
Charting the Galactic acceleration field II. A global mass model of the Milky Way from the STREAMFINDER Atlas of Stellar Streams detected in Gaia DR3
Authors:
Rodrigo Ibata,
Khyati Malhan,
Wassim Tenachi,
Anke Ardern-Arentsen,
Michele Bellazzini,
Paolo Bianchini,
Piercarlo Bonifacio,
Elisabetta Caffau,
Foivos Diakogiannis,
Raphael Errani,
Benoit Famaey,
Salvatore Ferrone,
Nicolas Martin,
Paola di Matteo,
Giacomo Monari,
Florent Renaud,
Else Starkenburg,
Guillaume Thomas,
Akshara Viswanathan,
Zhen Yuan
Abstract:
We present an atlas and follow-up spectroscopic observations of 87 thin stream-like structures detected with the STREAMFINDER algorithm in Gaia DR3, of which 29 are new discoveries. Here we focus on using these streams to refine mass models of the Galaxy. Fits with a double power law halo with the outer power law slope set to $-β_h=3$ yield an inner power law slope $-γ_h=0.97^{+0.17}_{-0.21}$, a s…
▽ More
We present an atlas and follow-up spectroscopic observations of 87 thin stream-like structures detected with the STREAMFINDER algorithm in Gaia DR3, of which 29 are new discoveries. Here we focus on using these streams to refine mass models of the Galaxy. Fits with a double power law halo with the outer power law slope set to $-β_h=3$ yield an inner power law slope $-γ_h=0.97^{+0.17}_{-0.21}$, a scale radius of $r_{0, h}=14.7^{+4.7}_{-1.0}$ kpc, a halo density flattening $q_{m, h}=0.75\pm0.03$, and a local dark matter density of $ρ_{h, \odot}=0.0114\pm0.0007 {\rm M_\odot pc^{-3}}$. Freeing $β$ yields $β=2.53^{+0.42}_{-0.16}$, but this value is heavily influenced by our chosen virial mass limit. The stellar disks are found to have a combined mass of $4.20^{+0.44}_{-0.53}\times10^{10} {\rm M_\odot}$, with the thick disk contributing $12.4\pm0.7$\% to the local stellar surface density. The scale length of the thin and thick disks are $2.17^{+0.18}_{-0.08}$ kpc and $1.62^{+0.72}_{-0.13}$ kpc, respectively, while their scale heights are $0.347^{+0.007}_{-0.010}$ kpc and $0.86^{+0.03}_{-0.02}$ kpc, respectively. The virial mass of the favored model is $M_{200}=1.09^{+0.19}_{-0.14}\times 10^{12} {\rm M_\odot}$, while the mass inside of 50 kpc is $M_{R<50}=0.46\pm0.03\times 10^{12} {\rm M_\odot}$. We introduce the Large Magellanic Cloud (LMC) into the derived potential models, and fit the "Orphan" stream therein, finding a mass for the LMC that is consistent with recent estimates. Some highlights of the atlas include the nearby trailing arm of $ω$-Cen, and a nearby very metal-poor stream that was once a satellite of the Sagittarius dwarf galaxy. Finally, we unambiguously detect a hot component around the GD-1 stream, consistent with it having been tidally pre-processed within its own DM subhalo.
△ Less
Submitted 28 November, 2023;
originally announced November 2023.
-
An end-to-end strategy for recovering a free-form potential from a snapshot of stellar coordinates
Authors:
Wassim Tenachi,
Rodrigo Ibata,
Foivos I. Diakogiannis
Abstract:
New large observational surveys such as Gaia are leading us into an era of data abundance, offering unprecedented opportunities to discover new physical laws through the power of machine learning. Here we present an end-to-end strategy for recovering a free-form analytical potential from a mere snapshot of stellar positions and velocities. First we show how auto-differentiation can be used to capt…
▽ More
New large observational surveys such as Gaia are leading us into an era of data abundance, offering unprecedented opportunities to discover new physical laws through the power of machine learning. Here we present an end-to-end strategy for recovering a free-form analytical potential from a mere snapshot of stellar positions and velocities. First we show how auto-differentiation can be used to capture an agnostic map of the gravitational potential and its underlying dark matter distribution in the form of a neural network. However, in the context of physics, neural networks are both a plague and a blessing as they are extremely flexible for modeling physical systems but largely consist in non-interpretable black boxes. Therefore, in addition, we show how a complementary symbolic regression approach can be used to open up this neural network into a physically meaningful expression. We demonstrate our strategy by recovering the potential of a toy isochrone system.
△ Less
Submitted 26 May, 2023;
originally announced May 2023.
-
Deep symbolic regression for physics guided by units constraints: toward the automated discovery of physical laws
Authors:
Wassim Tenachi,
Rodrigo Ibata,
Foivos I. Diakogiannis
Abstract:
Symbolic Regression is the study of algorithms that automate the search for analytic expressions that fit data. While recent advances in deep learning have generated renewed interest in such approaches, the development of symbolic regression methods has not been focused on physics, where we have important additional constraints due to the units associated with our data. Here we present $Φ$-SO, a P…
▽ More
Symbolic Regression is the study of algorithms that automate the search for analytic expressions that fit data. While recent advances in deep learning have generated renewed interest in such approaches, the development of symbolic regression methods has not been focused on physics, where we have important additional constraints due to the units associated with our data. Here we present $Φ$-SO, a Physical Symbolic Optimization framework for recovering analytical symbolic expressions from physics data using deep reinforcement learning techniques by learning units constraints. Our system is built, from the ground up, to propose solutions where the physical units are consistent by construction. This is useful not only in eliminating physically impossible solutions, but because the "grammatical" rules of dimensional analysis restrict enormously the freedom of the equation generator, thus vastly improving performance. The algorithm can be used to fit noiseless data, which can be useful for instance when attempting to derive an analytical property of a physical model, and it can also be used to obtain analytical approximations to noisy data. We test our machinery on a standard benchmark of equations from the Feynman Lectures on Physics and other physics textbooks, achieving state-of-the-art performance in the presence of noise (exceeding 0.1%) and show that it is robust even in the presence of substantial (10%) noise. We showcase its abilities on a panel of examples from astrophysics.
△ Less
Submitted 9 October, 2023; v1 submitted 6 March, 2023;
originally announced March 2023.
-
Typhon: a polar stream from the outer halo raining through the Solar neighborhood
Authors:
Wassim Tenachi,
Pierre-Antoine Oria,
Rodrigo Ibata,
Benoit Famaey,
Zhen Yuan,
Anke Arentsen,
Nicolas Martin,
Akshara Viswanathan
Abstract:
We report on the discovery in the Gaia DR3 astrometric and spectroscopic catalog of a new polar stream that is found as an over-density in action space. This structure is unique as it has an extremely large apocenter distance, reaching beyond 100 kpc, and yet is detected as a coherent moving structure in the Solar neighborhood with a width of $\sim 4$ kpc. A sub-sample of these stars that was fort…
▽ More
We report on the discovery in the Gaia DR3 astrometric and spectroscopic catalog of a new polar stream that is found as an over-density in action space. This structure is unique as it has an extremely large apocenter distance, reaching beyond 100 kpc, and yet is detected as a coherent moving structure in the Solar neighborhood with a width of $\sim 4$ kpc. A sub-sample of these stars that was fortuitously observed by LAMOST has a mean spectroscopic metallicity of $\langle {\rm [Fe/H]}\rangle = -1.60^{+0.15}_{-0.16}$ dex and possesses a resolved metallicity dispersion of $σ({\rm [Fe/H]}) = 0.32^{+0.17}_{-0.06}$ dex. The physical width of the stream, the metallicity dispersion and the vertical action spread indicate that the progenitor was a dwarf galaxy. The existence of such a coherent and highly radial structure at their pericenters in the vicinity of the Sun suggests that many other dwarf galaxy fragments may be lurking in the outer halo.
△ Less
Submitted 29 February, 2024; v1 submitted 21 June, 2022;
originally announced June 2022.
-
Antaeus: a retrograde group of tidal debris in the Milky Way's disk plane
Authors:
Pierre-Antoine Oria,
Wassim Tenachi,
Rodrigo Ibata,
Benoit Famaey,
Zhen Yuan,
Anke Arentsen,
Nicolas Martin,
Akshara Viswanathan
Abstract:
We present the discovery of a wide retrograde moving group in the disk plane of the Milky Way using action-angle coordinates derived from the \textit{Gaia} DR3 catalog. The structure is identified from a sample of its members that are currently almost at the pericenter of their orbit and are passing through the Solar neighborhood. The motions of the stars in this group are highly correlated, indic…
▽ More
We present the discovery of a wide retrograde moving group in the disk plane of the Milky Way using action-angle coordinates derived from the \textit{Gaia} DR3 catalog. The structure is identified from a sample of its members that are currently almost at the pericenter of their orbit and are passing through the Solar neighborhood. The motions of the stars in this group are highly correlated, indicating that the system is probably not phase mixed. With a width of at least 1.5 kpc and with a probable intrinsic spread in metallicity, this structure is most likely the wide remnant of a tidal stream of a disrupted ancient dwarf galaxy (age $\sim 12$ Gyr, $\langle {\rm [Fe/H]} \rangle \sim -1.74$). The structure presents many similarities (e.g. in energy, angular momentum, metallicity, and eccentricity) with the Sequoia merging event. However, it possesses extremely low vertical action $J_z$ which makes it unique even amongst Sequoia dynamical groups. As the low $J_z$ may be attributable to dynamical friction, we speculate that the these stars may be the remnants of the dense core of the Sequoia progenitor.
△ Less
Submitted 21 June, 2022;
originally announced June 2022.
