Showing 1–12 of 12 results for author: Sandhu, J S

Sum-of-Squares & Gaussian Processes I: Certification

Authors: J. S. Sandhu, J. Shi

Abstract: We propose a new SoS hierarchy for a broad class of average-case problems. The new high-entropy steps (HES) SoS hierarchy certifies properties of input instances only over a certain family of well behaved distributions of possible solutions. This allows the hierarchy to circumvent the oft-cited impossibility of providing a low-degree SoS proof of concentration of measure \cite[Lec.~1.2, Marley Par… ▽ More We propose a new SoS hierarchy for a broad class of average-case problems. The new high-entropy steps (HES) SoS hierarchy certifies properties of input instances only over a certain family of well behaved distributions of possible solutions. This allows the hierarchy to circumvent the oft-cited impossibility of providing a low-degree SoS proof of concentration of measure \cite[Lec.~1.2, Marley Paradigm]{barak2016proofs} by vastly reducing the set of solution concepts they need to be proven over. We evaluate the HES SoS hierarchy on a well-studied Gaussian process: the spherical spin-glass, which is a particular ensemble of random low-degree polynomials (LDPs). We prove that, given an instance sampled from the problem distribution, with high probability there exists a HES distribution that achieves value arbitrarily close to the conjectured optimal algorithmic threshold and, at the same time, there are low-degree SoS certificates that no HES distribution can achieve value above a constant factor times this threshold, improving over standard SoS whose certificates are loose by a factor of $n^{\lfloor p/2 - 1\rfloor/2}$ for a degree-$p$ random polynomial \cite{bhattiprolu2017sum, hopkins2017power}. We also give examples of anisotropic modifications to the spherical spin glass problem where there is proof (or strong evidence) that the HES SoS hierarchy does better than the only other known algorithm \cite{subag2021following}. We conjecture that the HES SoS hierarchy can be applied to any LDP approximate Gaussian process to obtain value that is optimal among polynomial-time algorithms. The rounding algorithm is introduced and analyzed in a companion paper. △ Less

Submitted 9 June, 2024; v1 submitted 25 January, 2024; originally announced January 2024.

Comments: 141 pages, 0 figures. Updates to Section-4 & Appendix-E; The updated bound of Theorem-1.1, which is off from ALG by a constant factor, is due to lemma-5.26 giving a suboptimal bound, and was incorrectly stated in the earlier version. The proof is strengthened with a tight bound in the upcoming companion paper which introduces the rounding algorithm

Random Max-CSPs Inherit Algorithmic Hardness from Spin Glasses

Authors: Chris Jones, Kunal Marwaha, Juspreet Singh Sandhu, Jonathan Shi

Abstract: We study random constraint satisfaction problems (CSPs) in the unsatisfiable regime. We relate the structure of near-optimal solutions for any Max-CSP to that for an associated spin glass on the hypercube, using the Guerra-Toninelli interpolation from statistical physics. The noise stability polynomial of the CSP's predicate is, up to a constant, the mixture polynomial of the associated spin glass… ▽ More We study random constraint satisfaction problems (CSPs) in the unsatisfiable regime. We relate the structure of near-optimal solutions for any Max-CSP to that for an associated spin glass on the hypercube, using the Guerra-Toninelli interpolation from statistical physics. The noise stability polynomial of the CSP's predicate is, up to a constant, the mixture polynomial of the associated spin glass. We prove two main consequences: 1) We relate the maximum fraction of constraints that can be satisfied in a random Max-CSP to the ground state energy density of the corresponding spin glass. Since the latter value can be computed with the Parisi formula, we provide numerical values for some popular CSPs. 2) We prove that a Max-CSP possesses generalized versions of the overlap gap property if and only if the same holds for the corresponding spin glass. We transfer results from Huang et al. [arXiv:2110.07847, 2021] to obstruct algorithms with overlap concentration on a large class of Max-CSPs. This immediately includes local classical and local quantum algorithms. △ Less

Submitted 10 January, 2023; v1 submitted 6 October, 2022; originally announced October 2022.

Comments: 41 pages, 1 table

Journal ref: 14th Innovations in Theoretical Computer Science Conference (ITCS 2023); Article No. 77

Efficient Quantum Voting with Information-Theoretic Security

Authors: Emil T. Khabiboulline, Juspreet Singh Sandhu, Marco Ugo Gambetta, Mikhail D. Lukin, Johannes Borregaard

Abstract: Ensuring security and integrity of elections constitutes an important challenge with wide-ranging societal implications. Classically, security guarantees can be ensured based on computational complexity, which may be challenged by quantum computers. We show that the use of quantum networks can enable information-theoretic security for the desirable aspects of a distributed voting scheme in a resou… ▽ More Ensuring security and integrity of elections constitutes an important challenge with wide-ranging societal implications. Classically, security guarantees can be ensured based on computational complexity, which may be challenged by quantum computers. We show that the use of quantum networks can enable information-theoretic security for the desirable aspects of a distributed voting scheme in a resource-efficient manner. In our approach, ballot information is encoded in quantum states that enable an exponential reduction in communication complexity compared to classical communication. In addition, we provide an efficient and secure anonymous queuing protocol. As a result, our scheme only requires modest quantum memories with size scaling logarithmically with the number of voters. This intrinsic efficiency together with certain noise-robustness of our protocol paves the way for its physical implementation in realistic quantum networks. △ Less

Submitted 28 December, 2021; originally announced December 2021.

Comments: 11 + 15 pages, 4 + 1 figures

Limitations of Local Quantum Algorithms on Random Max-k-XOR and Beyond

Authors: Chi-Ning Chou, Peter J. Love, Juspreet Singh Sandhu, Jonathan Shi

Abstract: We introduce a notion of \emph{generic local algorithm} which strictly generalizes existing frameworks of local algorithms such as \emph{factors of i.i.d.} by capturing local \emph{quantum} algorithms such as the Quantum Approximate Optimization Algorithm (QAOA). Motivated by a question of Farhi et al. [arXiv:1910.08187, 2019] we then show limitations of generic local algorithms including QAOA o… ▽ More We introduce a notion of \emph{generic local algorithm} which strictly generalizes existing frameworks of local algorithms such as \emph{factors of i.i.d.} by capturing local \emph{quantum} algorithms such as the Quantum Approximate Optimization Algorithm (QAOA). Motivated by a question of Farhi et al. [arXiv:1910.08187, 2019] we then show limitations of generic local algorithms including QAOA on random instances of constraint satisfaction problems (CSPs). Specifically, we show that any generic local algorithm whose assignment to a vertex depends only on a local neighborhood with $o(n)$ other vertices (such as the QAOA at depth less than $ε\log(n)$) cannot arbitrarily-well approximate boolean CSPs if the problem satisfies a geometric property from statistical physics called the coupled overlap-gap property (OGP) [Chen et al., Annals of Probability, 47(3), 2019]. We show that the random MAX-k-XOR problem has this property when $k\geq4$ is even by extending the corresponding result for diluted $k$-spin glasses. Our concentration lemmas confirm a conjecture of Brandao et al. [arXiv:1812.04170, 2018] asserting that the landscape independence of QAOA extends to logarithmic depth -- in other words, for every fixed choice of QAOA angle parameters, the algorithm at logarithmic depth performs almost equally well on almost all instances. One of these concentration lemmas is a strengthening of McDiarmid's inequality, applicable when the random variables have a highly biased distribution, and may be of independent interest. △ Less

Submitted 21 February, 2022; v1 submitted 12 August, 2021; originally announced August 2021.

Comments: 59 pages, 2 figures. Third version has an updated abstract, an introduction with a more complete literature review, and open questions, as well as a fix to some typos in Section-5 and Section-6. The second version was updated with a new proof that demonstrated a coupled OGP for Random Max-k-XOR (signed)

A General Framework for Analyzing Stochastic Dynamics in Learning Algorithms

Authors: Chi-Ning Chou, Juspreet Singh Sandhu, Mien Brabeeba Wang, Tiancheng Yu

Abstract: One of the challenges in analyzing learning algorithms is the circular entanglement between the objective value and the stochastic noise. This is also known as the "chicken and egg" phenomenon and traditionally, there is no principled way to tackle this issue. People solve the problem by utilizing the special structure of the dynamic, and hence the analysis would be difficult to generalize. In t… ▽ More One of the challenges in analyzing learning algorithms is the circular entanglement between the objective value and the stochastic noise. This is also known as the "chicken and egg" phenomenon and traditionally, there is no principled way to tackle this issue. People solve the problem by utilizing the special structure of the dynamic, and hence the analysis would be difficult to generalize. In this work, we present a streamlined three-step recipe to tackle the "chicken and egg" problem and give a general framework for analyzing stochastic dynamics in learning algorithms. Our framework composes standard techniques from probability theory, such as stop** time and martingale concentration. We demonstrate the power and flexibility of our framework by giving a unifying analysis for three very different learning problems with the last iterate and the strong uniform high probability convergence guarantee. The problems are stochastic gradient descent for strongly convex functions, streaming principal component analysis, and linear bandit with stochastic gradient descent updates. We either improve or match the state-of-the-art bounds on all three dynamics. △ Less

Submitted 28 September, 2022; v1 submitted 10 June, 2020; originally announced June 2020.

Accurate Ground-based Near-Earth-Asteroid Astrometry using Synthetic Tracking

Authors: Chengxing Zhai, Michael Shao, Navtej S. Saini, Jagmit S. Sandhu, Phil Choi, William M. Owen, Thomas A. Werne, Todd A. Ely, Joseph Lazio, Tomas J. Martin-Mur, Robert A. Preston, Slava G. Turyshev, Adam W. Mitchell, Kutay Nazli, Isaac Cui, Rachel M. Mochama

Abstract: Accurate astrometry is crucial for determining orbits of near-Earth-asteroids (NEAs) and therefore better tracking them. This paper reports on a demonstration of 10 milliarcsecond-level astrometric precision on a dozen NEAs using the Pomona College 40 inch telescope, at the JPL's Table Mountain Facility. We used the technique of synthetic tracking, in which many short exposure (1 second) images ar… ▽ More Accurate astrometry is crucial for determining orbits of near-Earth-asteroids (NEAs) and therefore better tracking them. This paper reports on a demonstration of 10 milliarcsecond-level astrometric precision on a dozen NEAs using the Pomona College 40 inch telescope, at the JPL's Table Mountain Facility. We used the technique of synthetic tracking, in which many short exposure (1 second) images are acquired and then combined in post-processing to track both target asteroid and reference stars across the field of view. This technique avoids the trailing loss and keeps the jitter effects from atmosphere and telescope pointing common between the asteroid and reference stars, resulting in higher astrometric precision than the 100 mas level astrometry from traditional approach of using long exposure images. Treating our synthetic tracking of near-Earth asteroids as a proxy for observations of future spacecraft while they are downlinking data via their high rate optical communication laser beams, our approach shows precision plane-of-sky measurements can be obtained by the optical ground terminals for navigation. We also discuss how future data releases from the Gaia mission can improve our results. △ Less

Submitted 3 May, 2018; originally announced May 2018.

Comments: 26 pages, 10 figures, submitted to the Astronomical Journal

Journal ref: Astron. J. 156(2): 65 (2018)

Parallax of PSR J1744-1134 and the Local Interstellar Medium

Authors: M. Toscano, M. C. Britton, R. N. Manchester, M. Bailes, J. S. Sandhu, S. R. Kulkarni, S. B. Anderson

Abstract: We present the annual trigonometric parallax of PSR J1744-1134 derived from an analysis of pulse times of arrival. The measured parallax, pi = 2.8+/-0.3 mas ranks among the most precisely determined distances to any pulsar. The parallax distance of 357+/-39 pc is over twice that derived from the dispersion measure using the Taylor & Cordes model for the Galactic electron distribution. The mean e… ▽ More We present the annual trigonometric parallax of PSR J1744-1134 derived from an analysis of pulse times of arrival. The measured parallax, pi = 2.8+/-0.3 mas ranks among the most precisely determined distances to any pulsar. The parallax distance of 357+/-39 pc is over twice that derived from the dispersion measure using the Taylor & Cordes model for the Galactic electron distribution. The mean electron density in the path to the pulsar, n_e = (0.0088 +/- 0.0009) cm^{-3}, is the lowest for any disk pulsar. We have compared the n_e for PSR J1744-1134 with those for another 11 nearby pulsars with independent distance estimates. We conclude that there is a striking asymmetry in the distribution of electrons in the local interstellar medium. The electron column densities for pulsars in the third Galactic quadrant are found to be systematically higher than for those in the first. The former correlate with the position of the well known local HI cavity in quadrant three. The excess electrons within the cavity may be in the form of HII clouds marking a region of interaction between the local hot bubble and a nearby superbubble. △ Less

Submitted 28 July, 1999; v1 submitted 24 June, 1999; originally announced June 1999.

Comments: revised version accepted for publication in ApJ Letters; reanalysis of uncertainty in parallax measure and changes to fig 1

Report number: SASG-3

Millisecond Pulsar Velocities

Authors: M. Toscano, J. S. Sandhu, M. Bailes, R. N. Manchester, M. C. Britton, S. R. Kulkarni, S. B. Anderson, B. W. Stappers

Abstract: We present improved timing parameters for 13 millisecond pulsars (MSPs) including 9 new proper motion measurements. These new proper motions bring to 23 the number of MSPs with measured transverse velocities. In light of these new results we present and compare the kinematic properties of MSPs with those of ordinary pulsars. The mean transverse velocity of MSPs was found to be 85+/-13 km/s; a va… ▽ More We present improved timing parameters for 13 millisecond pulsars (MSPs) including 9 new proper motion measurements. These new proper motions bring to 23 the number of MSPs with measured transverse velocities. In light of these new results we present and compare the kinematic properties of MSPs with those of ordinary pulsars. The mean transverse velocity of MSPs was found to be 85+/-13 km/s; a value consistent with most models for the origin and evolution of MSPs and approximately a factor of four lower than that of ordinary pulsars. We also find that, in contrast to young ordinary pulsars, the vast majority of which are moving away from the Galactic plane, almost half of the MSPs are moving towards the plane. This near isotropy would be expected of a population that has reached dynamic equilibrium. Accurate measurements of MSP velocities have allowed us to correct their measured spin-down rates for Doppler acceleration effects, and thereby derive their intrinsic magnetic field strengths and characteristic ages. We find that close to half of our sample of MSPs have a characteristic age comparable to or greater than the age of the Galaxy. △ Less

Submitted 24 November, 1998; originally announced November 1998.

Comments: 10 pages LaTeX including 2 LaTeX tables and 3 postscript figures; submitted to MNRAS

Report number: SASG-2

Spectra of Southern Pulsars

Authors: M. Toscano, M. Bailes, R. N. Manchester, J. S. Sandhu

Abstract: We compare the spectral properties of the millisecond and slow pulsars detected in the Parkes 70 cm survey. The mean spectral index for the millisecond pulsars (MSPs) is -1.9 +/- 0.1 whereas the mean spectral index for the slow pulsars is a surprisingly steep -1.72 +/- 0.04. A Kolmogorov-Smirnov test indicates that there is only a 72% probability that the two distributions differ. As a class, MS… ▽ More We compare the spectral properties of the millisecond and slow pulsars detected in the Parkes 70 cm survey. The mean spectral index for the millisecond pulsars (MSPs) is -1.9 +/- 0.1 whereas the mean spectral index for the slow pulsars is a surprisingly steep -1.72 +/- 0.04. A Kolmogorov-Smirnov test indicates that there is only a 72% probability that the two distributions differ. As a class, MSPs are therefore only fractionally steeper-spectrum objects than slow pulsars, as recent literature would suggest. We then model the expected distribution of millisecond pulsars in the Galaxy and find that high-frequency surveys, with sensitivities similar to the current Parkes multibeam survey, are likely to detect MSPs in large numbers. The observed distribution of MSPs will be much less isotropic than that resulting from low-frequency surveys, with 50% of detectable MSPs residing within 11 degrees of the Galactic plane in an all-sky survey. △ Less

Submitted 18 May, 1998; originally announced May 1998.

Comments: 17 pages LaTeX, 4 postscript figures; accepted for publication in Astrophysical Journal

Report number: SASG-1

The Orbital Evolution and Proper Motion of PSR J2051-0827

Authors: B. W. Stappers, M. Bailes, R. N. Manchester, J. S. Sandhu, M. Toscano

Abstract: We have carried out high-precision timing observations of the eclipsing binary PSR J2051-0827 in the 3.3 years since its discovery. These data indicate that the orbital period is decreasing at a rate of dPb/dt = (-11+-1)X10^-12. If secular, this orbital period derivative implies a decay time for the orbit of only 25 Myr which is much shorter than the expected timescale for ablation of the compan… ▽ More We have carried out high-precision timing observations of the eclipsing binary PSR J2051-0827 in the 3.3 years since its discovery. These data indicate that the orbital period is decreasing at a rate of dPb/dt = (-11+-1)X10^-12. If secular, this orbital period derivative implies a decay time for the orbit of only 25 Myr which is much shorter than the expected timescale for ablation of the companion. We have also measured the proper motion of the pulsar to be 5+-3 mas/yr. Assuming the pulsar is at the dispersion-measure distance this implies a very slow transverse velocity vt=(30+-20) km/s. This combination of low velocity and short orbital period argue against formation of the system in the standard manner and we discuss the implications for its evolutionary history. △ Less

Submitted 10 April, 1998; originally announced April 1998.

Comments: 7 pages, 1 figure, accepted for publication in Astrophysical Journal Letters

Discovery of the Young, Energetic Radio Pulsar PSR J1105-6107

Authors: V. M. Kaspi, M. Bailes, R. N. Manchester, B. W. Stappers, J. S. Sandhu, J. Navarro, N. D'Amico

Abstract: We report the discovery and follow-up timing observations of the 63 ms radio pulsar, PSR J1105-6107. The pulsar is young, having a characteristic age of only 63 kyr and, from its dispersion measure, is estimated to be at a distance of ~7 kpc from the Sun. We consider its possible association with the nearby supernova remnant G290.1-0.8 (MSH 11-61A); an association requires that the pulsar's prop… ▽ More We report the discovery and follow-up timing observations of the 63 ms radio pulsar, PSR J1105-6107. The pulsar is young, having a characteristic age of only 63 kyr and, from its dispersion measure, is estimated to be at a distance of ~7 kpc from the Sun. We consider its possible association with the nearby supernova remnant G290.1-0.8 (MSH 11-61A); an association requires that the pulsar's proper motion be ~22 mas/yr (corresponding to ~650 km/s for a distance of 7 kpc) directed away from the remnant center, assuming that the characteristic age is the true age. The pulsar's spin-down luminosity, 2.5 x 10^{36} erg/s, is in the top 1% of all known pulsar spin-down luminosities. Given its estimated distance, PSR J1105-6107 is therefore likely to be observable at high energies. Indeed, it is coincident with the known CGRO/EGRET source 2EG J1103-6106; we consider the possible association and conclude that it is likely. △ Less

Submitted 18 March, 1997; originally announced March 1997.

Comments: 18 pages (LaTex), 4 figures, to be published in The Astrophysical Journal, preprint also available at http://space.mit.edu/~vicky/papers.html

Timing Measurements and Their Implications for Four Binary Millisecond Pulsars

Authors: J. F. Bell, M. Bailes, R. N. Manchester, A. G. Lyne, F. Camilo, J. S. Sandhu

Abstract: We present timing observations of four millisecond pulsars, using data obtained over three years at the ATNF Parkes and NRAL Jodrell Bank radio telescopes. Astrometric, spin, and binary parameters are updated, and substantially improved for three pulsars, PSRs J0613-0200, J1045-4509 and J1643-1224. We have measured the time variation of the projected semi-major axis of the PSR J0437-4715 orbit d… ▽ More We present timing observations of four millisecond pulsars, using data obtained over three years at the ATNF Parkes and NRAL Jodrell Bank radio telescopes. Astrometric, spin, and binary parameters are updated, and substantially improved for three pulsars, PSRs J0613-0200, J1045-4509 and J1643-1224. We have measured the time variation of the projected semi-major axis of the PSR J0437-4715 orbit due to its proper motion, and use it to constrain the inclination of the orbit and the mass of the companion. Some evidence is found for changes in the dispersion measures of PSRs J1045-4509 and J1643-1224. Limits are placed on the existence of planetary mass companions, ruling out companions with masses and orbits similar to the terrestrial planets of the solar system for eight pulsars. △ Less

Submitted 11 November, 1996; originally announced November 1996.

Comments: 8 pages, Latex, etc, requires mn.sty epsf.sty, Accepted for publication in MNRAS, 6 postscript figures included