-
Bulk Johnson-Lindenstrauss Lemmas
Authors:
Michael P. Casey
Abstract:
For a set $X$ of $N$ points in $\mathbb{R}^D$, the Johnson-Lindenstrauss lemma provides random linear maps that approximately preserve all pairwise distances in $X$ -- up to multiplicative error $(1\pm ε)$ with high probability -- using a target dimension of $O(ε^{-2}\log(N))$. Certain known point sets actually require a target dimension this large -- any smaller dimension forces at least one dist…
▽ More
For a set $X$ of $N$ points in $\mathbb{R}^D$, the Johnson-Lindenstrauss lemma provides random linear maps that approximately preserve all pairwise distances in $X$ -- up to multiplicative error $(1\pm ε)$ with high probability -- using a target dimension of $O(ε^{-2}\log(N))$. Certain known point sets actually require a target dimension this large -- any smaller dimension forces at least one distance to be stretched or compressed too much. What happens to the remaining distances? If we only allow a fraction $η$ of the distances to be distorted beyond tolerance $(1\pm ε)$, we show a target dimension of $O(ε^{-2}\log(4e/η)\log(N)/R)$ is sufficient for the remaining distances. With the stable rank of a matrix $A$ as $\lVert{A\rVert}_F^2/\lVert{A\rVert}^2$, the parameter $R$ is the minimal stable rank over certain $\log(N)$ sized subsets of $X-X$ or their unit normalized versions, involving each point of $X$ exactly once. The linear maps may be taken as random matrices with i.i.d. zero-mean unit-variance sub-gaussian entries. When the data is sampled i.i.d. as a given random vector $ξ$, refined statements are provided; the most improvement happens when $ξ$ or the unit normalized $\widehat{ξ-ξ'}$ is isotropic, with $ξ'$ an independent copy of $ξ$, and includes the case of i.i.d. coordinates.
△ Less
Submitted 15 July, 2023;
originally announced July 2023.
-
Linear Dimension Reduction Approximately Preserving a Function of the 1-Norm
Authors:
Michael P. Casey
Abstract:
For any finite point set in $D$-dimensional space equipped with the 1-norm, we present random linear embeddings to $k$-dimensional space, with a new metric, having the following properties. For any pair of points from the point set that are not too close, the distance between their images is a strictly concave increasing function of their original distance, up to multiplicative error. The target d…
▽ More
For any finite point set in $D$-dimensional space equipped with the 1-norm, we present random linear embeddings to $k$-dimensional space, with a new metric, having the following properties. For any pair of points from the point set that are not too close, the distance between their images is a strictly concave increasing function of their original distance, up to multiplicative error. The target dimension $k$ need only be quadratic in the logarithm of the size of the point set to ensure the result holds with high probability. The linear embeddings are random matrices composed of standard Cauchy random variables, and the proofs rely on Chernoff bounds for sums of iid random variables. The new metric is translation invariant, but is not induced by a norm.
△ Less
Submitted 5 November, 2020; v1 submitted 8 June, 2019;
originally announced June 2019.
-
MOST observations of the Herbig Ae δ-Scuti star HD 34282
Authors:
M. P. Casey,
K. Zwintz,
D. B. Guenther,
W. W. Weiss,
P. J. Amado,
D. Díaz-Fraile,
E. Rodriguez,
R. Kuschnig,
J. M. Matthews,
A. F. J. Moffat,
J. F. Rowe,
S. M. Rucinski,
D. Sasselov
Abstract:
MOST observations and model analysis of the Herbig Ae star HD 34282 (V1366 Ori) reveal δ-Scuti pulsations. 22 frequencies are observed, 10 of which confirm those previously identified by Amado et al. (2006), and 12 of which are newly discovered in this work. We show that the weighted-average frequency in each group fits the radial p-mode frequencies of viable models. We argue that the observed pul…
▽ More
MOST observations and model analysis of the Herbig Ae star HD 34282 (V1366 Ori) reveal δ-Scuti pulsations. 22 frequencies are observed, 10 of which confirm those previously identified by Amado et al. (2006), and 12 of which are newly discovered in this work. We show that the weighted-average frequency in each group fits the radial p-mode frequencies of viable models. We argue that the observed pulsation spectrum extends just to the edge to the acoustic cut-off frequency and show that this also is consistent with our best-fitting models.
△ Less
Submitted 18 October, 2012; v1 submitted 17 October, 2012;
originally announced October 2012.
-
Intrinsic symmetry groups of links with 8 and fewer crossings
Authors:
Michael Berglund,
Jason Cantarella,
Meredith Perrie Casey,
Ellie Dannenberg,
Whitney George,
Aja Johnson,
Amelia Kelly,
Al LaPointe,
Matt Mastin,
Jason Parsley,
Jacob Rooney,
Rachel Whitaker
Abstract:
We present an elementary derivation of the "intrinsic" symmetry groups for knots and links of 8 or fewer crossings. The standard symmetry group for a link is the map** class group $\MCG(S^3,L)$ or $\Sym(L)$ of the pair $(S^3,L)$. Elements in this symmetry group can (and often do) fix the link and act nontrivially only on its complement. We ignore such elements and focus on the "intrinsic" symmet…
▽ More
We present an elementary derivation of the "intrinsic" symmetry groups for knots and links of 8 or fewer crossings. The standard symmetry group for a link is the map** class group $\MCG(S^3,L)$ or $\Sym(L)$ of the pair $(S^3,L)$. Elements in this symmetry group can (and often do) fix the link and act nontrivially only on its complement. We ignore such elements and focus on the "intrinsic" symmetry group of a link, defined to be the image $Σ(L)$ of the natural homomorphism $\MCG(S^3,L) \rightarrow \MCG(S^3) \cross \MCG(L)$. This different symmetry group, first defined by Whitten in 1969, records directly whether $L$ is isotopic to a link $L'$ obtained from $L$ by permuting components or reversing orientations.
For hyperbolic links both $\Sym(L)$ and $Σ(L)$ can be obtained using the output of \texttt{SnapPea}, but this proof does not give any hints about how to actually construct isotopies realizing $Σ(L)$. We show that standard invariants are enough to rule out all the isotopies outside $Σ(L)$ for all links except $7^2_6$, $8^2_{13}$ and $8^3_5$ where an additional construction is needed to use the Jones polynomial to rule out "component exchange" symmetries. On the other hand, we present explicit isotopies starting with the positions in Cerf's table of oriented links which generate $Σ(L)$ for each link in our table. Our approach gives a constructive proof of the $Σ(L)$ groups.
△ Less
Submitted 4 September, 2011; v1 submitted 15 October, 2010;
originally announced October 2010.
-
MOST photometry of the enigmatic PMS pulsator HD 142666
Authors:
K. Zwintz,
T. Kallinger,
D. B. Guenther,
M. Gruberbauer,
D. Huber,
J. Rowe,
R. Kuschnig,
W. W. Weiss,
J. M. Matthews,
A. F. J. Moffat,
S. M. Rucinski,
D. Sasselov,
G. A. H. Walker,
M. P. Casey
Abstract:
We present precise photometry of the pulsating Herbig Ae star HD 142666 obtained in two consecutive years with the MOST (Microvariability & Oscilations of STars) satellite.
Previously, only a single pulsation period was known for HD 142666. The MOST photometry reveals that HD 142666 is multi-periodic. However, the unique identification of pulsation frequencies is complicated by the presence of…
▽ More
We present precise photometry of the pulsating Herbig Ae star HD 142666 obtained in two consecutive years with the MOST (Microvariability & Oscilations of STars) satellite.
Previously, only a single pulsation period was known for HD 142666. The MOST photometry reveals that HD 142666 is multi-periodic. However, the unique identification of pulsation frequencies is complicated by the presence of irregular variability caused by the star's circumstellar dust disk. The two light curves obtained with MOST in 2006 and 2007 provided data of unprecedented quality to study the pulsations in HD 142666 and also to monitor the circumstellar variability.
We attribute 12 frequencies to pulsation. Model fits to the three frequencies with the highest amplitudes lie well outside the uncertainty box for the star's position in the HR diagram based on published values.
The models suggest that either (1) the published estimate of the luminosity of HD 142666, based on a relation between circumstellar disk radius and stellar luminosity, is too high and/or (2) additional physics such as mass accretion may be needed in our models to accurately fit both the observed frequencies and HD 142666's position in the HR diagram.
△ Less
Submitted 10 December, 2008;
originally announced December 2008.
-
The DDO IVC Distance Project
Authors:
Michael D. Gladders,
Tracy E. Clarke,
Christopher R. Burns,
Allen Attard,
Michael P. Casey,
Devon Hamilton,
Gabriela Mallen-Ornelas,
Jennifer L. Karr,
Sara M. Poirier,
Marcin Sawicki,
Felipe Barrientos,
Wayne Barkhouse,
Mark Brodwin,
Jason Clark,
Rosemary McNaughton,
Marcelo Ruetalo-Pacheco,
Stefan W. Mochnacki
Abstract:
We present the first set of distance limits from the David Dunlap Observatory Intermediate Velocity Cloud (DDO IVC) distance project. Such distance measures are crucial to understanding the origins and dynamics of IVCs, as the distances set most of the basic physical parameters for the clouds. Currently there are very few IVCs with reliably known distances. This paper describes in some detail th…
▽ More
We present the first set of distance limits from the David Dunlap Observatory Intermediate Velocity Cloud (DDO IVC) distance project. Such distance measures are crucial to understanding the origins and dynamics of IVCs, as the distances set most of the basic physical parameters for the clouds. Currently there are very few IVCs with reliably known distances. This paper describes in some detail the basic techniques used to measure distances, with particular emphasis on the the analysis of interstellar absorption line data, which forms the basis of our distance determinations. As an example, we provide a detailed description of our distance determination for the Draco Cloud. Preliminary distance limits for a total of eleven clouds are provided.
△ Less
Submitted 18 November, 1998;
originally announced November 1998.
-
The Distance to the Draco Cloud
Authors:
Michael D. Gladders,
T. E. Clarke,
Christopher R. Burns,
A. Attard,
M. P. Casey,
Devon Hamilton,
Gabriela Mallén-Ornelas,
J. L. Karr,
Sara M. Poirier,
Marcin Sawicki,
L. Felipe Barrientos,
Stefan W. Mochnacki
Abstract:
The understanding of the nature of intermediate and high velocity gas in the Milky Way is hampered by a paucity of distance estimates to individual clouds. A project has been started at the David Dunlap Observatory to address this lack of distance measures by observing early-type stars along the line of sight towards these clouds and searching for sodium doublet absorption at the clouds' systemi…
▽ More
The understanding of the nature of intermediate and high velocity gas in the Milky Way is hampered by a paucity of distance estimates to individual clouds. A project has been started at the David Dunlap Observatory to address this lack of distance measures by observing early-type stars along the line of sight towards these clouds and searching for sodium doublet absorption at the clouds' systemic velocities. Distances to foreground stars (no absorption) and background stars (with absorption) are estimated from spectroscopic parallax, and thus the distance to the bracketed cloud is estimated. In this Letter, we present the first result from this ongoing project, a measurement of the distance to the Draco Cloud, which is the most studied of the intermediate velocity clouds. The result presented here is the first distance bracket which tightly constrains the position of the Draco Cloud. We briefly describe our target selection and observing methodology, and then demonstrate absorption at the velocity of the Draco Cloud for one star (TYC~4194~2188), and a lack of absorption for several other stars. We derive a distance bracket to the Draco Cloud of 463$^{+192}_{-136}$ to 618$^{+243}_{-174}$~pc.
△ Less
Submitted 14 September, 1998;
originally announced September 1998.
