Papers made digestable
Our architecture simplifies the obstacle-perception
problem to that of place-dependent change detection. While we use the method with VT&R, it
can be generalized to suit arbitrary path-following applications.
Visual Teach and Repeat 3 (VT&R3), a generalization of stereo VT&R, achieves
long-term autonomous path-following using topometric mapping and localization
from a single rich sensor stream. In this paper, we improve the capabilities of
a LiDAR implementation of VT&R3 to reliably detect and avoid obstacles in
changing environments. Our architecture simplifies the obstacle-perception
problem to that of place-dependent change detection. We then extend the
behaviour of generic sample-based motion planners to better suit the
teach-and-repeat problem structure by introducing a new edge-cost metric paired
with a curvilinear planning space. The resulting planner generates naturally
smooth paths that avoid local obstacles while minimizing lateral path deviation
to best exploit prior terrain knowledge. While we use the method with VT&R, it
can be generalized to suit arbitrary path-following applications. Experimental
results from online run-time analysis, unit testing, and qualitative
experiments on a differential drive robot show the promise of the technique for
reliable long-term autonomous operation in complex unstructured environments.
Authors: Jordy Sehn, Yuchen Wu, Timothy D. Barfoot.
The statistical and design considerations that pertain to
dose optimization are discussed. The sample size savings range from 16.6% to 27.3%,
depending on the design and scenario, with a mean savings of 22.1%.
The traditional more-is-better dose selection paradigm, developed based on
cytotoxic chemotherapeutics, is often problematic When applied to the
development of novel molecularly targeted agents (e.g., kinase inhibitors,
monoclonal antibodies, and antibody-drug conjugates). The US Food and Drug
Administration (FDA) initiated Project Optimus to reform the dose optimization
and dose selection paradigm in oncology drug development and call for more
attention to benefit-risk consideration.
We systematically investigated the operating characteristics of the seamless
phase 2-3 design as a strategy for dose optimization, where in stage 1
(corresponding to phase 2) patients are randomized to multiple doses, with or
without a control; and in stage 2 (corresponding to phase 3) the efficacy of
the selected optimal dose is evaluated with a randomized concurrent control or
historical control. Depending on whether the concurrent control is included and
the type of endpoints used in stages 1 and 2, we describe four types of
seamless phase 2-3 dose-optimization designs, which are suitable for different
clinical settings. The statistical and design considerations that pertain to
dose optimization are discussed. Simulation shows that dose optimization phase
2-3 designs are able to control the familywise type I error rates and yield
appropriate statistical power with substantially smaller sample size than the
conventional approach. The sample size savings range from 16.6% to 27.3%,
depending on the design and scenario, with a mean savings of 22.1%. Due to the
interim dose selection, the phase 2-3 dose-optimization design is logistically
and operationally more challenging, and should be carefully planned and
implemented to ensure trial integrity.
Authors: Liyun Jiang, Ying Yuan.
We significantly improve performance using properties of the posterior
in our active learning scheme and for the definition of the GP prior. In
particular we account for the expected dynamical range of the posterior in
different dimensionalities. We test our model against a number of synthetic and
cosmological examples.
We present the GPry algorithm for fast Bayesian inference of general
(non-Gaussian) posteriors with a moderate number of parameters. GPry does not
need any pre-training, special hardware such as GPUs, and is intended as a
drop-in replacement for traditional Monte Carlo methods for Bayesian inference.
Our algorithm is based on generating a Gaussian Process surrogate model of the
log-posterior, aided by a Support Vector Machine classifier that excludes
extreme or non-finite values. An active learning scheme allows us to reduce the
number of required posterior evaluations by two orders of magnitude compared to
traditional Monte Carlo inference. Our algorithm allows for parallel
evaluations of the posterior at optimal locations, further reducing wall-clock
times. We significantly improve performance using properties of the posterior
in our active learning scheme and for the definition of the GP prior. In
particular we account for the expected dynamical range of the posterior in
different dimensionalities. We test our model against a number of synthetic and
cosmological examples. GPry outperforms traditional Monte Carlo methods when
the evaluation time of the likelihood (or the calculation of theoretical
observables) is of the order of seconds; for evaluation times of over a minute
it can perform inference in days that would take months using traditional
methods. GPry is distributed as an open source Python package (pip install
gpry) and can also be found at https://github.com/jonaselgammal/GPry.
Authors: Jonas El Gammal, Nils Schöneberg, Jesús Torrado, Christian Fidler.
We consider the fundamental scheduling problem of minimizing the sum of
weighted completion times on a single machine in the non-clairvoyant setting. However, to the best of our knowledge, this concept has never been considered
for the total completion time objective in the non-clairvoyant model. This implies
a performance guarantee of $(1+3\sqrt{3})\approx 6.197$ for the deterministic
algorithm and of $\approx 3.032$ for the randomized version.
We consider the fundamental scheduling problem of minimizing the sum of
weighted completion times on a single machine in the non-clairvoyant setting.
While no non-preemptive algorithm is constant competitive, Motwani, Phillips,
and Torng (SODA '93) proved that the simple preemptive round robin procedure is
$2$-competitive and that no better competitive ratio is possible, initiating a
long line of research focused on preemptive algorithms for generalized variants
of the problem. As an alternative model, Shmoys, Wein, and Williamson (FOCS
'91) introduced kill-and-restart schedules, where running jobs may be killed
and restarted from scratch later, and analyzed then for the makespan objective.
However, to the best of our knowledge, this concept has never been considered
for the total completion time objective in the non-clairvoyant model.
We contribute to both models: First we give for any $b > 1$ a tight analysis
for the natural $b$-scaling kill-and-restart strategy for scheduling jobs
without release dates, as well as for a randomized variant of it. This implies
a performance guarantee of $(1+3\sqrt{3})\approx 6.197$ for the deterministic
algorithm and of $\approx 3.032$ for the randomized version. Second, we show
that the preemptive Weighted Shortest Elapsed Time First (WSETF) rule is
$2$-competitive for jobs released in an online fashion over time, matching the
lower bound by Motwani et al. Using this result as well as the competitiveness
of round robin for multiple machines, we prove performance guarantees of
adaptions of the $b$-scaling algorithm to online release dates and unweighted
jobs on identical parallel machines.
Authors: Sven Jäger, Guillaume Sagnol, Daniel Schmidt genannt Waldschmidt, Philipp Warode.
Frozen pretrained models have become a viable alternative to the
pretraining-then-finetuning paradigm for transfer learning. With this work, we hope to
bring greater attention to this promising path of freezing pretrained image
models.
Frozen pretrained models have become a viable alternative to the
pretraining-then-finetuning paradigm for transfer learning. However, with
frozen models there are relatively few parameters available for adapting to
downstream tasks, which is problematic in computer vision where tasks vary
significantly in input/output format and the type of information that is of
value. In this paper, we present a study of frozen pretrained models when
applied to diverse and representative computer vision tasks, including object
detection, semantic segmentation and video action recognition. From this
empirical analysis, our work answers the questions of what pretraining task
fits best with this frozen setting, how to make the frozen setting more
flexible to various downstream tasks, and the effect of larger model sizes. We
additionally examine the upper bound of performance using a giant frozen
pretrained model with 3 billion parameters (SwinV2-G) and find that it reaches
competitive performance on a varied set of major benchmarks with only one
shared frozen base network: 60.0 box mAP and 52.2 mask mAP on COCO object
detection test-dev, 57.6 val mIoU on ADE20K semantic segmentation, and 81.7
top-1 accuracy on Kinetics-400 action recognition. With this work, we hope to
bring greater attention to this promising path of freezing pretrained image
models.
Authors: Yutong Lin, Ze Liu, Zheng Zhang, Han Hu, Nanning Zheng, Stephen Lin, Yue Cao.
The Szegedy quantum walk is a discrete time quantum walk model which defines
a quantum analogue of any Markov chain. We prove a formula for our mixing matrix in terms of the
spectral decomposition of the Markov chain and show a relationship with the
mixing matrix of a continuous quantum walk on the chain. In particular, we
prove that average uniform mixing in the continuous walk implies average
uniform mixing in the Szegedy walk.
The Szegedy quantum walk is a discrete time quantum walk model which defines
a quantum analogue of any Markov chain. The long-term behavior of the quantum
walk can be encoded in a matrix called the average mixing matrix, whose columns
give the limiting probability distribution of the walk given an initial state.
We define a version of the average mixing matrix of the Szegedy quantum walk
which allows us to more readily compare the limiting behavior to that of the
chain it quantizes. We prove a formula for our mixing matrix in terms of the
spectral decomposition of the Markov chain and show a relationship with the
mixing matrix of a continuous quantum walk on the chain. In particular, we
prove that average uniform mixing in the continuous walk implies average
uniform mixing in the Szegedy walk. We conclude by giving examples of Markov
chains of arbitrarily large size which admit average uniform mixing in both the
continuous and Szegedy quantum walk.
Authors: Julien Sorci.
\sim \! 0.2-0.4$ is consistent with each system we model.
Common envelope (CE) evolution, which is crucial in creating short period
binaries and associated astrophysical events, can be constrained by reverse
modeling of such binaries' formation histories. Through analysis of a sample of
well-constrained white dwarf (WD) binaries with low-mass primaries (7 eclipsing
double WDs, 2 non-eclipsing double WDs, 1 WD-brown dwarf), we estimate the CE
energy efficiency $\alpha_{\rm{CE}}$ needed to unbind the hydrogen envelope. We
use grids of He- and CO-core WD models to determine the masses and cooling ages
that match each primary WD's radius and temperature. Assuming gravitational
wave-driven orbital decay, we then calculate the associated ranges in post-CE
orbital period. By mapping WD models to a grid of red giant progenitor stars,
we determine the total envelope binding energies and possible orbital periods
at the point CE evolution is initiated, thereby constraining $\alpha_{\rm CE}$.
Assuming He-core WDs with progenitors of 0.9 - 2.0 $M_\odot$, we find
$\alpha_{\rm CE} \! \sim \! 0.2-0.4$ is consistent with each system we model.
Significantly higher values of $\alpha_{\rm{CE}}$ are required for higher mass
progenitors and for CO-core WDs, so these scenarios are deemed unlikely. Our
values are mostly consistent with previous studies of post-CE WD binaries, and
they suggest a nearly constant and low envelope ejection efficiency for CE
events that produce He-core WDs.
Authors: Peter Scherbak, Jim Fuller.
Online adaptive proton therapy (APT) is an ideal solution theoretically,
which however is challenging to proton clinics. Although multiple groups have
been endeavoring to develop online APT technology, there is a concern in the
radiotherapy community about the necessity of online APT because of its unknown
impact on treatment outcomes. The cumulative dose of simulated online APT courses
was compared to actual offline APT courses and the initially designed treatment
plan dose. Future studies are needed to help identify the patients
with large potential benefits prior to treatment to conserve scarce clinical
resources.
Online adaptive proton therapy (APT) is an ideal solution theoretically,
which however is challenging to proton clinics. Although multiple groups have
been endeavoring to develop online APT technology, there is a concern in the
radiotherapy community about the necessity of online APT because of its unknown
impact on treatment outcomes. Hence, we have performed a retrospective study to
investigate the potential clinical effects of online APT for HN cancer patients
in relative to the current offline APT via simulations. To mimic an online APT
treatment course, we have recalculated and evaluated the actual dose of the
current treatment plan on patient actual treatment anatomy captured by cone
beam CT for each fraction. The cumulative dose of simulated online APT courses
was compared to actual offline APT courses and the initially designed treatment
plan dose. For patients 1 and 2, the simulated online ART course maintained a
relatively higher CTV dose coverages than the offline ART course, particularly
for CTV-Low, which led to an improvement of 2.66% and 4.52% in TCP of CTV-Low.
For patients 3 and 4, with clinically comparable CTV dose coverages, the
simulated online ART course achieved better OAR sparing than the offline ART
course. The mean doses of right parotid and oral cavity were decreased from
29.52 Gy relative biological effectiveness (RBE) and 41.89 Gy RBE to 22.16 Gy
RBE and 34.61 Gy RBE for patient 3, leading to a reduce of 1.67% and 3.40% in
NTCP for the two organs. Compared to the current clinical practice, the
retrospective study indicated that online APT tended to spare more normal
tissues by achieving the clinical goal with merely half of the positional
uncertainty margin. Future studies are needed to help identify the patients
with large potential benefits prior to treatment to conserve scarce clinical
resources.
Authors: Chih-Wei Chang, Duncan Bohannon, Zhen Tian, Yinan Wang, Mark W. Mcdonald, David S. Yu, Tian Liu, Jun Zhou, Xiaofeng Yang.
This is the natural dynamical analogue of the result for a fixed time by Hughes, Keating and O'Connell [1]. In the course of this research we also proved a Wick-type identity, which we include in this paper, as it might be of independent interest.
We prove that the real and imaginary parts of the logarithm of the
characteristic polynomial of unitary Brownian motion converge to Gaussian free
fields on the cylinder, as the matrix dimension goes to infinity. This is the
natural dynamical analogue of the result for a fixed time by Hughes, Keating
and O'Connell [1]. Further it complements a result by Spohn [2] on linear
statistics of unitary Brownian motion, and a recent result by Bourgade and
Falconet [3] connecting the characteristic polynomial of unitary Brownian
motion to a Gaussian multiplicative chaos measure. In the course of this
research we also proved a Wick-type identity, which we include in this paper,
as it might be of independent interest.
Authors: Johannes Forkel, Isao Sauzedde.
The depolarization-field-screening-driven
expansion is separate from a photoacoustic pulse launched from the bottom
electrode on which the superlattice was epitaxially grown. The magnitude of
expansion in BaTiO3 layers is larger than the contraction in CaTiO3. The depolarization-field-screening-driven polarization
reduction in the CaTiO3 layers points to a new direction for the manipulation
of polarization in the component layers of a strongly coupled
ferroelectric/dielectric superlattice.
Above-bandgap femtosecond optical excitation of a ferroelectric/dielectric
BaTiO3/CaTiO3 superlattice leads to structural responses that are a consequence
of the screening of the strong electrostatic coupling between the component
layers. Time-resolved x-ray free-electron laser diffraction shows that the
structural response to optical excitation includes a net lattice expansion of
the superlattice consistent with depolarization-field screening driven by the
photoexcited charge carriers. The depolarization-field-screening-driven
expansion is separate from a photoacoustic pulse launched from the bottom
electrode on which the superlattice was epitaxially grown. The distribution of
diffracted intensity of superlattice x-ray reflections indicates that the
depolarization-field-screening-induced strain includes a photoinduced expansion
in the ferroelectric BaTiO3 and a contraction in CaTiO3. The magnitude of
expansion in BaTiO3 layers is larger than the contraction in CaTiO3. The
difference in the magnitude of depolarization-field-screening-driven strain in
the BaTiO3 and CaTiO3 components can arise from the contribution of the oxygen
octahedral rotation patterns at the BaTiO3/CaTiO3 interfaces to the
polarization of CaTiO3. The depolarization-field-screening-driven polarization
reduction in the CaTiO3 layers points to a new direction for the manipulation
of polarization in the component layers of a strongly coupled
ferroelectric/dielectric superlattice.
Authors: Deepankar Sri Gyan, Hyeon Jun Lee, Youngjun Ahn, Jerome Carnis, Tae Yeon Kim, Sanjith Unithrattil, Jun Young Lee, Sae Hwan Chun, Sunam Kim, Intae Eom, Minseok Kim, Sang-Youn Park, Kyung Sook Kim, Ho Nyung Lee, Ji Young Jo, Paul G. Evans.
We are interested in the weak noise regime for the observation equation. In particular, we demonstrate that there is a sharp phase transition between a spiking regime and a regime with perfect smoothing.
We study the celebrated Shiryaev-Wonham filter in its historical setup of
Wonham (1964) where the hidden Markov jump process has two states. We are
interested in the weak noise regime for the observation equation.
Interestingly, this becomes a strong noise regime for the filtering equations.
Earlier results of the authors show the appearance of spikes in the filtered
process, akin to a metastability phenomenon. This paper is aimed at
understanding the smoothed optimal filter, which is relevant for any system
with feedback. In particular, we demonstrate that there is a sharp phase
transition between a spiking regime and a regime with perfect smoothing.
Authors: Bernardin Cédric, Chhaibi Reda, Najnudel Joseph, Pellegrini Clément.
Recent ARPES experiments on doped 1D cuprates revealed the importance of
effective near-neighbor (NN) attractions in explaining certain features in
spectral functions. From state-of-the-art TDVP calculations, we find that the spectral
weights of the holon-folding and $3k_F$ branches evolve oppositely as a
function of $V$. This peculiar dichotomy may be explained in bosonization
analysis from the opposite dependence of exponent that determines the spectral
weights on Luttinger parameter $K_{\rho}$.
Recent ARPES experiments on doped 1D cuprates revealed the importance of
effective near-neighbor (NN) attractions in explaining certain features in
spectral functions. Here we investigate spectral properties of the extended
Hubbard model with the on-site repulsion $U$ and NN interaction $V$, by
employing bosonization analysis and the high-precision time-dependent
variational principle (TDVP) calculations of the model on 1D chain with up to
300 sites. From state-of-the-art TDVP calculations, we find that the spectral
weights of the holon-folding and $3k_F$ branches evolve oppositely as a
function of $V$. This peculiar dichotomy may be explained in bosonization
analysis from the opposite dependence of exponent that determines the spectral
weights on Luttinger parameter $K_{\rho}$. Moreover, our TDVP calculations of
models with fixed $U=8t$ and different $V$ show that $V\approx -1.7t$ may fit
the experimental results best, indicating a moderate effective NN attraction in
1D cuprates that might provide some hints towards understanding
superconductivity in 2D cuprates.
Authors: Hao-Xin Wang, Yi-Ming Wu, Yi-Fan Jiang, Hong Yao.
The hunt for such elusive signals requires accurate simulations to characterise the detector response and estimate the experimental sensitivity.
The physics programme of the MEG II experiment can be extended with the
search for new invisible particles produced in rare muon decays. The hunt for
such elusive signals requires accurate simulations to characterise the detector
response and estimate the experimental sensitivity. This work presents an
improved simulation of muon decay in MEG II, based on McMule and Geant4.
Authors: A. Gurgone, A. Papa, P. Schwendimann, A. Signer, Y. Ulrich, A. M. Baldini, F. Cei, M. Chiappini, M. Francesconi, L. Galli, M. Grassi, D. Nicolò, G. Signorelli.
The CERN Large Hadron Collider was built to uncover fundamental particles and
their interactions at the energy frontier. We propose an
algorithm with a greatly enhanced capability of disentangling individual proton
collisions to obtain a new global event description, considerably improving
over the current state-of-the-art.
The CERN Large Hadron Collider was built to uncover fundamental particles and
their interactions at the energy frontier. Upon entering its High Luminosity
phase at the end of this decade, the unprecedented interaction rates when
colliding two proton bunches will pose a significant challenge to the
reconstruction algorithms of experiments such as ATLAS and CMS. We propose an
algorithm with a greatly enhanced capability of disentangling individual proton
collisions to obtain a new global event description, considerably improving
over the current state-of-the-art. Employing a metric inspired by optimal
transport problems as the cost function of a graph neural network enhanced with
attention mechanisms, our algorithm is able to compare two particle collections
with different noise levels, thereby learning to correctly flag particles
originating from the main proton interaction amidst products from up to 200
simultaneous pileup collisions. The adoption of such an algorithm will lead to
a quasi-global sensitivity improvement for searches for new physics phenomena
and will enable precision measurements at the percent level in the High
Luminosity era of the Large Hadron Collider.
Authors: Loukas Gouskos, Fabio Iemmi, Sascha Liechti, Benedikt Maier, Vinicius Mikuni, Huilin Qu.
In this paper we provide the first neutrino-related explanation of the most energetic 18 TeV event reported by LHAASO. We find that the minimal viable scenario involves both mixing and transition magnetic moment portal between light and sterile neutrinos. Our explanation of this event, while being consistent with the terrestrial constraints, points to the non-standard cosmology.
LHAASO collaboration detected photons with energy above 10 TeV from the most
recent gamma-ray burst (GRB), GRB221009A. Given the redshift of this event,
$z\sim 0.15$, photons of such energy are expected to interact with the diffuse
extragalactic background light (EBL) well before reaching Earth. In this paper
we provide the first neutrino-related explanation of the most energetic 18 TeV
event reported by LHAASO. We find that the minimal viable scenario involves
both mixing and transition magnetic moment portal between light and sterile
neutrinos. The production of sterile neutrinos occurs efficiently via mixing
while the transition magnetic moment portal governs the decay rate in the
parameter space where tree-level decays via mixing to non-photon final states
are suppressed. Our explanation of this event, while being consistent with the
terrestrial constraints, points to the non-standard cosmology.
Authors: Vedran Brdar, Ying-Ying Li.