arXiv subjects

Daniel George

Publications and source records attributed to Daniel George.

At least 19 recordsLinked to original sources

arXiv2026

A compact low-frequency optomechanical triaxial inertial sensor

Triaxial optomechanical accelerometers offer compact, ground-testable alternatives to electrostatic sensors for satellite geodesy, seismometry, and various other applications. We demonstrate a low-CSWaP triaxial sensor using monolithic fused-silica resonators with dual heterodyne interferometric readout. The X axis reaches a 60~pico-g/$\sqrt{\mathrm{Hz}}$ readout noise floor, and the in-plane axes resolve ambient seismic ground motion in agreement with a co-located commercial seismometer from 4~mHz to 8~Hz. The Z axis exhibits higher noise under $1g$ loading due to gravity-induced geometric stiffening (44.45~Hz versus 11.07~Hz in a 0g-equivalent configuration), with measurements indicating in-orbit performance on par with the in-plane sensors. These results support the feasibility of triaxial optomechanical accelerometry for space missions and ground-based applications.

physics.optics↗

arXiv2026

SLIP: Segmentation with Low-latency Interactive Prompting for 3D Medical Images

Interactive deep image segmentation enables efficient medical image annotation by iteratively refining predictions from user prompts, such as positive and negative clicks. Recent patch-based methods, including nnInteractive, achieve strong segmentation performance but remain limited in annotation workflows by high interaction latency, limited responsiveness to successive interactions, and the lack of support for reversible prompting. Furthermore, evaluation relies predominantly on simulated rather than controlled real-user interaction studies. We present SLIP, an end-to-end trainable framework for interactive 3D medical image segmentation that decouples image encoding from prompt-guided refinement. Image features are computed once and reused, while a lightweight patch memory bank maintains an interaction-aware segmentation state shared across patches. This representation enables prediction updates by propagating interaction context throughout the image, supports reversible prompting without recomputing image features, and substantially reduces interaction latency. By separating image representation from interactive reasoning, SLIP remains compatible with a wide range of image encoders. We train a single SLIP model for general interactive segmentation across diverse anatomical structures and imaging modalities. Beyond standard simulated evaluation, we conduct a controlled prospective user study comparing manual segmentation, nnInteractive, and SLIP across three clinical annotation tasks, six expert participants, and subjective usability measures, addressing the limited human validation of interactive segmentation methods. SLIP achieves SOTA interactive segmentation performance across 13 public datasets while providing lower interaction latency, greater responsiveness, support for reversible prompting, and higher user preference than existing approaches.

cs.CV↗

arXiv2026

Not Every Rubric Teaches Equally: Policy-Aware Rubric Rewards for RLVR

Reinforcement learning with verifiable rewards has made post-training highly effective when correctness can be checked automatically. However, many important model behaviors require satisfying several qualitative criteria at once. Rubric-based rewards address this setting by grading prompt-specific criteria and aggregating them into a scalar reward. Yet standard static aggregations conflate a criterion's human-assigned importance with its current usefulness as an optimization signal. We show that this assumption breaks down in rubric RL: many important criteria are already saturated or currently unreachable, while criteria that distinguish rollouts are not necessarily those with the largest human weights. We introduce POW3R, a policy-aware rubric reward framework that preserves human weights and category balance as the rubric objective while adapting criterion-level reward weights during training. POW3R uses rollout-level contrast to emphasize criteria that currently separate the policy's outputs, making the GRPO reward more informative without changing the underlying evaluation target. Across three base policies on two datasets spanning multimodal and text-only settings, POW3R wins $24$ of $30$ base-policy/metric comparisons, improving both mean rubric reward and strict completion (the fraction of prompts whose response satisfies every required rubric criterion) over vanilla GRPO with rubric rewards, and reaches the same plateau in $2.5$--$4\times$ fewer training steps. Rubric rewards should therefore distinguish what should matter in the final answer from what can teach the current policy.

cs.AI↗

arXiv2026

Layout-Aware Representation Learning for Open-Set ID Fraud Discovery

Identity-document fraud detection is not a stationary binary classification problem. Adaptive attackers modify templates and fabrication pipelines, making historical fraud labels stale, and successful forgeries recur at scale as coherent campaigns. We therefore study layout-aware representation learning for open-set fraud discovery rather than only closed-set classification. We adapt DINOv3 to the document domain via context-aware SimMIM fine-tuning and supervised metric learning with composite loss that encourages inter-class separability and intra-class compactness. The model is trained with U.S. IDs only. With a lightweight MLP and softmax classifier, the embedding achieves 99.83% layout classification accuracy on Canadian layouts. Moreover, on a dataset of 20,448 Canadian IDs, embedding-space analysis surfaces 276 adaptive physical-fraud cases, including 222 not surfaced by incumbent detectors. The embedding supports similarity-based expansion from a single confirmed seed to additional related cases not linked by conventional metadata graphs. The layout-aware document embeddings provide a production-aligned basis for discovering novel and campaign-scale fraud under distribution shift.

cs.CV↗

arXiv2026

From Measurement to Mitigation: Quantifying and Reducing Identity Leakage in Image Representation Encoders with Linear Subspace Removal

Frozen visual embeddings (e.g., CLIP, DINOv2/v3, SSCD) power retrieval and integrity systems, yet their use on face-containing data is constrained by unmeasured identity leakage and a lack of deployable mitigations. We take an attacker-aware view and contribute: (i) a benchmark of visual embeddings that reports open-set verification at low false-accept rates, a calibrated diffusion-based template inversion check, and face-context attribution with equal-area perturbations; and (ii) propose a one-shot linear projector that removes an estimated identity subspace while preserving the complementary space needed for utility, which for brevity we denote as the identity sanitization projection ISP. Across CelebA-20 and VGGFace2, we show that these encoders are robust under open-set linear probes, with CLIP exhibiting relatively higher leakage than DINOv2/v3 and SSCD, robust to template inversion, and are context-dominant. In addition, we show that ISP drives linear access to near-chance while retaining high non-biometric utility, and transfers across datasets with minor degradation. Our results establish the first attacker-calibrated facial privacy audit of non-FR encoders and demonstrate that linear subspace removal achieves strong privacy guarantees while preserving utility for visual search and retrieval.

cs.CV↗

arXiv2024

Tilt-To-Length Coupling in LISA -- Uncertainty and Biases

The coupling of the angular jitter of the spacecraft and their sub-assemblies with the optical bench and the telescope into the interferometric length readout will be a major noise source in the LISA mission. We refer to this noise as tilt-to-length (TTL) coupling. It will be reduced directly by realignments, and the residual noise will then be subtracted in post-processing. The success of these mitigation strategies depends on an accurate computation of the TTL coupling coefficients. We present here a thorough analysis of the accuracy of the coefficient estimation under different jitter characteristics, angular readout noise levels, and gravitational wave sources. We analyze in which cases the estimates degrade using two estimators, the common least squares estimator and the instrumental variables estimator. Our investigations show that angular readout noise leads to a systematic bias of the least squares estimator, depending on the TTL coupling coefficients, jitter and readout noise level, while the instrumental variable estimator converges to an unbiased result as the data set length increases. We present an equation that predicts the estimation bias of the least squares method due to angular readout noise.

astro-ph.IM↗

arXiv2024

Multiscale Mechanical Modeling of Skeletal Muscle: A Systemic Review of the Literature

Purpose: From the myofibrils to the whole muscle scale, muscle micro-constituents exhibit passive and active mechanical properties, potentially coupled to electrical, chemical, and thermal properties. Experimental characterization of some of these properties is currently not available for all muscle constituents. Multiscale multiphysics models have recently gained interest as a numerical alternative to investigate the healthy and diseased physiological behavior of the skeletal muscle. Methods: This paper refers to the multiscale mechanical models proposed in the literature to investigate the mechanical properties and behavior of skeletal muscles. More specifically, we focus on the scale transition methods, constitutive laws and experimental data implemented in these models. Results: Using scale transition methods such as homogenization, coupled to appropriate constitutive behavior of the constituents, these models explore the mechanisms of ageing, myopathies, sportive injuries, and muscle contraction. Conclusion: Emerging trends include the development of multiphysics simulations and the coupling of modeling with the acquisition of experimental data at different scales, with increasing focus to little known constituents such as the extracellular matrix and the protein titin.

physics.med-ph↗

arXiv2023

TRUSTED: The Paired 3D Transabdominal Ultrasound and CT Human Data for Kidney Segmentation and Registration Research

Inter-modal image registration (IMIR) and image segmentation with abdominal Ultrasound (US) data has many important clinical applications, including image-guided surgery, automatic organ measurement and robotic navigation. However, research is severely limited by the lack of public datasets. We propose TRUSTED (the Tridimensional Renal Ultra Sound TomodEnsitometrie Dataset), comprising paired transabdominal 3DUS and CT kidney images from 48 human patients (96 kidneys), including segmentation, and anatomical landmark annotations by two experienced radiographers. Inter-rater segmentation agreement was over 94 (Dice score), and gold-standard segmentations were generated using the STAPLE algorithm. Seven anatomical landmarks were annotated, important for IMIR systems development and evaluation. To validate the dataset's utility, 5 competitive Deep Learning models for automatic kidney segmentation were benchmarked, yielding average DICE scores from 83.2% to 89.1% for CT, and 61.9% to 79.4% for US images. Three IMIR methods were benchmarked, and Coherent Point Drift performed best with an average Target Registration Error of 4.53mm. The TRUSTED dataset may be used freely researchers to develop and validate new segmentation and IMIR methods.

eess.IV↗

arXiv2022

Calculating the precision of tilt-to-length coupling estimation and noise subtraction in LISA using Fisher information

Tilt-to-length (TTL) noise from angular jitter in LISA is projected to be the dominant noise source in the milli-Hertz band unless corrected in post-processing. The correction is only possible after removing the overwhelming laser phase noise using time-delay interferometry (TDI). We present here a frequency domain model that describes the effect of angular motion of all three spacecraft on the interferometric signals after propagating through TDI. We then apply a Fisher information matrix analysis to this model to calculate the minimum uncertainty with which TTL coupling coefficients may be estimated. Furthermore, we show the impact of these uncertainties on the residual TTL noise in the gravitational wave readout channel, and compare it to the impact of the angular witness sensors' readout noise. We show that the residual TTL noise post-subtraction in the TDI variables for a case using the LISA angular jitter requirement and integration time of one day is limited to the 8\,pm/$\sqrt{\rm Hz}$ level by angular sensing noise. However, using a more realistic model for the angular jitter we find that the TTL coupling uncertainties are 70 times larger, and the noise subtraction is limited by these uncertainties to the 14\,pm/$\sqrt{\rm Hz}$ level.

gr-qc↗

arXiv2021

QUOTAS: A new research platform for the data-driven investigation of black holes

We present QUOTAS, a novel research platform for the data-driven investigation of super-massive black hole (SMBH) populations. While SMBH data sets -- observations and simulations -- have grown rapidly in complexity and abundance, our computational environments and analysis tools have not matured commensurately to exhaust opportunities for discovery. Motivated to explore BH host galaxy and the parent dark matter halo connection, in this pilot version of QUOTAS, we assemble and co-locate the high-redshift, luminous quasar population at $z \geq 3$ alongside simulated data of the same epochs. Leveraging machine learning algorithms (ML) we expand simulation volumes that successfully replicate halo populations beyond the training set. Training ML on the Illustris-TNG300 simulation that includes baryonic physics, we populate the larger LEGACY Expanse dark matter-only box with quasars. Our first science results comparing observational and ML simulated quasars at $z \sim 3$, reveal that while the recovered Black Hole Mass Functions and clustering are in good agreement, simulated SMBHs fail to accrete, shine and grow at high enough rates to match observed quasars. We conclude that sub-grid models of mass accretion and SMBH feedback implemented in Illustris-TNG300 do not reproduce their observed mass growth. QUOTAS, demonstrates the power of ML, both for analyzing large complex datasets, and offering a unique opportunity to interrogate our theoretical model assumptions. We deploy ML again to derive and devise an optimal survey strategy for bringing the undetected lower luminosity quasar population into view. QUOTAS, and all related materials are publicly available at the Google Kaggle platform.

astro-ph.CO↗

arXiv2019

Enabling real-time multi-messenger astrophysics discoveries with deep learning

Multi-messenger astrophysics is a fast-growing, interdisciplinary field that combines data, which vary in volume and speed of data processing, from many different instruments that probe the Universe using different cosmic messengers: electromagnetic waves, cosmic rays, gravitational waves and neutrinos. In this Expert Recommendation, we review the key challenges of real-time observations of gravitational wave sources and their electromagnetic and astroparticle counterparts, and make a number of recommendations to maximize their potential for scientific discovery. These recommendations refer to the design of scalable and computationally efficient machine learning algorithms; the cyber-infrastructure to numerically simulate astrophysical sources, and to process and interpret multi-messenger astrophysics data; the management of gravitational wave detections to trigger real-time alerts for electromagnetic and astroparticle follow-ups; a vision to harness future developments of machine learning and cyber-infrastructure resources to cope with the big-data requirements; and the need to build a community of experts to realize the goals of multi-messenger astrophysics.

gr-qc↗

arXiv2019

Deep Learning for Multi-Messenger Astrophysics: A Gateway for Discovery in the Big Data Era

This report provides an overview of recent work that harnesses the Big Data Revolution and Large Scale Computing to address grand computational challenges in Multi-Messenger Astrophysics, with a particular emphasis on real-time discovery campaigns. Acknowledging the transdisciplinary nature of Multi-Messenger Astrophysics, this document has been prepared by members of the physics, astronomy, computer science, data science, software and cyberinfrastructure communities who attended the NSF-, DOE- and NVIDIA-funded "Deep Learning for Multi-Messenger Astrophysics: Real-time Discovery at Scale" workshop, hosted at the National Center for Supercomputing Applications, October 17-19, 2018. Highlights of this report include unanimous agreement that it is critical to accelerate the development and deployment of novel, signal-processing algorithms that use the synergy between artificial intelligence (AI) and high performance computing to maximize the potential for scientific discovery with Multi-Messenger Astrophysics. We discuss key aspects to realize this endeavor, namely (i) the design and exploitation of scalable and computationally efficient AI algorithms for Multi-Messenger Astrophysics; (ii) cyberinfrastructure requirements to numerically simulate astrophysical sources, and to process and interpret Multi-Messenger Astrophysics data; (iii) management of gravitational wave detections and triggers to enable electromagnetic and astro-particle follow-ups; (iv) a vision to harness future developments of machine and deep learning and cyberinfrastructure resources to cope with the scale of discovery in the Big Data Era; (v) and the need to build a community that brings domain experts together with data scientists on equal footing to maximize and accelerate discovery in the nascent field of Multi-Messenger Astrophysics.

astro-ph.IM↗

arXiv2018

Fusing numerical relativity and deep learning to detect higher-order multipole waveforms from eccentric binary black hole mergers

We determine the mass-ratio, eccentricity and binary inclination angles that maximize the contribution of the higher-order waveform multipoles $(\ell, \, |m|)= \{(2,\,2),\, (2,\,1),\, (3,\,3),\, (3,\,2), \, (3,\,1),\, (4,\,4),\, (4,\,3),\, (4,\,2),\,(4,\,1)\}$ for the gravitational wave detection of eccentric binary black hole mergers. We carry out this study using numerical relativity waveforms that describe non-spinning black hole binaries with mass-ratios $1\leq q \leq 10$, and orbital eccentricities as high as $e_0=0.18$ fifteen cycles before merger. For stellar-mass, asymmetric mass-ratio, binary black hole mergers, and assuming LIGO's Zero Detuned High Power configuration, we find that in regions of parameter space where black hole mergers modeled with $\ell=|m|=2$ waveforms have vanishing signal-to-noise ratios, the inclusion of $(\ell, \, |m|)$ modes enables the observation of these sources with signal-to-noise ratios that range between 30\% to 45\% the signal-to-noise ratio of optimally oriented binary black hole mergers modeled with $\ell=|m|=2$ numerical relativity waveforms. Having determined the parameter space where $(\ell, \, |m|)$ modes are important for gravitational wave detection, we construct waveform signals that describe these astrophysically motivate scenarios, and demonstrate that these topologically complex signals can be detected and characterized in real LIGO noise with deep learning algorithms.

gr-qc↗

arXiv2018

Real-time regression analysis with deep convolutional neural networks

We discuss the development of novel deep learning algorithms to enable real-time regression analysis for time series data. We showcase the application of this new method with a timely case study, and then discuss the applicability of this approach to tackle similar challenges across science domains.

cs.LG↗

arXiv2018

A Carillon of Black Holes

Scientists collaborating internationally have developed a new way to learn about our universe through gravitational waves, which are ripples in space-time caused by the motion and vibration of celestial bodies. By analogy, gravitational waves are akin to the vibrations carried through the air as sound. Quite remarkably, black holes, which are the densest objects in the universe formed from dead stars can vibrate and emit gravitational waves at frequencies that are within the range of human hearing once the gravitational waves are detected and amplified by instruments such as the LIGO and Virgo gravitational wave detectors. In this work, we explore how to make musical instruments based on gravitational waves by mapping a different gravitational wave pattern to each of the 88 keys of a piano, much like a carillon, which has its bells mapped to the batons of a carillon-keyboard. We rely on theoretical calculations for black hole vibrations to construct our digital black hole instruments. Our software and music samples are freely available to those who want to explore the music of gravitational waves.

gr-qc↗

arXiv2017

Denoising Gravitational Waves using Deep Learning with Recurrent Denoising Autoencoders

Gravitational wave astronomy is a rapidly growing field of modern astrophysics, with observations being made frequently by the LIGO detectors. Gravitational wave signals are often extremely weak and the data from the detectors, such as LIGO, is contaminated with non-Gaussian and non-stationary noise, often containing transient disturbances which can obscure real signals. Traditional denoising methods, such as principal component analysis and dictionary learning, are not optimal for dealing with this non-Gaussian noise, especially for low signal-to-noise ratio gravitational wave signals. Furthermore, these methods are computationally expensive on large datasets. To overcome these issues, we apply state-of-the-art signal processing techniques, based on recent groundbreaking advancements in deep learning, to denoise gravitational wave signals embedded either in Gaussian noise or in real LIGO noise. We introduce SMTDAE, a Staired Multi-Timestep Denoising Autoencoder, based on sequence-to-sequence bi-directional Long-Short-Term-Memory recurrent neural networks. We demonstrate the advantages of using our unsupervised deep learning approach and show that, after training only using simulated Gaussian noise, SMTDAE achieves superior recovery performance for gravitational wave signals embedded in real non-Gaussian LIGO noise.

gr-qc↗

arXiv2017

Deep Learning for Real-time Gravitational Wave Detection and Parameter Estimation with LIGO Data

The recent Nobel-prize-winning detections of gravitational waves from merging black holes and the subsequent detection of the collision of two neutron stars in coincidence with electromagnetic observations have inaugurated a new era of multimessenger astrophysics. To enhance the scope of this emergent science, we proposed the use of deep convolutional neural networks for the detection and characterization of gravitational wave signals in real-time. This method, Deep Filtering, was initially demonstrated using simulated LIGO noise. In this article, we present the extension of Deep Filtering using real data from the first observing run of LIGO, for both detection and parameter estimation of gravitational waves from binary black hole mergers with continuous data streams from multiple LIGO detectors. We show for the first time that machine learning can detect and estimate the true parameters of a real GW event observed by LIGO. Our comparisons show that Deep Filtering is far more computationally efficient than matched-filtering, while retaining similar sensitivity and lower errors, allowing real-time processing of weak time-series signals in non-stationary non-Gaussian noise, with minimal resources, and also enables the detection of new classes of gravitational wave sources that may go unnoticed with existing detection algorithms. This approach is uniquely suited to enable coincident detection campaigns of gravitational waves and their multimessenger counterparts in real-time.

gr-qc↗

arXiv2017

Eccentric, nonspinning, inspiral, Gaussian-process merger approximant for the detection and characterization of eccentric binary black hole mergers

We present $\texttt{ENIGMA}$, a time domain, inspiral-merger-ringdown waveform model that describes non-spinning binary black holes systems that evolve on moderately eccentric orbits. The inspiral evolution is described using a consistent combination of post-Newtonian theory, self-force and black hole perturbation theory. Assuming eccentric binaries that circularize prior to coalescence, we smoothly match the eccentric inspiral with a stand-alone, quasi-circular merger, which is constructed using machine learning algorithms that are trained with quasi-circular numerical relativity waveforms. We show that $\texttt{ENIGMA}$ reproduces with excellent accuracy the dynamics of quasi-circular compact binaries. We validate $\texttt{ENIGMA}$ using a set of $\texttt{Einstein Toolkit}$ eccentric numerical relativity waveforms, which describe eccentric binary black hole mergers with mass-ratios between $1 \leq q \leq 5.5$, and eccentricities $e_0 \lesssim 0.2$ ten orbits before merger. We use this model to explore in detail the physics that can be extracted with moderately eccentric, non-spinning binary black hole mergers. We use $\texttt{ENIGMA}$ to show that GW150914, GW151226, GW170104, GW170814 and GW170608 can be effectively recovered with spinning, quasi-circular templates if the eccentricity of these events at a gravitational wave frequency of 10Hz satisfies $e_0\leq \{0.175,\, 0.125,\,0.175,\,0.175,\, 0.125\}$, respectively. We show that if these systems have eccentricities $e_0\sim 0.1$ at a gravitational wave frequency of 10Hz, they can be misclassified as quasi-circular binaries due to parameter space degeneracies between eccentricity and spin corrections. Using our catalog of eccentric numerical relativity simulations, we discuss the importance of including higher-order waveform multipoles in gravitational wave searches of eccentric binary black hole mergers.

gr-qc↗