arXiv subjects

David Carlson

Publications and source records attributed to David Carlson.

At least 37 records · Page 2Linked to original sources

arXiv2022

Universal visible emitters in nanoscale integrated photonics

Visible wavelengths of light control the quantum matter of atoms and molecules and are foundational for quantum technologies, including computers, sensors, and clocks. The development of visible integrated photonics opens the possibility for scalable circuits with complex functionalities, advancing both the scientific and technological frontiers. We experimentally demonstrate an inverse design approach based on superposition of guided-mode sources, allowing the generation and full control of free-space radiation directly from within a single 150 nm layer Ta2O5, showing low loss across visible and near-infrared spectra. We generate diverging circularly-polarized beams at the challenging 461 nm wavelength that can be directly used for magneto-optical traps of strontium atoms, constituting a fundamental building block for a range of atomic-physics-based quantum technologies. Our generated topological vortex beams and spatially-varying polarization emitters could open unexplored light-matter interaction pathways, enabling a broad new photonic-atomic paradigm. Our platform highlights the generalizability of nanoscale devices for visible-laser emission and will be critical for scaling quantum technologies.

physics.optics↗

arXiv2022

Multiple Domain Causal Networks

Observational studies are regarded as economic alternatives to randomized trials, often used in their stead to investigate and determine treatment efficacy. Due to lack of sample size, observational studies commonly combine data from multiple sources or different sites/centers. Despite the benefits of an increased sample size, a naive combination of multicenter data may result in incongruities stemming from center-specific protocols for generating cohorts or reactions towards treatments distinct to a given center, among other things. These issues arise in a variety of other contexts, including capturing a treatment effect related to an individual's unique biological characteristics. Existing methods for estimating heterogeneous treatment effects have not adequately addressed the multicenter context, but rather treat it simply as a means to obtain sufficient sample size. Additionally, previous approaches to estimating treatment effects do not straightforwardly generalize to the multicenter design, especially when required to provide treatment insights for patients from a new, unobserved center. To address these shortcomings, we propose Multiple Domain Causal Networks (MDCN), an approach that simultaneously strengthens the information sharing between similar centers while addressing the selection bias in treatment assignment through learning of a new feature embedding. In empirical evaluations, MDCN is consistently more accurate when estimating the heterogeneous treatment effect in new centers compared to benchmarks that adjust solely based on treatment imbalance or general center differences. Finally, we justify our approach by providing theoretical analyses that demonstrate that MDCN improves on the generalization bound of the new, unobserved target center.

stat.ML↗

arXiv2022

AugmentedPCA: A Python Package of Supervised and Adversarial Linear Factor Models

Deep autoencoders are often extended with a supervised or adversarial loss to learn latent representations with desirable properties, such as greater predictivity of labels and outcomes or fairness with respects to a sensitive variable. Despite the ubiquity of supervised and adversarial deep latent factor models, these methods should demonstrate improvement over simpler linear approaches to be preferred in practice. This necessitates a reproducible linear analog that still adheres to an augmenting supervised or adversarial objective. We address this methodological gap by presenting methods that augment the principal component analysis (PCA) objective with either a supervised or an adversarial objective and provide analytic and reproducible solutions. We implement these methods in an open-source Python package, AugmentedPCA, that can produce excellent real-world baselines. We demonstrate the utility of these factor models on an open-source, RNA-seq cancer gene expression dataset, showing that augmenting with a supervised objective results in improved downstream classification performance, produces principal components with greater class fidelity, and facilitates identification of genes aligned with the principal axes of data variance with implications to development of specific types of cancer.

stat.ML↗

arXiv2021

Estimating Uncertainty Intervals from Collaborating Networks

Effective decision making requires understanding the uncertainty inherent in a prediction. In regression, this uncertainty can be estimated by a variety of methods; however, many of these methods are laborious to tune, generate overconfident uncertainty intervals, or lack sharpness (give imprecise intervals). We address these challenges by proposing a novel method to capture predictive distributions in regression by defining two neural networks with two distinct loss functions. Specifically, one network approximates the cumulative distribution function, and the second network approximates its inverse. We refer to this method as Collaborating Networks (CN). Theoretical analysis demonstrates that a fixed point of the optimization is at the idealized solution, and that the method is asymptotically consistent to the ground truth distribution. Empirically, learning is straightforward and robust. We benchmark CN against several common approaches on two synthetic and six real-world datasets, including forecasting A1c values in diabetic patients from electronic health records, where uncertainty is critical. In the synthetic data, the proposed approach essentially matches ground truth. In the real-world datasets, CN improves results on many performance metrics, including log-likelihood estimates, mean absolute errors, coverage estimates, and prediction interval widths.

stat.ML↗

arXiv2021

Adversarial Factor Models for the Generation of Improved Autism Diagnostic Biomarkers

Discovering reliable measures that inform on autism spectrum disorder (ASD) diagnosis is critical for providing appropriate and timely treatment for this neurodevelopmental disorder. In this work we present applications of adversarial linear factor models in the creation of improved biomarkers for ASD diagnosis. First, we demonstrate that an adversarial linear factor model can be used to remove confounding information from our biomarkers, ensuring that they contain only pertinent information on ASD. Second, we show this same model can be used to learn a disentangled representation of multimodal biomarkers that results in an increase in predictive performance. These results demonstrate that adversarial methods can address both biomarker confounds and improve biomarker predictive performance.

cs.LG↗

arXiv2021

Ookami: Deployment and Initial Experiences

Ookami is a computer technology testbed supported by the United States National Science Foundation. It provides researchers with access to the A64FX processor developed by Fujitsu in collaboration with RIKΞN for the Japanese path to exascale computing, as deployed in Fugaku, the fastest computer in the world. By focusing on crucial architectural details, the ARM-based, multi-core, 512-bit SIMD-vector processor with ultrahigh-bandwidth memory promises to retain familiar and successful programming models while achieving very high performance for a wide range of applications. We review relevant technology and system details, and the main body of the paper focuses on initial experiences with the hardware and software ecosystem for micro-benchmarks, mini-apps, and full applications, and starts to answer questions about where such technologies fit into the NSF ecosystem.

cs.DC↗

arXiv2020

Mid-infrared frequency combs at 10 GHz

We demonstrate mid-infrared (MIR) frequency combs at 10 GHz repetition rate via intra-pulse difference-frequency generation (DFG) in quasi-phase-matched nonlinear media. Few-cycle pump pulses ($\mathbf{\lesssim}$15 fs, 100 pJ) from a near-infrared (NIR) electro-optic frequency comb are provided via nonlinear soliton-like compression in photonic-chip silicon-nitride waveguides. Subsequent intra-pulse DFG in periodically-poled lithium niobate waveguides yields MIR frequency combs in the 3.1--4.1 $μ$m region, while orientation-patterned gallium phosphide provides coverage across 7--11 $μ$m. Cascaded second-order nonlinearities simultaneously provide access to the carrier-envelope-offset frequency of the pump source via in-line f-2f nonlinear interferometry. The high-repetition rate MIR frequency combs introduced here can be used for condensed phase spectroscopy and applications such as laser heterodyne radiometry.

physics.optics↗

arXiv2019

Dynamic Embedding on Textual Networks via a Gaussian Process

Textual network embedding aims to learn low-dimensional representations of text-annotated nodes in a graph. Prior work in this area has typically focused on fixed graph structures; however, real-world networks are often dynamic. We address this challenge with a novel end-to-end node-embedding model, called Dynamic Embedding for Textual Networks with a Gaussian Process (DetGP). After training, DetGP can be applied efficiently to dynamic graphs without re-training or backpropagation. The learned representation of each node is a combination of textual and structural embeddings. Because the structure is allowed to be dynamic, our method uses the Gaussian process to take advantage of its non-parametric properties. To use both local and global graph structures, diffusion is used to model multiple hops between neighbors. The relative importance of global versus local structure for the embeddings is learned automatically. With the non-parametric nature of the Gaussian process, updating the embeddings for a changed graph structure requires only a forward pass through the learned model. Considering link prediction and node classification, experiments demonstrate the empirical effectiveness of our method compared to baseline approaches. We further show that DetGP can be straightforwardly and efficiently applied to dynamic textual networks.

cs.LG↗

arXiv2019

Ultranarrow linewidth photonic-atomic laser

Lasers with high spectral purity can enable a diverse application space, including precision spectroscopy, coherent high-speed communications, physical sensing, and manipulation of quantum systems. Already, meticulous design and construction of bench Fabry-Perot cavities has made possible dramatic achievements in active laser-linewidth reduction, predominantly for optical-atomic clocks. Yet there is increasing demand for miniaturized laser systems operating with high performance in ambient environments. Here, we report a compact and robust photonic-atomic laser comprising a 2.5 cm long, 20,000 finesse, monolithic Fabry-Perot cavity integrated with a micromachined rubidium vapor cell. By leveraging the short-time frequency stability of the cavity and the long-time frequency stability of atoms, we realize an ultranarrow-linewidth laser that enables integration for extended measurements. Specifically, our laser supports a fractional-frequency stability of $1\times 10^{-13}$ at an averaging time of 20 ms, $7 \times 10^{-13}$ at 300 s, an integrated linewidth of 25 Hz that results from thermal noise, a Lorentzian linewidth as low as 0.06 Hz$^2$/Hz, and a passive vibration immunity as low as $10^{-10}$/g. Our work explores hybrid laser systems with monolithic photonic and atomic packages based on physical design.

physics.optics↗

arXiv2019

The Secrets of Machine Learning: Ten Things You Wish You Had Known Earlier to be More Effective at Data Analysis

Despite the widespread usage of machine learning throughout organizations, there are some key principles that are commonly missed. In particular: 1) There are at least four main families for supervised learning: logical modeling methods, linear combination methods, case-based reasoning methods, and iterative summarization methods. 2) For many application domains, almost all machine learning methods perform similarly (with some caveats). Deep learning methods, which are the leading technique for computer vision problems, do not maintain an edge over other methods for most problems (and there are reasons why). 3) Neural networks are hard to train and weird stuff often happens when you try to train them. 4) If you don't use an interpretable model, you can make bad mistakes. 5) Explanations can be misleading and you can't trust them. 6) You can pretty much always find an accurate-yet-interpretable model, even for deep neural networks. 7) Special properties such as decision making or robustness must be built in, they don't happen on their own. 8) Causal inference is different than prediction (correlation is not causation). 9) There is a method to the madness of deep neural architectures, but not always. 10) It is a myth that artificial intelligence can do anything.

cs.LG↗

arXiv2019

StoryGAN: A Sequential Conditional GAN for Story Visualization

We propose a new task, called Story Visualization. Given a multi-sentence paragraph, the story is visualized by generating a sequence of images, one for each sentence. In contrast to video generation, story visualization focuses less on the continuity in generated images (frames), but more on the global consistency across dynamic scenes and characters -- a challenge that has not been addressed by any single-image or video generation methods. We therefore propose a new story-to-image-sequence generation model, StoryGAN, based on the sequential conditional GAN framework. Our model is unique in that it consists of a deep Context Encoder that dynamically tracks the story flow, and two discriminators at the story and image levels, to enhance the image quality and the consistency of the generated sequences. To evaluate the model, we modified existing datasets to create the CLEVR-SV and Pororo-SV datasets. Empirically, StoryGAN outperforms state-of-the-art models in image quality, contextual consistency metrics, and human evaluation.

cs.CV↗

arXiv2018

Tunable mid-infrared generation via wide-band four wave mixing in silicon nitride waveguides

We experimentally demonstrate wide-band (>100 THz) frequency down-conversion of near-infrared (NIR) femtosecond-scale pulses from an Er:fiber laser to the mid-infrared (MIR) using four-wave-mixing (FWM) in photonic-chip silicon-nitride waveguides. The engineered dispersion in the nanophotonic geometry, along with the wide transparency range of silicon nitride, enables large-detuning FWM phase-matching and results in tunable MIR from 2.6-3.6 um on a single chip with 100-pJ-scale pump-pulse energies. Additionally, we observe > 20 dB broadband parametric gain for the NIR pulses when the FWM process is operated in a frequency up-conversion configuration. Our results demonstrate how integrated photonic circuits could realize multiple nonlinear optical phenomena on the same chip and lead to engineered synthesis of broadband, tunable, and coherent light across the NIR and MIR wavelength bands from fiber-based pumps.

physics.app-ph↗

arXiv2018

Thermal and Nonlinear Dissipative-Soliton Dynamics in Kerr Microresonator Frequency Combs

We explore the dynamical response of dissipative Kerr solitons to changes in pump power and detuning and show how thermal and nonlinear processes couple these parameters to the frequency-comb degrees of freedom. Our experiments are enabled by a Pound-Drever-Hall (PDH) stabilization approach that provides on-demand, radiofrequency control of the frequency comb. PDH locking not only guides Kerr-soliton formation from a cold microresonator, but opens a path to decouple the repetition and carrier-envelope-offset frequencies. In particular, we demonstrate phase stabilization of both Kerr-comb degrees-of-freedom to a fractional frequency precision below 10^-16, compatible with optical-timekeeping technology. Moreover, we investigate the fundamental role that residual laser-resonator detuning noise plays in the spectral purity of microwave generation with Kerr combs.

physics.optics↗

arXiv2018

Deuterated silicon nitride photonic devices for broadband optical frequency comb generation

We report and characterize low-temperature, plasma-deposited deuterated silicon nitride thin films for nonlinear integrated photonics. With a peak processing temperature less than 300$^\circ$C, it is back-end compatible with pre-processed CMOS substrates. We achieve microresonators with a quality factor of up to $1.6\times 10^6 $ at 1552 nm, and $>1.2\times 10^6$ throughout $λ$ = 1510 -- 1600 nm, without annealing or stress management. We then demonstrate the immediate utility of this platform in nonlinear photonics by generating a 1 THz free spectral range, 900-nm-bandwidth modulation-instability microresonator Kerr comb and octave-spanning, supercontinuum-broadened spectra.

physics.optics↗

arXiv2017

Video Generation From Text

Generating videos from text has proven to be a significant challenge for existing generative models. We tackle this problem by training a conditional generative model to extract both static and dynamic information from text. This is manifested in a hybrid framework, employing a Variational Autoencoder (VAE) and a Generative Adversarial Network (GAN). The static features, called "gist," are used to sketch text-conditioned background color and object layout structure. Dynamic features are considered by transforming input text into an image filter. To obtain a large amount of data for training the deep-learning model, we develop a method to automatically create a matched text-video corpus from publicly available online videos. Experimental results show that the proposed framework generates plausible and diverse videos, while accurately reflecting the input text information. It significantly outperforms baseline models that directly adapt text-to-image generation procedures to produce videos. Performance is evaluated both visually and by adapting the inception score used to evaluate image generation in GANs.

cs.MM↗

arXiv2017

Stochastic Bouncy Particle Sampler

We introduce a novel stochastic version of the non-reversible, rejection-free Bouncy Particle Sampler (BPS), a Markov process whose sample trajectories are piecewise linear. The algorithm is based on simulating first arrival times in a doubly stochastic Poisson process using the thinning method, and allows efficient sampling of Bayesian posteriors in big datasets. We prove that in the BPS no bias is introduced by noisy evaluations of the log-likelihood gradient. On the other hand, we argue that efficiency considerations favor a small, controllable bias in the construction of the thinning proposals, in exchange for faster mixing. We introduce a simple regression-based proposal intensity for the thinning method that controls this trade-off. We illustrate the algorithm in several examples in which it outperforms both unbiased, but slowly mixing stochastic versions of BPS, as well as biased stochastic gradient-based samplers.

stat.CO↗

arXiv2017

Photonic-chip supercontinuum with tailored spectra for precision frequency metrology

Supercontinuum generation using chip-integrated photonic waveguides is a powerful approach for spectrally broadening pulsed laser sources with very low pulse energies and compact form factors. When pumped with a mode-locked laser frequency comb, these waveguides can coherently expand the comb spectrum to more than an octave in bandwidth to enable self-referenced stabilization. However, for applications in frequency metrology and precision spectroscopy, it is desirable to not only support self-referencing, but also to generate low-noise combs with customizable broadband spectra. In this work, we demonstrate dispersion-engineered waveguides based on silicon nitride that are designed to meet these goals and enable precision optical metrology experiments across large wavelength spans. We perform a clock comparison measurement and report a clock-limited relative frequency instability of $3.8\times10^{-15}$ at $τ= 2$ seconds between a 1550 nm cavity-stabilized reference laser and NIST's calcium atomic clock laser at 657 nm using a two-octave waveguide-supercontinuum comb.

physics.optics↗

arXiv2016

Bridging the Gap between Stochastic Gradient MCMC and Stochastic Optimization

Stochastic gradient Markov chain Monte Carlo (SG-MCMC) methods are Bayesian analogs to popular stochastic optimization methods; however, this connection is not well studied. We explore this relationship by applying simulated annealing to an SGMCMC algorithm. Furthermore, we extend recent SG-MCMC methods with two key components: i) adaptive preconditioners (as in ADAgrad or RMSprop), and ii) adaptive element-wise momentum weights. The zero-temperature limit gives a novel stochastic optimization method with adaptive element-wise momentum weights, while conventional optimization methods only have a shared, static momentum weight. Under certain assumptions, our theoretical analysis suggests the proposed simulated annealing approach converges close to the global optima. Experiments on several deep neural network models show state-of-the-art results compared to related stochastic optimization algorithms.

stat.ML↗