Search

More than 50 years ago, Bongard introduced 100 visual concept learning problems as a challenge for artificial vision systems. These problems are now known as Bongard problems. Although they are well known in cognitive science and artificial intelligence, only very little progress has been made toward building systems that can solve a substantial subset of them. In the system presented here, visual features are extracted through image processing and then translated into a symbolic visual vocabulary. We introduce a formal language that allows representing compositional visual concepts based on this vocabulary. Using this language and Bayesian inference, concepts can be induced from the examples that are provided in each problem. We find a reasonable agreement between the concepts with high posterior probability and the solutions formulated by Bongard himself for a subset of 35 problems. While this approach is far from solving Bongard problems like humans, it does considerably better than previous approaches. We discuss the issues we encountered while developing this system and their continuing relevance for understanding visual cognition. For instance, contrary to other concept learning problems, the examples are not random in Bongard problems; instead they are carefully chosen to ensure that the concept can be induced, and we found it helpful to take the resulting pragmatic constraints into account.

Abstract Goal-directed navigation requires continuously integrating uncertain self-motion and landmark cues into an internal sense of location and direction, concurrently planning future paths, and sequentially executing motor actions. Here, we provide a unified account of these processes with a computational model of probabilistic path planning in the framework of optimal feedback control under uncertainty. This model gives rise to diverse human navigational strategies previously believed to be distinct behaviors and predicts quantitatively both the errors and the variability of navigation across numerous experiments. This furthermore explains how sequential egocentric landmark observations form an uncertain allocentric cognitive map, how this internal map is used both in route planning and during execution of movements, and reconciles seemingly contradictory results about cue-integration behavior in navigation. Taken together, the present work provides a parsimonious explanation of how patterns of human goal-directed navigation behavior arise from the continuous and dynamic interactions of spatial uncertainties in perception, cognition, and action.

Despite being a minimally invasive sample source, dried blood spots (DBS) generally pose the challenge of efficient sample extraction for broad metabolomics applications. This is particularly true for the quantification of steroids using non-derivatized liquid chromatography tandem mass spectrometry assays. To address these limitations, we have demonstrated the use of electric field as a driver for sample preparation from DBS samples to simultaneously quantify testosterone (T), 17α-hydroxyprogesterone (17-OHP), progesterone (P), and cortisol (C), using both standard electroporation cuvettes, as well as a novel custom vertical electric field setup. Our findings, backed by computational modeling, show that a 10V DC application for 180 s can draw out twice the amount of the aforementioned steroids from both Whatman-903 and DMPK-C sample collection cards using our novel device when compared to standard solvent-based collection methods. This study not only introduces the use of electric field for sample preparation for metabolomics, but additionally introduces a novel device that eliminates the electric double layer effect in the process.

We construct a discrete non-Hermitian momentum operator, which implements faithfully the non-self-adjoint nature of momentum for a particle in a box. Its eigenfunctions are strictly limited to the interior of the box in the continuum limit, with the quarter wave as first nontrivial eigenstate. We show how to construct the corresponding Hermitian Hamiltonian for the infinite well as a concrete example to realize unitary dynamics. The resulting Hilbert space can be decomposed into a physical and unphysical subspace, which are mutually orthogonal. The physical subspace in the continuum limit reproduces that of the continuum theory and we give numerical evidence that the correct probability distributions for momentum and energy are recovered.

Abstract We present a novel strategy aimed at restoring correct convergence in complex Langevin simulations. The central idea is to incorporate system-specific prior knowledge into the simulations, in order to circumvent the NP-hard sign problem. In order to do so, we modify complex Langevin using kernels and propose the use of modern auto-differentiation methods to learn optimal kernel values. The optimization process is guided by functionals encoding relevant prior information, such as symmetries or Euclidean correlator data. Our approach recovers correct convergence in the non-interacting theory on the Schwinger-Keldysh contour for any real-time extent. For the strongly coupled quantum anharmonic oscillator we achieve correct convergence up to three-times the real-time extent of the previous benchmark study. An appendix sheds light on the fact that for correct convergence not only the absence of boundary terms, but in addition the correct Fokker-Plank spectrum is crucial.

Abstract Depression is a growing global crisis, with females at a higher rate of diagnosis than males. While the percentage of patients on prescribed antidepressants have tripled over the last two decades, we are still at a crossroad where the discrepancy lies between finding a drug to suit a patient and monitoring the abundance of it in the body to prevent unwanted side effects. Liquid Chromatography tandem mass spectrometry (LC–MS/MS) has garnered the attention of clinicians as a technique to accurately monitor therapeutic drugs in human serum with high specificity and accuracy. This may be a potential solution, but the challenge persists in the realm of sample preparation, where a method is automatable. We have developed and validated an LC–MS/MS-based assay for simultaneous quantification of 4 different classes of commonly prescribed antidepressants in women that is automated using a JANUS G3 Robotic Liquid Handler. Our method utilizes a simple sample preparation technique, utilizing only 20 μL of a serum sample, to accurately measure Bupropion, Citalopram, Desipramine, Imipramine, Olanzapine, Sertraline and Vilazodone across a range of 1.0 to 230 ng/mL. Our method exhibits a linearity of R2 ≥ 0.99 when detected in MRM mode and % CV of ≤ 20% for all analytes across the board. In addition, we have designed a prototype that can be utilized at a clinical mass spectrometry lab and assessed the long-term use of this prototype using an accelerated stability study. Overall, our developed method has the potential to be translated to clinical settings to monitor postpartum depression for a large number of patient samples using automation.

Abstract We generalize a complex heavy-quark potential model from an isotropic QCD plasma to an anisotropic one by replacing the Debye mass m D with an anisotropic screening mass depending on the quark pair alignment with respect to the direction of anisotropy. Such an angle-dependent mass is determined by matching the perturbative contributions in the potential model to the exact result obtained in the Hard-Thermal-Loop resummed perturbation theory. An advantage of the resulting potential model is that its angular dependence can be effectively described by using a set of angle-averaged screening masses as proposed in our previous work. Consequently, one could solve a one-dimensional Schrödinger equation with a potential model built by changing the anisotropic screening masses into the corresponding angle-averaged ones, and reproduce the full three-dimensional results for the binding energies and decay widths of low-lying quarkonium bound states to very high accuracy. Finally, turning to dynamics, we demonstrate that the one-dimensional effective potential can accurately describe the time evolution of the vacuum overlaps obtained using the full three-dimensional anisotropic potential. This includes the splitting of different p-wave polarizations.

Background:. Postoperative hand therapy (HT) is important for regaining function and preventing complications in patients undergoing tendon repair of the hand and wrist. Loss to follow-up (LTFU) can hinder this process; so we sought to determine factors that predict attrition of these patients. Methods:. Charts were retrospectively reviewed for patients who underwent extensor or flexor tendon repair of the hand, wrist, or forearm between 2014 and 2019. Demographic data, including age, sex, zip code, employment status, education level, and insurance type, were collected, and the rate of LTFU was calculated. Logistic regression was used to analyze factors. Results:. A total of 149 patients were identified and analyzed. The rate of LTFU was 42%. Factors that predicted loss were younger age, male gender, lower educational degree, and a documented psychiatric history. Employment status, insurance type, and distance from the HT center did not predict attrition. The number of HT weeks recommended by the occupational therapist did not differ between those who were lost and those who were not. Lost patients completed, on average, 57% of their suggested HT course. Conclusions:. The current study identified demographic factors associated with attrition in patients undergoing tendon repair of the distal upper extremity. Factors included patients who were younger, male gender, less educated, and had a documented psychiatric history. By identifying factors that predict LTFU, specific strategies can be developed to reduce attrition rates, particularly for at-risk populations, to improve patient care after tendon repair.

Combining experts’ subjective probability estimates is a fundamental task with broad applicability in domains ranging from finance to public health. However, it is still an open question how to combine such estimates optimally. Since the beta distribution is a common choice for modeling uncertainty about probabilities, here we propose a family of normative Bayesian models for aggregating probability estimates based on beta distributions. We systematically derive and compare different variants, including hierarchical and non-hierarchical as well as asymmetric and symmetric beta fusion models. Using these models, we show how the beta calibration function naturally arises in this normative framework and how it is related to the widely used Linear-in-Log-Odds calibration function. For evaluation, we provide the new Knowledge Test Confidence data set consisting of subjective probability estimates of 85 forecasters on 180 queries. On this and another data set, we show that the hierarchical symmetric beta fusion model performs best of all beta fusion models and outperforms related Bayesian fusion models in terms of mean absolute error.

The computation of dynamical properties of nuclear matter, ranging from parton distribution functions of nucleons and nuclei to transport properties in the quark-gluon plasma, constitutes a central goal of modern theoretical physics. This real-time physics often defies a perturbative treatment and the most successful strategy so far is to deploy lattice QCD simulations. These numerical computations are based on Monte-Carlo sampling and formulated in an artificial Euclidean time. Real-time physics is most conveniently formulated in terms of spectral functions, which are hidden in lattice QCD behind an ill-posed inverse problem. I will discuss state-of-the art methods in the extraction of spectral functions from lattice QCD simulations, based on Bayesian inference and emphasize the importance of prior domain knowledge, vital to regularizing the otherwise ill-posed extraction task. With Bayesian inference allowing us to make explicit the uncertainty in both observations and in our prior knowledge, a systematic estimation of the total uncertainties in the extracted spectral functions is nowadays possible. Two implementations of the Bayesian Reconstruction (BR) method for spectral function extraction, one for MAP point estimates and one based on an open access Monte-Carlo sampler are provided. I will briefly touch on the use of machine learning for spectral function reconstruction and discuss some new insight it has brought to the Bayesian community.

Abstract This study explores the potential of modern implicit solvers for stochastic partial differential equations in the simulation of real-time complex Langevin dynamics. Not only do these methods offer asymptotic stability, rendering the issue of runaway solution moot, but they also allow us to simulate at comparatively large Langevin time steps, leading to lower computational cost. We compare different ways of regularizing the underlying path integral and estimate the errors introduced due to the finite Langevin time steps. Based on that insight, we implement benchmark (non-)thermal simulations of the quantum anharmonic oscillator on the canonical Schwinger-Keldysh contour of short real-time extent.

Abstract We compute the proper real-time interaction potential between a static quark and antiquark in classical lattice gauge theory at finite temperature. Our central result is the determination of the screened real-part of this potential, and we reconfirm the presence of an imaginary part. The real part is intimately related to the back-reaction of the static sources onto the gauge fields, incorporated via Gauss’s law. Differences in the treatment of static sources in quantum and classical lattice gauge theory are discussed.

Psychophysical methods are a cornerstone of psychology, cognitive science, and neuroscience where they have been used to quantify behavior and its neural correlates for a vast range of mental phenomena. Their power derives from the combination of controlled experiments and rigorous analysis through signal detection theory. Unfortunately, they require many tedious trials and preferably highly trained participants. A recently developed approach, continuous psychophysics, promises to transform the field by abandoning the rigid trial structure involving binary responses and replacing it with continuous behavioral adjustments to dynamic stimuli. However, what has precluded wide adoption of this approach is that current analysis methods do not account for the additional variability introduced by the motor component of the task and therefore recover perceptual thresholds that are larger compared to equivalent traditional psychophysical experiments. Here, we introduce a computational analysis framework for continuous psychophysics based on Bayesian inverse optimal control. We show via simulations and previously published data that this not only recovers the perceptual thresholds but additionally estimates subjects’ action variability, internal behavioral costs, and subjective beliefs about the experimental stimulus dynamics. Taken together, we provide further evidence for the importance of including acting uncertainties, subjective beliefs, and, crucially, the intrinsic costs of behavior, even in experiments seemingly only investigating perception.

More than 50 years ago, Bongard introduced 100 visual concept learning problems as a challenge for artificial vision systems. These problems are now known as Bongard problems. Although they are well known in cognitive science and artificial intelligence, only very little progress has been made toward building systems that can solve a substantial subset of them. In the system presented here, visual features are extracted through image processing and then translated into a symbolic visual vocabulary. We introduce a formal language that allows representing compositional visual concepts based on this vocabulary. Using this language and Bayesian inference, concepts can be induced from the examples that are provided in each problem. We find a reasonable agreement between the concepts with high posterior probability and the solutions formulated by Bongard himself for a subset of 35 problems. While this approach is far from solving Bongard problems like humans, it does considerably better than previous approaches. We discuss the issues we encountered while developing this system and their continuing relevance for understanding visual cognition. For instance, contrary to other concept learning problems, the examples are not random in Bongard problems; instead they are carefully chosen to ensure that the concept can be induced, and we found it helpful to take the resulting pragmatic constraints into account. [ABSTRACT FROM AUTHOR]

Coronavirus disease 2019 (COVID-19) has placed an unprecedented strain on healthcare systems worldwide, but while high-income countries (HICs) have been able to adapt, low- and middle-income countries (LMICs) have been much slower to do so due to a lack of funding, skilled healthcare providers, equipment, and facilities. The redistribution of resources to combat the pandemic in LMICs has resulted in decreased surgical volumes at local surgical centers as well as a dramatic reduction in the number of humanitarian aid missions. Despite recent global investment in improving the surgical capacities of LMICs, even in the pre-COVID-19 era there was a vast unmet surgical need. This deficit in surgical capacity has grown during the pandemic and it will be a significant struggle to overcome the resulting backlog of patients. A topic of particular concern to the authors is the effect that the pandemic will have on the delivery of time-sensitive surgical care to patients with cleft palate deformities as delay in providing care can have enormous physical and psychosocial consequences. This paper draws increased attention to the lasting impact that the COVID-19 pandemic may have on cleft palate patients in LMICs. SSRN Pre-print server link: 'https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3898055'

We present first results on finite system size corrections to the equation of state, trace anomaly, and speed of sound for two model systems: 1) free, massless scalar theory and 2) quenched QCD with periodic boundary conditions (PBC). We further present work-in-progress results for quenched QCD with Dirichlet boundary conditions.

Abstract In this paper we study the interactions among a static quark-antiquark pair in the presence of dense two-color quark matter with lattice simulation. To this end we compute Polyakov line correlation functions and determine the renormalized color averaged, color singlet and color triplet grand potentials. The color singlet grand potential allows us to elucidate the number of quarks induced by a static quark antiquark source, as well as the internal energy of such a pair in dense quark matter. We furthermore determine the screening length, which in the confinement phase is synonymous with the string breaking distance. The screening length is a decreasing function of baryon density, due to the possibility to break the interquark string via a scalar diquark condensate at high density. We also study the large distance properties of the color singlet grand potential i a dense medium and find that it is well described by a simple Debye screening formula, parameterized by a Debye mass and an effective coupling constant. The latter is of order of unity, i.e. even at large density two-color quark matter is a strongly correlated system.

Abstract The computation of the parton distribution functions (PDF) or distribution amplitudes (DA) of hadrons from first principles lattice QCD constitutes a central open problem in high energy nuclear physics. In this study, we present and evaluate the efficiency of several numerical methods, well established in the study of inverse problems, to reconstruct the full x-dependence of PDFs. Our starting point are the so called Ioffe time PDFs, which are accessible from Euclidean time simulations in conjunction with a matching procedure. Using realistic mock data tests, we find that the ill-posed incomplete Fourier transform underlying the reconstruction requires careful regularization, for which both the Bayesian approach as well as neural networks are efficient and flexible choices.