Browse Articles
Discover research articles across all indexed journals
Fast Interpretable Greedy-Tree Sums
Modern machine learning has achieved impressive prediction performance, but often sacrifices interpretability, a critical consideration in high-stakes domains such as medicine. In such settings, practitioners often use highly interpretable decision tree models, but these suffer from inductive bias against additive structure. To overcome this bias, we propose Fast Interpretable Greedy-Tree Sums (FIGS), which generalizes the Classification and Regression Trees (CART) algorithm to simultaneously grow a flexible number of trees in summation. By combining logical rules with addition, FIGS adapts to additive structure while remaining highly interpretable. Experiments on real-world datasets show FIGS achieves state-of-the-art prediction performance. To demonstrate the usefulness of FIGS in high-stakes domains, we adapt FIGS to learn clinical decision instruments (CDIs), which are tools for guiding decision-making. Specifically, we introduce a variant of FIGS known as Group Probability-Weighted Tree Sums (G-FIGS) that accounts for heterogeneity in medical data. G-FIGS derives CDIs that reflect domain knowledge and enjoy improved specificity (by up to 20% over CART) without sacrificing sensitivity or interpretability. Theoretically, we prove that FIGS learns components of additive models, a property we refer to as disentanglement. Further, we show (under oracle conditions) that tree-sum models leverage disentanglement to generalize more efficiently than single tree models when fitted to additive regression functions. Finally, to avoid overfitting with an unconstrained number of splits, we develop Bagging-FIGS, an ensemble version of FIGS that borrows the variance reduction techniques of random forests. Bagging-FIGS performs competitively with random forests and XGBoost on real-world datasets.
Mechanical properties and biodegradability of samples obtained by 3D printing using FDM technology from PLA filament with by-products
Bioinspired energy-free temperature gradient regulator for significant enhancement of thermoelectric conversion efficiency
Enhancing thermoelectric conversion efficiency (TCE) is pivotal for advancing global energy conservation and emission reduction initiatives. Traditional approaches primarily focus on microscopic strategies such as bandgap engineering, chemical potential adjustments, and entropy engineering. However, these methods face substantial limitations in practical applications due to challenging material property requirements. Additionally, the efficiencies achieved remain modest, constrained by the interdependent nature of electrical and thermal conductivities, which typically vary concurrently. Inspired by the thermoregulation mechanisms in biological organisms, we propose here a macroscopic strategy based on an expanded-plane (EP) meta-structure. Such a meta-structure, of the thermal gradient regulator type, exploits the temperature gradient concentrating effect to significantly boost TCE. We prove, both numerically and experimentally, how the proposed device can achieve temperature gradient concentration under diverse conditions. Remarkably, under undistorted temperature background conditions, we measured an striking efficiency enhancement of 59.0 % . This work highlights the EP thermal gradient regulator’s ability to boost thermoelectric efficiency, valuable across domains: aiding efficient chip cooling in microelectronics and powering wearable medical devices.
A web-based artificial intelligence system for label-free virus classification and detection of cytopathic effects
Double nanowire quantum dots and machine learning
Abstract We present an approach to estimate the single-particle energies in double InAs/InP nanowire quantum dots by combining an atomistic tight-binding approach with machine learning. The method works particularly well with a neural network and transfer learning, where we can accurately recover ground state energies with root-mean-square deviation around 1 meV by using only a small training set and capitalizing on earlier, smaller-scale computations. The training set is only a fraction of the multidimensional search space of possible dot sizes and inter-dot spacings. Besides the cases presented in this work, we expect this technique will interest other researchers involved in solving the inverse computational problem of matching spectra to nanostructure morphological properties.
Turbulent mixing controls fixation of growing antagonistic populations
Unlike coffee and cream that homogenize when stirred, growing micro-organisms (e.g., bacteria, baker’s yeast) can actively kill each other and avoid mixing. How do such antagonistic interactions impact the growth and survival of competing strains, while being spatially advected by turbulent flows? By using numerical simulations of a continuum model, we study the dynamics of two antagonistic strains that are dispersed by incompressible turbulent flows in two spatial dimensions. A key parameter is the ratio of the fluid transport time to that of biological reproduction, which determines the winning organism that ultimately takes over the whole population from an initial heterogeneous state, a process known as fixation. By quantifying the probability and mean time for fixation along with the spatial structure of concentration fluctuations, we demonstrate how turbulence raises the threshold for biological nucleation and antagonism suppresses flow-induced mixing by depleting the population at interfaces. Our work highlights the unusual biological consequences of the interplay of turbulent fluid flows with antagonistic population dynamics, with potential implications for marine microbial ecology and origins of biological chirality.
Topological transition on a conformal manifold for the quantum Ising model with a longer range interaction
A preliminary study on dynamic response of cold-region tunnel systems considering the frozen soil in plastic zone
Variable-strength nonlocal measurements reveal quantum violations of classical counting principles
We implement a variant of the quantum pigeonhole paradox thought experiment to study whether classical counting principles survive in the quantum domain. We observe strong measurements significantly violate the pigeonhole principle (that among three pigeons in two holes, at least one pair must be in the same hole) and the sum rule (that the number of pigeon pairs in the same hole is the sum of the number of pairs across each of the holes) in an ensemble that is pre- and postselected into particular separable states. To investigate whether measurement disturbance is a viable explanation for these counterintuitive phenomena, we employ a we employ variable-strength nonlocal measurements. As we decrease the measurement strength, we find the violation of the sum rule decreases, yet the pigeonhole principle remains violated. In the weak limit, the sum rule is restored due to the cancellation between two weak values with equal and opposite imaginary parts. We observe the same kind of cancellation at higher measurement strengths, thus raising the question: do strong measurements have imaginary parts?
Towards full integration of explainable artificial intelligence in colon capsule endoscopy’s pathway
Integration of 101 machine learning algorithm combinations to unveil m6A/m1A/m5C/m7G-associated prognostic signature in colorectal cancer
Self-assembled proteomimetic (SAP) with antibody-like binding from short PNA–peptide conjugates
Affinity proteins based on a three-helix bundle (affibodies, alphabodies, and computationally de novo designed ones) have been shown to be a general platform to discover binders with properties reminiscent of antibodies, combining high target specificity with affinities reaching well below the nanomolar. Herein, we report a strategy, coined self-assembled proteomimetic (SAP), to mimic such three-helix bundle architecture with a hybridization-enforced two-helix coiled coil that is obtained by templated native chemical ligation (T-NCL) of PNA–peptide conjugates. This SAP strategy stands out by its synthetic accessibility, reducing the length on the longest synthetic peptide to less than 30 amino acids which is readily attainable by standard SPPS methodologies. We show that the T-NCL dramatically accelerates the ligation, enabling this chemistry to proceed in a combinatorial fashion at low micromolar concentrations. We demonstrate that small combinatorial libraries of SAPs can be prepared in one operation and used directly in affinity selections against a target of interest with an LC–MS analysis of the fittest binders. Moreover, we show that the underlying design paradigm is functional for SAPs based on structurally distinct three-helix peptides aimed at different therapeutic targets, namely HER2 and spike’s RBD, reaching picomolar affinities. We further illustrate that the affinity of the SAP can be allosterically regulated using a toehold displacement of the hybridizing PNAs to disrupt the coiled coil stabilization. Finally, we show that an RBD-targeting SAP effectively inhibits viral entry of SARS-CoV-2 with an IC 50 of 2.8 nM.
Synergistic effect of nanosilver fluoride with L-arginine on remineralization of early carious lesions
Molecular characterization of Streptococcus agalactiae strains isolated from pregnant women
A method for unsupervised learning of coherent spatiotemporal patterns in multiscale data
The unsupervised and principled diagnosis of multiscale data is a fundamental obstacle in modern scientific problems from, for instance, weather and climate prediction, neurology, epidemiology, and turbulence. Multiscale data are characterized by a combination of processes acting along multiple dimensions simultaneously, spatiotemporal scales across orders of magnitude, nonstationarity, and/or invariances such as translation and rotation. Existing methods are not well-suited to multiscale data, usually requiring supervised strategies such as human intervention, extensive tuning, or selection of ideal time periods. We present the multiresolution coherent spatio-temporal scale separation (mrCOSTS), a hierarchical and automated algorithm for the diagnosis of coherent patterns or modes in multiscale data. mrCOSTS is a variant of dynamic mode decomposition which decomposes data into bands of spatial patterns with shared time dynamics, thereby providing a robust method for analyzing multiscale data. It requires no training but instead takes advantage of the hierarchical nature of multiscale systems. We demonstrate mrCOSTS using complex multiscale datasets that are canonically difficult to analyze: 1) climate patterns of sea surface temperature, 2) electrophysiological observations of neural signals of the motor cortex, and 3) horizontal wind in the mountain boundary layer. With mrCOSTS, we trivially retrieve complex dynamics that were previously difficult to resolve while additionally extracting hitherto unknown patterns of activity embedded in the dynamics, allowing for advancing the understanding of these fields of study. This method is an important advancement for addressing the multiscale data which characterize many of the grand challenges in science and engineering.
An improved method of AUD-YOLO for surface damage detection of wind turbine blades
Superfluid weight cross-over and critical temperature enhancement in singular flat bands
Nonanalytic Bloch eigenstates at isolated band degeneracy points exhibit singular behavior in the quantum metric. Here, a description of superfluid weight for zero-energy flat bands in proximity to other high-energy bands is presented, where they together form a singular band gap system. When the singular band gap closes, the geometric and conventional contributions to the superfluid weight as a function of the superconducting gap exhibit different cross-over behaviors. The scaling behavior of superfluid weight with the band gap is studied in detail, and the effect on the Berezinskii–Kosterlitz–Thouless transition temperature is explored. It is found that tuning the singular band gap provides a unique mechanism for enhancing the supercurrent and critical temperature of two-dimensional superconductors.
The impact of 24-forms Tai Chi on alpha band power and physical fitness in young adults: a randomized controlled trial
Detrimental influence of Arginase-1 in infiltrating macrophages on poststroke functional recovery and inflammatory milieu
Poststroke inflammation critically influences functional outcomes following ischemic stroke. Arginase-1 (Arg1) is considered a marker for anti-inflammatory macrophages, associated with the resolution of inflammation and promotion of tissue repair in various pathological conditions. However, its specific role in poststroke recovery remains to be elucidated. This study investigates the functional impact of Arg1 expressed in macrophages on poststroke recovery and inflammatory milieu. We observed a time-dependent increase in Arg1 expression, peaking at 7 d after photothrombotic stroke in mice. Cellular mapping analysis revealed that Arg1 was predominantly expressed in LysM-positive infiltrating macrophages. Using a conditional knockout (cKO) mouse model, we examined the role of Arg1 expressed in infiltrating macrophages. Contrary to its presumed beneficial effects, Arg1 cKO in LysM-positive macrophages significantly improved skilled forelimb motor function recovery after stroke. Mechanistically, Arg1 cKO attenuated fibrotic scar formation, enhanced peri-infarct remyelination, and increased synaptic density while reducing microglial synaptic elimination in the peri-infarct cortex. Gene expression analysis of fluorescence-activated single cell sorting (FACS)-sorted CD45 low microglia revealed decreased transforming growth factor-β (TGF-β) signaling and proinflammatory cytokine activity in peri-infarct microglia from Arg1 cKO animals. In vitro coculture experiments demonstrated that Arg1 activity in macrophages modulates microglial synaptic phagocytosis, providing evidence for macrophage–microglia interaction. These findings present unique insights into the function of Arg1 in central nervous system injury and highlight an interaction between infiltrating macrophages and resident microglia in shaping the poststroke inflammatory milieu. Our study identifies Arg1 in macrophages as a potential therapeutic target for modulating poststroke inflammation and improving functional recovery.