623509 research outputs found
Sort by
BarcodeMamba: State Space Models for Biodiversity Analysis
No full textDNA barcodes are crucial in biodiversity analysis for building automatic identification systems that recognize known species and discover unseen species. Unlike human genome modeling, barcode-based invertebrate identification poses challenges in the vast diversity of species and taxonomic complexity. Among Transformer-based foundation models, BarcodeBERT excelled in species-level identification of invertebrates, highlighting the effectiveness of self-supervised pretraining on barcode-specific datasets. Recently, structured state space models (SSMs) have emerged, with a time complexity that scales sub-quadratically with the context length. SSMs provide an efficient parameterization of sequence modeling relative to attention-based architectures. Given the success of Mamba and Mamba-2 in natural language, we designed BarcodeMamba, a performant and efficient foundation model for DNA barcodes in biodiversity analysis. We conducted a comprehensive ablation study on the impacts of self-supervised training and tokenization methods, and compared both versions of Mamba layers in terms of expressiveness and their capacity to identify unseen species held back from training. Our study shows that BarcodeMamba has better performance than BarcodeBERT even when using only 8.3% as many parameters, and improves accuracy to 99.2% on species-level accuracy in linear probing without fine-tuning for seen species. In our scaling study, BarcodeMamba with 63.6% of BarcodeBERT\u27s parameters achieved 70.2% genus-level accuracy in 1-nearest neighbor (1-NN) probing for unseen species. The code repository to reproduce our experiments is available at https://github.com/bioscan-ml/BarcodeMamba.9 pages, 2 figures, accepted at Foundation Models for Science: Progress, Opportunities, and Challenges Workshop (NeurIPS 2024
Decays and accounting for the contribution of
No full textIn the quark NJL model, lepton decays with the production of scalar mesons and neutrinos are studied, where . It is shown that these decays mainly occur via contact channels and channels with axial-vector mesons , and . All mesons are considered as quark-antiquark states in this case. The obtained results can be considered as predictions for future experiments. The obtained estimates for the branching fractions of the decay taking into account the contributions of the and states are in satisfactory agreement with experimental data.7 page
Quasinormal mode as a foundational framework for all electromagnetic Fano resonances
No full textFano profiles are observed across various fields of wave physics. They emerge from interference phenomena and are quantified by the asymmetry parameter q. In optics, q is usually considered as a phenomenological coefficient obtained by fitting experimental or numerical data. In this work, we introduce an ab initio Maxwellian approach using quasinormal modes to analytically describe line shapes in light scattering problems. We show that the response of each individual quasinormal mode inherently exhibits a Fano profile and derive an explicit analytical formula for the Fano parameter. Experimental and numerical validations confirm the formula\u27s accuracy across a broad spectrum of electromagnetic systems. The general expression for q opens new possibilities for fine-tuning and optimizing spectral line shapes in electromagnetism
Thermodynamics, structure and dynamics of cylindrically confined hard spheres: The role of excess helical twist
No full textHard spheres confined to narrow quasi-one-dimensional cylindrical channels form perfect helical structures at close packing. Here, we use molecular dynamics simulation to show that the thermodynamics, structure and dynamics of the fluid below close packing are dominated by the presence of topological defects that reverse the local twist direction of the helix. When compressed from a random, low density state, or decompressed from high density ordered states with zero excess helical twist, the system equilibrates to an achiral fluid that exhibits two heat capacity maxima along the equation of state. The low density heat capacity maximum corresponds to the onset of helix formation and the high density maximum occurs when the system rapidly loses defects in a Schottky-like anomaly. The local twist auto-correlation function in the achiral fluid exhibits a stretched exponential decay and the structural relaxation times undergo a fragile-to-strong crossover located at the high density heat capacity maximum. We also study the effect of excess helical twist by using initial starting configurations consisting of two helical domains with opposite twist directions of different lengths. This leads to the formation of topologically protected states that are characterized by the presence of loosely bound defect pairs which become more tightly bound with increasing excess helical twist. The local twist auto-correlation function in the chiral fluid decays as a power law at long times. The possible kinetic or thermodynamic origin of this topological protection is discussed.New results, extended discussion. 13 figures, 16 page
Single vibronic level fluorescence spectra from Hagedorn wavepacket dynamics
No full textIn single vibronic level (SVL) fluorescence experiments, the electronically excited initial state is also excited in one or several vibrational modes. Whereas computing all contributing Franck-Condon factors individually becomes impractical in large systems, a time-dependent formalism has not been applied to simulate emission from arbitrary initial vibrational levels. Here, we use Hagedorn functions, which are products of a Gaussian and carefully generated polynomials, to represent SVL initial states. In systems where the potential is at most quadratic, the Hagedorn functions are exact solutions to the time-dependent Schrödinger equation and can be propagated with the same equations of motion as a simple Gaussian wavepacket. Having developed an efficient recursive algorithm to compute the overlaps between two Hagedorn wavepackets, we can now evaluate emission spectra from arbitrary vibronic levels using a single trajectory. We validate the method in two-dimensional global harmonic models by comparing it with quantum split-operator calculations. Additionally, we study the effects of displacement, distortion (squeezing), and Duschinsky rotation on SVL spectra. Finally, we demonstrate the applicability of the Hagedorn approach to high-dimensional systems on an example of displaced, distorted, and Duschinsky-rotated harmonic model with 100 degrees of freedom.9 pages, 5 figure
Estimating the history of a random recursive tree
No full textThis paper studies the problem of estimating the order of arrival of the vertices in a random recursive tree. Specifically, we study two fundamental models: the uniform attachment model and the linear preferential attachment model. We propose an order estimator based on the Jordan centrality measure and define a family of risk measures to quantify the quality of the ordering procedure. Moreover, we establish a minimax lower bound for this problem, and prove that the proposed estimator is nearly optimal. Finally, we numerically demonstrate that the proposed estimator outperforms degree-based and spectral ordering procedures
Hybrid weakly over-penalised symmetric interior penalty method on anisotropic meshes
No full textIn this study, we investigate a hybrid-type anisotropic weakly over-penalised symmetric interior penalty method for the Poisson equation on convex domains. Compared with the well-known hybrid discontinuous Galerkin methods, our approach is simple and easy to implement. Our primary contributions are the proposal of a new scheme and the demonstration of a proof for the consistency term, which allows us to estimate the anisotropic consistency error. The key idea of the proof is to apply the relation between the Raviart--Thomas finite element space and a discontinuous space. In numerical experiments, we compare the calculation results for standard and anisotropic mesh partitions.30 page
COBRA -- COnfidence score Based on shape Regression Analysis for method-independent quality assessment of object pose estimation from single images
No full textWe propose a generic procedure for assessing 6D object pose estimates. Our approach relies on the evaluation of discrepancies in the geometry of the observed object, in particular its respective estimated back-projection in 3D, against a putative functional shape representation comprising mixtures of Gaussian Processes, that act as a template. Each Gaussian Process is trained to yield a fragment of the object\u27s surface in a radial fashion with respect to designated reference points. We further define a pose confidence measure as the average probability of pixel back-projections in the Gaussian mixture. The goal of our experiments is two-fold. a) We demonstrate that our functional representation is sufficiently accurate as a shape template on which the probability of back-projected object points can be evaluated, and, b) we show that the resulting confidence scores based on these probabilities are indeed a consistent quality measure of pose
Investigating Self-Supervised Image Denoising with Denaturation
No full textSelf-supervised learning for image denoising problems in the presence of denaturation for noisy data is a crucial approach in machine learning. However, theoretical understanding of the performance of the approach that uses denatured data is lacking. To provide better understanding of the approach, in this paper, we analyze a self-supervised denoising algorithm that uses denatured data in depth through theoretical analysis and numerical experiments. Through the theoretical analysis, we discuss that the algorithm finds desired solutions to the optimization problem with the population risk, while the guarantee for the empirical risk depends on the hardness of the denoising task in terms of denaturation levels. We also conduct several experiments to investigate the performance of an extended algorithm in practice. The results indicate that the algorithm training with denatured images works, and the empirical performance aligns with the theoretical results. These results suggest several insights for further improvement of self-supervised image denoising that uses denatured data in future directions
Effect of Torsion on Neutron Star Structure in Einstein-Cartan Gravity
No full textEinstein-Cartan gravity is a close historical sibling of general relativity that allows for spacetime torsion. As a result, angular momentum couples to spacetime geometry in a similar way to energy. While consequences of this are well studied on cosmological scales, their role in neutron star physics is largely under-explored. We study the effects that torsion, sourced by either microphysical spin or macroscopic angular momentum, has on neutron stars. For this, we use a simplified polytropic model to quantify the microphysical coupling to torsion. We also derive expressions to model rotation-induced torsion effects and estimate the consequences for rotating neutron stars with different rotation rates. We find that the presence of torsion in general leads to neutron stars with smaller radii and masses, but higher central densities. Realistic models for microphysical spin lead to torsion effects that have no relevant influence on the neutron star structure. Rotation-induced torsion effects however, can decrease the radius by up to , which is comparable to the increase due to centrifugal forces. Depending on which effect dominates, this leads to a torsion-induced spin-up or spin-down of the neutron star. We conclude that torsion effects due to rotation can not be neglected and are large enough to be tested using current or near-future technology.17+1 pages, 4 figures. v2: updated to match published versio