搜索结果：AANA

Originally, an exotic Diophantine triple is a set $\{a,b,c\}$ of distinct nonzero rational numbers for which \[ a+1,\quad b+1,\quad c+1,\quad ab+1,\quad ac+1,\quad bc+1,\quad abc+1 \] are all perfect squares. We prove that there is no such triple in $\mathbb{R}[X]$, with at least one nonconstant element, if none of $a,b,c$ is equal to $1$. Equivalently, under the distinct nonzero convention, every exotic Diophantine triple in $\mathbb{R}[X]$ with a nonconstant element must contain the element $1$.

Adeno-associated viral (AAV) vectors are widely used delivery platforms in gene therapy, and the design of improved capsids is key to expanding their therapeutic potential. A central challenge in AAV bioengineering, as in protein design more broadly, is the vast sequence design space relative to the scale of feasible experimental screening. Machine-guided generative approaches provide a powerful means of navigating this landscape and proposing novel protein sequences that satisfy functional constraints. Here, we develop a generative design framework based on protein language models and reinforcement learning to generate highly novel yet functionally plausible AAV capsids. A pretrained model was fine-tuned on experimentally validated capsid sequences to learn patterns associated with viability. Reinforcement learning was then used to guide sequence generation, with a reward function that jointly promoted predicted viability and sequence novelty, thereby enabling exploration beyond regions represented in the training data. Comparative analyses showed that fine-tuning alone produces sequences with high predicted viability but remains biased toward the training distribution, whereas re

In analogy with bilinear Riesz potentials, we introduce bilinear Bessel potentials and characterize their boundedness from $L^p\times L^q$ into Lebesgue and Lorentz spaces $L^{r,α}.$ In several cases we identify the optimal Lorentz indices by constructing explicit counterexamples.

We construct three nonconstant polynomials $a,b,c\in \mathbb{R}[X]$ such that \[ab+1,\qquad ac+1,\qquad bc+1,\qquad abc+1 \] are all squares in $\mathbb{R}[X]$. The entries in the construction are quadratic, so the construction has the smallest possible positive degrees. By composition with nonconstant polynomials, this gives infinitely many examples over $\mathbb{R}[X]$ with all entries nonconstant.

We develop a convergent scheme to train neural networks involving analytic activation functions based on gradient flows. Convergence properties are guaranteed by Lojasiewicz theory. The main advantage of this approach is its simplicity of implementation. The coefficients of the network are approximated by solving a system of ordinary differential equations. We test the method by constructing residual neural network approximations of solutions of parametric problems. The dependence of the solutions of simple ordinary differential equations on a few parameters is correctly reproduced. The solutions of inverse problems involving wave constraints which depend on a few parameters can be reasonably approximated, even in regions in which the problem is severely ill posed.

We show that non-positively curved $k$-fold triangle groups have finitely many cone types, and hence a regular language of all geodesics. Further, we prove that the language of lexicographically first geodesics is both regular and satisfies the fellow traveller property, giving an automatic structure for this family of groups.

The lack of standardization in seizure forecasting slows progress in the field and limits the clinical translation of forecasting models. In this work, we introduce a Python-based framework aimed at streamlining the development, assessment, and documentation of individualized seizure forecasting algorithms. The framework automates data labeling, cross-validation splitting, forecast post-processing, performance evaluation, and reporting. It supports various forecasting horizons and includes a model card that documents implementation details, training and evaluation settings, and performance metrics. Three different models were implemented as a proof-of-concept. The models leveraged features extracted from time series data and seizure periodicity. Model performance was assessed using time series cross-validation and key deterministic and probabilistic metrics. Implementation of the three models was successful, demonstrating the flexibility of the framework. The results also emphasize the importance of careful model interpretation due to variations in probability scaling, calibration, and subject-specific differences. Although formal usability metrics were not recorded, empirical obse

The search for ``dark galaxies,'' a key prediction of the lambda cold dark matter, has yielded few viable candidates. Recently, FAST J0139+4328 was reported as the first isolated dark galaxy in the nearby universe, based on a neutral hydrogen (HI) detection and a non-detection in the Pan-STARRS1 survey. To verify the nature of this candidate, we obtained deep optical imaging, using the $1.4\,\mathrm{m}$ \textit{Milanković} and $0.6\,\mathrm{m}$ \textit{Nedeljković} telescopes, and spectroscopic follow-up of the field. We report the unambiguous discovery of a low-surface-brightness (LSB) optical counterpart at the location of the HI cloud. Furthermore, the detection of H$α$ emission via the $6\,\mathrm{m}$ Big Telescope Alt-Azimuthal (BTA) confirms that the stellar system lies at a redshift consistent with the HI source, establishing their physical association. Through detailed photometry and employing color-dependent mass-to-light scaling relations, we derive a total stellar mass of $M_\star = (7.2 \pm 3.7) \times 10^6\, M_{\odot}$, about an order of magnitude higher than the previously estimated upper limit. Using the literature HI mass, this implies a gas-to-stellar mass ratio of

Generative Artificial Intelligence (GenAI) is driving significant environmental impacts. The rapid development and deployment of increasingly larger algorithmic models capable of analysing vast amounts of data are contributing to rising carbon emissions, water withdrawal, and waste generation. Generative models often consume substantially more energy than traditional models, with major tech firms increasingly turning to nuclear power to sustain these systems -- an approach that could have profound environmental consequences. This paper introduces seven data ecofeminist principles delineating a pathway for developing technological alternatives of eco-societal transformations within the AI research context. Rooted in data feminism and ecofeminist frameworks, which interrogate about the historical and social construction of epistemologies underlying the hegemonic development of science and technology that disrupt communities and nature, these principles emphasise the integration of social and environmental justice within a critical AI agenda. The paper calls for an urgent reassessment of the GenAI innovation race, advocating for ecofeminist algorithmic and infrastructural projects tha

We study the stable pair theory on toric surfaces and determine the virtual tangent space over the fixed point loci. Further, we present a program to compute the virtual Euler characteristic, illustrated by the case of the projective plane. As an application, conjectures regarding rationality and symmetry are supported by verification of a special case.

We resolve the question of the boundedness of the composition of dyadic paraproducts, first posed by Pott, Reguera, Sawyer, and Wick in~\cite{PotCarSawWic}, by providing necessary and sufficient conditions for their boundedness.

In this paper we offer alternate upper bound for the operator $Π_b^*Π_d$ to the ones present in literature, thus extending the known upper bounds from the $L^2(\mathbb{R})$ setting to $L^p(w)$, for $1<p<\infty,$ and a Muckenhoupt weight $w$. In the $L^2(w)$ setting, we fully characterize the boundedness of the operator.

Let $P$ be a planar $n$-gon with the sidelengths $a_1, \ldots, a_n$ and let us denote by $L=L(P)$ the corresponding planar polygonal linkage. We are concerned with the problem of finding conditions on the sidelengths $a_i$ which guarantee the existence of a bicentric configuration of $L$. Specifically, we present some related results on tangential and chordal polygons and characterize pentagonal linkages having bicentric configurations.

We give new examples of geometric constructions of non-natural non-symplectic involutions of IHS manifolds whose existence is guaranteed by previous results of Bossière-Cattaneo-Nieper-Wiesskirchen-Sarti in arXiv:1410.8387 and Bossiére-Camere-Sarti in arXiv:1402.5154.

We construct an analogue of Whittaker reduction for Poisson actions of a semisimple complex Poisson-Lie group G. The reduction takes place along a class of transversal slices to unipotent orbits in G, which are generalizations of the Steinberg cross-section and are indexed by conjugacy classes in the Weyl group. We give an interpretation of these reductions in the framework of Dirac geometry, and we use this to describe their symplectic leaves.

PLATO (PLAnetary Transits and Oscillations of stars) is ESA's M3 mission designed to detect and characterise extrasolar planets and perform asteroseismic monitoring of a large number of stars. PLATO will detect small planets (down to <2 R_(Earth)) around bright stars (<11 mag), including terrestrial planets in the habitable zone of solar-like stars. With the complement of radial velocity observations from the ground, planets will be characterised for their radius, mass, and age with high accuracy (5 %, 10 %, 10 % for an Earth-Sun combination respectively). PLATO will provide us with a large-scale catalogue of well-characterised small planets up to intermediate orbital periods, relevant for a meaningful comparison to planet formation theories and to better understand planet evolution. It will make possible comparative exoplanetology to place our Solar System planets in a broader context. In parallel, PLATO will study (host) stars using asteroseismology, allowing us to determine the stellar properties with high accuracy, substantially enhancing our knowledge of stellar structure and evolution. The payload instrument consists of 26 cameras with 12cm aperture each. For at least f

The technology of autonomous driving is currently attracting a great deal of interest in both research and industry. In this paper, we present a deep learning dual-model solution that uses two deep neural networks for combined braking and steering in autonomous vehicles. Steering control is achieved by applying the NVIDIA's PilotNet model to predict the steering wheel angle, while braking control relies on the use of MobileNet SSD. Both models rely on a single front-facing camera for image input. The MobileNet SSD model is suitable for devices with constrained resources, whereas PilotNet struggles to operate efficiently on smaller devices with limited resources. To make it suitable for such devices, we modified the PilotNet model using our own original network design and reduced the number of model parameters and its memory footprint by approximately 60%. The inference latency has also been reduced, making the model more suitable to operate on resource-constrained devices. The modified PilotNet model achieves similar loss and accuracy compared to the original PilotNet model. When evaluated in a simulated environment, both autonomous driving systems, one using the modified PilotNet

These are expanded lecture notes from the author's minicourse at the 2022 Poisson Geometry Summer School, which took place at the Centre de Recerca Matematica in Barcelona, Spain. After giving a general introduction to wonderful varieties, and an explicit construction of the wonderful compactification of a semisimple adjoint group, we outline several connections to Poisson geometry and to varieties of Lagrangian subalgebras. This survey is intended to be accessible to readers without an extensive background in algebraic geometry.

In this note we show that the multiplicative Grothendieck-Springer space has a natural quasi-Poisson structure. The associated group-valued moment map is the resolution morphism, and the quasi-Hamiltonian leaves are the connected components of the preimages of Steinberg fibers. This is a multiplicative analogue of the standard Poisson structure on the additive Grothendieck-Springer resolution, and an explicit illustration of a more general procedure of reduction along Dirac realizations which is developed in recent work of the author and Mayrand.

Model explainability has become an important problem in machine learning (ML) due to the increased effect that algorithmic predictions have on humans. Explanations can help users understand not only why ML models make certain predictions, but also how these predictions can be changed. In this thesis, we examine the explainability of ML models from three vantage points: algorithms, users, and pedagogy, and contribute several novel solutions to the explainability problem.