A document where I would keep sentences and phrases from the articles I'm reading:
https://arxiv.org/abs/1707.06347
https://arxiv.org/abs/2010.08895
https://arxiv.org/pdf/2109.03697.pdf
advection
noun
the transfer of heat or matter by the flow of a fluid, especially horizontally in the atmosphere or the sea.
Example: Super-resolution (SR) reconstruction of an advection-diffusion process has a direct relevance to many
important applications in atmospheric and environmental problems.
Wang, Chulin, Kyongmin Yeo, Xiao Jin, Andres Codas, Levente J. Klein, and Bruce Elmegreen. "S3RP: Self-Supervised Super-Resolution and Prediction for Advection-Diffusion Process." arXiv preprint arXiv:2111.04639 (2021).
ru: перенос
assimilation
noun
the science of combining different sources of information to estimate possible states of a system as it evolves in time.
Example: More accurate model physics means that forecasts compare better with observations and facilitate improved data assimilation; in turn this permits ingestion of more observations thereby further improving forecasts.
[1] P. Bauer, A. Thorpe, and G. Brunet, “The quiet revolution of numerical weather prediction,” Nature, vol. 525, no. 7567, pp. 47–55, 2015.
ru: ассимиляция
augment
verb
make (something) greater by adding to it; increase.
Example: This objective can further be AUGMENTED by adding an entropy bonus to ensure sufficient exploration, as suggested in past work.
Schulman, John, et al. "Proximal policy optimization algorithms." arXiv preprint arXiv:1707.06347 (2017).
ru: увеличить, дополнить, добавить
azimuthal
adjective
The horizontal angular distance from a reference direction, usually the northern point of the horizon, to the point where a vertical circle through a celestial body intersects the horizon, usually measured clockwise.
Example: To acquire a rain map, the two radars perform four azimuthal scans of 360◦ around a vertical axis beam elevation angles of 0.3◦, 1.1◦, 2.0◦, and 3.0◦
[1] K. Trebing, T. Stanczyk, and S. Mehrkanoon, “Smaat-unet: Precipitation nowcasting using a small attention-unet architecture,” ´ Pattern
Recognition Letters, vol. 145, pp. 178–186, 2021.
ru: азимутальный
bias
noun
a systematic distortion of a statistical result due to a factor not allowed for in its derivation.
Example: The bias towards blurriness is due to the fact that, given the many possibilities for future frames based on the input sequence, the model is trying to keep the error low by predicting a value that is closest to all possible outcomes.
[1] K. Trebing, T. Stanczyk, and S. Mehrkanoon, “Smaat-unet: Precipitation nowcasting using a small attention-unet architecture,” ´ Pattern Recognition Letters, vol. 145, pp. 178–186, 2021.
ru: смещение
bicubic
adjective
Of or pertaining to interpolation in two dimensions using cubic splines or other polynomials (technique for sharpening enlargements of digital images).
Example: We use bicubic upsampling to generate SR images as the baseline [1].
[1] Wang, Chulin, Kyongmin Yeo, Xiao Jin, Andres Codas, Levente J. Klein, and Bruce Elmegreen. "S3RP: Self-Supervised Super-Resolution and Prediction for Advection-Diffusion Process." arXiv preprint arXiv:2111.04639 (2021).
ru: бикубический
constituent
noun
a component part of something.
Example: Atmospheric constituents such as trace gases and aerosols directly affect radiative heating, but aerosols can also act as condensation nuclei in cloud formation and heterogeneous chemistry occurs at the surface of polar stratospheric clouds, accelerating ozone destruction.
[1] P. Bauer, A. Thorpe, and G. Brunet, “The quiet revolution of numerical weather prediction,” Nature, vol. 525, no. 7567, pp. 47–55, 2015.
ru: составляющая
convection
noun
process by which heat is transferred by movement of a heated fluid such as air or water.
Example: On the other hand, deep convection and specific boundary layer processes require a higher degree of parametric formulation as they only occur in small fractions of the grid scale; consequently these parameterizations critically depend on which resolution is actually used.
[1] P. Bauer, A. Thorpe, and G. Brunet, “The quiet revolution of numerical weather prediction,” Nature, vol. 525, no. 7567, pp. 47–55, 2015.
ru: конвекция
discrete
adjective
individually separate and distinct.
Example: Notice that since we assume κ is periodic, it admits a Fourier series expansion, so we may work with the discrete modes
[1] Li, Zongyi, et al. "Fourier neural operator for parametric partial differential equations." arXiv preprint arXiv:2010.08895 (2020).
ru: дискретный
divergence
noun
the process or state of diverging.
Example: ... the Earth Mover (EM) distance behaves in comparison to popular probability distances and divergences used in the context of learning distributions.
Arjovsky, Martin, Soumith Chintala, and Léon Bottou. "Wasserstein generative adversarial networks." International conference on machine learning. PMLR, 2017.
ru: расхождение
enforce
verb
compel observance of or compliance with (a law, rule, or obligation).
Example: Weight clipping is a clearly terrible way to enforce a Lipschitz constraint.
Arjovsky, Martin, Soumith Chintala, and Léon Bottou. "Wasserstein generative adversarial networks." International conference on machine learning. PMLR, 2017.
ru: обеспечить соблюдение
ensemble
noun
a group of items viewed as a whole rather than individually.
Example: Determining these perturbations consistently and seamlessly so that the ensemble provides a good estimate of uncertainty across a wide range of prediction scales is challenging, and the input of mathematics and statistical physics expertise was crucially important.
[1] P. Bauer, A. Thorpe, and G. Brunet, “The quiet revolution of numerical weather prediction,” Nature, vol. 525, no. 7567, pp. 47–55, 2015.
ru: ансамбль
facilitate
verb
make (an action or process) easy or easier.
Example: This facilitates research labs and individuals that do not possess a lot of computing power to also work on these computationally intensive calculations.
[1] K. Trebing, T. Stanczyk, and S. Mehrkanoon, “Smaat-unet: Precipitation nowcasting using a small attention-unet architecture,” ´ Pattern Recognition Letters, vol. 145, pp. 178–186, 2021.
ru: содействовать
flux
noun
the action or process of flowing or flowing out.
Example: The greatest scientific challenge for coupling is associated with matching fluxes at the interfaces where systematic errors in each component interact and can produce model shocks and compensating changes of mean state at every coupling time step and through feedbacks in longer integrations.
[1] P. Bauer, A. Thorpe, and G. Brunet, “The quiet revolution of numerical weather prediction,” Nature, vol. 525, no. 7567, pp. 47–55, 2015.
ru: поток
gauge
verb
estimate or determine the amount, level, or volume of.
Example: Gauging the relative contributions to success and progress from model development, data assimilation algorithms and observational
data usage is difficult because they are interdependent.
[1] P. Bauer, A. Thorpe, and G. Brunet, “The quiet revolution of numerical weather prediction,” Nature, vol. 525, no. 7567, pp. 47–55, 2015.
ru: измерять
hazardous
adjective
risky; dangerous.
Example: Such problems are commonly encountered in Earth and Environmental Sciences, such as
pollution prediction and source identification from satellite images or for autonomous vehicle/robot
navigation under hazardous conditions.
Wang, Chulin, Kyongmin Yeo, Xiao Jin, Andres Codas, Levente J. Klein, and Bruce Elmegreen. "S3RP: Self-Supervised Super-Resolution and Prediction for Advection-Diffusion Process." arXiv preprint arXiv:2111.04639 (2021).
ru: опасный
ill-posed
adjective
Not well-posed. an ill-posed problem.
Example: Generating SR from low resolution data is an ill-posed problem as multiple high resolution solutions may exist corresponding to a low resolution data.
Wang, Chulin, Kyongmin Yeo, Xiao Jin, Andres Codas, Levente J. Klein, and Bruce Elmegreen. "S3RP: Self-Supervised Super-Resolution and Prediction for Advection-Diffusion Process." arXiv preprint arXiv:2111.04639 (2021).
ru: плохо поставленные
likelihood
noun
the state or fact of something's being likely; probability.
Example: This is often done by defining a parametric family of densities and finding the one that maximized the likelihood on our data.
Arjovsky, Martin, Soumith Chintala, and Léon Bottou. "Wasserstein generative adversarial networks." International conference on machine learning. PMLR, 2017.
ru: правдоподобие
manifold
noun
a topological space that locally resembles Euclidean space near each point.
Example: ... we can learn a probability distribution over a low dimensional manifold by doing gradient descent on the EM distance
Arjovsky, Martin, Soumith Chintala, and Léon Bottou. "Wasserstein generative adversarial networks." International conference on machine learning. PMLR, 2017.
ru: многообразие
mapping
noun
an operation that associates each element of a given set (the domain) with one or more elements of a second set (the range).
Example: The classical development of neural networks has primarily focused on learning
mappings between finite-dimensional Euclidean spaces.
[1] Li, Zongyi, et al. "Fourier neural operator for parametric partial differential equations." arXiv preprint arXiv:2010.08895 (2020).
ru: отображение
mesh
noun
an interlaced structure
Example: The neural operator is MESH-invariant, so it can be trained on a lower resolution and evaluated at a
higher resolution, without seeing any higher resolution data (zero-shot super-resolution).
Li, Zongyi, et al. "Fourier neural operator for parametric partial differential equations." arXiv preprint arXiv:2010.08895 (2020).
ru: сетка
nowcasting
noun
Nowcasting is weather forecasting on a very short term mesoscale period of up to 2 hours
Example: Due to their high computational and time requirements, NWP models are less suitable for short-term forecasts ranging from minutes to up to 6 hours, also referred to as nowcasting.
[1] K. Trebing, T. Stanczyk, and S. Mehrkanoon, “Smaat-unet: Precipitation nowcasting using a small attention-unet architecture,” ´ Pattern Recognition Letters, vol. 145, pp. 178–186, 2021.
ru: прогнозирование (краткосрочное)
overfitting
noun
the production of an analysis that corresponds too closely or exactly to a particular set of data, and may therefore fail to fit to additional data or predict future observations reliably
Example: Having fewer parameters in a network reduces the chance of possible overfitting, because the model is simpler and can’t adapt too closely to the training set’s distribution.
[1] K. Trebing, T. Stanczyk, and S. Mehrkanoon, “Smaat-unet: Precipitation nowcasting using a small attention-unet architecture,” ´ Pattern Recognition Letters, vol. 145, pp. 178–186, 2021.
ru: переобучение (подгон)
perturbation
noun
a deviation of a system, moving object, or process from its regular or normal state or path, caused by an outside influence.
Example: Early in the twentieth century, Poincare´21 recognized that forecasts of nonlinear systems can vastly differ if small perturbations are applied to the initial conditions, and that this difficulty could be fundamental in limiting predictive skill.
[1] P. Bauer, A. Thorpe, and G. Brunet, “The quiet revolution of numerical weather prediction,” Nature, vol. 525, no. 7567, pp. 47–55, 2015.
ru: возмущение
porous
adjective
having minute interstices through which liquid or air may pass.
Example: Numerical simulation of multiphase flow in porous media is essential for
many geoscience applications.
Wen, Gege, et al. "U-FNO—An enhanced Fourier neural operator-based deep-learning model for multiphase flow." Advances in Water Resources 163 (2022): 104180.
ru: пористый
precipitation
noun
water that falls from the clouds towards the ground, especially as rain or snow: Hail and sleet are types of precipitation.
Example: The input data for our model consists of precipitation maps, i.e. cartographic radar images showing the accumulated rainfall over a period of time. In addition, the applicability of the proposed model is also shown on cloud cover nowcasting task.
[1] K. Trebing, T. Stanczyk, and S. Mehrkanoon, “Smaat-unet: Precipitation nowcasting using a small attention-unet architecture,” ´ Pattern Recognition Letters, vol. 145, pp. 178–186, 2021.
ru: осадки
prognostic
adjective
of, relating to, or serving as ground for prognostication or a prognosis
Example: Consequently, various numerical methods of solution emerged that addressed numerical stability, accuracy, computational speed and versatility to deal with more prognostic variables, and the interaction between resolved and unresolved scales. [1] P. Bauer, A. Thorpe, and G. Brunet, “The quiet revolution of numerical weather prediction,” Nature, vol. 525, no. 7567, pp. 47–55, 2015.
ru: прогностический
satellite
noun
an artificial body placed in orbit round the earth or moon or another planet in order to collect information or for communication.
Example: SR and prediction are two of the most important aspects of down scaling climate/weather modelling and satellite observations.
Wang, Chulin, Kyongmin Yeo, Xiao Jin, Andres Codas, Levente J. Klein, and Bruce Elmegreen. "S3RP: Self-Supervised Super-Resolution and Prediction for Advection-Diffusion Process." arXiv preprint arXiv:2111.04639 (2021).
ru: спутник
saturation
noun
the degree or extent to which something is dissolved or absorbed compared with the maximum possible, usually expressed as a percentage.
Example: For gas saturation, we do not normalize the data because the saturation values always range from 0 to 1.
Wen, Gege, et al. "U-FNO—An enhanced Fourier neural operator-based deep-learning model for multiphase flow." Advances in Water Resources 163 (2022): 104180.
ru: насыщенность
super-resolution
noun
is a class of techniques that enhance (increase) the resolution of an imaging system.
Example: This allows us to achieve zero-shot super-resolution as shown in Section 5. [1]
[1] Li, Zongyi, et al. "Fourier neural operator for parametric partial differential equations." arXiv preprint arXiv:2010.08895 (2020).
ru: супер-разрешение
truncate
verb
shorten the duration or extent of.
Example: We pick a finite-dimensional parameterization by truncating the Fourier series at
a maximal number of modes [1].
[1] Li, Zongyi, et al. "Fourier neural operator for parametric partial differential equations." arXiv preprint arXiv:2010.08895 (2020).
ru: обрезать
unsupervised
adjective
not done or acting under supervision.
Example: The problem this paper is concerned with is that of unsupervised learning.
Arjovsky, Martin, Soumith Chintala, and Léon Bottou. "Wasserstein generative adversarial networks." International conference on machine learning. PMLR, 2017.
ru: без учителя
versatility
noun
ability to adapt or be adapted to many different functions or activities.
Example: Consequently, various numerical methods of solution emerged that addressed numerical stability, accuracy, computational
speed and versatility to deal with more prognostic variables, and the interaction between resolved and unresolved scales.
[1] P. Bauer, A. Thorpe, and G. Brunet, “The quiet revolution of numerical weather prediction,” Nature, vol. 525, no. 7567, pp. 47–55,
2015.
ru: универсальность
viscous
adjective
having a thick, sticky consistency between solid and liquid; having a high viscosity.
Example: The 1-d Burgers’ equation is a non-linear PDE with various applications including modeling the one
dimensional flow of a viscous fluid.
[1] Li, Zongyi, et al. "Fourier neural operator for parametric partial differential equations." arXiv preprint arXiv:2010.08895 (2020).
ru: вязкий
vorticity
noun
a pseudovector field that describes the local spinning motion of a continuum near some point (the tendency of something to rotate), as would be seen by an observer located at that point and traveling along with the flow
Example: We consider the 2-d Navier-Stokes equation for a viscous, incompressible fluid in vorticity form on
the unit torus.
Li, Zongyi, et al. "Fourier neural operator for parametric partial differential equations." arXiv preprint arXiv:2010.08895 (2020).
