Reading Diary

59 Data Science terms

Accuracy
Noun
Ratio of correctly predicted observation to the total observations
Example: Among classification metrics, Accuracy is a good metric, but it is not enough for all classification problems. [2]
ru: Точность
Acquisition
Noun
To sample signals and convert them into digital values
Example: The acquisition of respiratory sounds was performed on adult and elderly patients. [4]
ru: Сбор
Active learning
phrase
Learning method in which the learner receives a set of unlabeled data (images/videos) for training a visual accurate model, which must be done minimizing the labeling effort by choosing the best training data out of the total amount.
Example: In active learning, the learner receives a set of unlabeled data (images/videos) for training a visual accurate model, which must be done minimizing the labeling effort by choosing the best training data out of the total amount. [1]
ru: Активное обучение
Adaptataion
Noun
The process of adapting something or becoming adapted to a situation; adjustment, modification.
Example: In this work, we also consider to align the prediction layer to improve the domain adaptation performance. [1]
ru: Адаптация
Auto-Tuned Models
phrase
A multi-tenant, multi-data system for automated machine learning
Example: Next, we have Auto-Tuned Models (ATM), which is meant to be a distributed, collaborative, and scalable system for AutoML. [6]
ru: Модели с автоподстройкой
Backpropagation
Noun
Algorithm for supervised learning of artificial neural networks using gradient descent
Example: Forward and backpropagation for that training case are done only on this thinned network. [7]
ru: Обратное распространение
Computer Vision
phrase
Interdisciplinary scientific field that deals with how computers can gain high-level understanding from digital images or videos.
Example: Since supervised (deep) machine learning became the key to solve computer vision tasks, the availability of task ground truth (i.e. supervision information) associated to the raw data (i.e. images and videos) has been a major practical problem. [1]
ru: Компьютерное зрение
Data
Noun
Information, especially in a scientific or computational context, or with the implication that it is organized.
Example: Such varieties in the data made it challenging to identify problems in the respiratory sounds. [4]
ru: Данные
Data mining
phrase
Process of extracting and discovering patterns in large data sets involving methods at the intersection of machine learning, statistics, and database systems.
Example: Data mining is used to extract useful information from large datasets and to display it in easy-to-interpret visualizations. [5]
ru: Добыча данных
Dataset
Noun
A set of data to be analyzed.
Example: Ling-Spam dataset is divided into 10 parts from the ‘bare’ distribution that includes individual emails as a text file (.txt). [3]
ru: Датасет
Decision tree
phrase
Non-parametric supervised learning method used for classification and regression.
Example: Decision Tree is a common supervised learning approach employed for both regression and classification problems. [2]
ru: Дерево решений
Decoder
Noun
Decoder is a device that converts the given data format back to the original format
Example: For semantic segmentation, we simplified the decoder by a single up-sampling layer.
ru: Декодер
Deviation
Noun
For interval variables and ratio variables, a measure of difference between the observed value and the mean.
Example: As audio clip contains high deviations across its entire length, its analysis is not trivial. [4]
ru: Отклонение
Domain adaptation
phrase
Example: In domain adaptation, a model is trained to perform a visual task in a specific domain (e.g. semantic segmentation in synthetic images), however, we need to apply it to perform the same task in a correlated, but significantly different, domain (e.g. semantic segmentation in real-world images); which is done by reusing the previous knowledge (in the form of model or labeled data) for minimizing the labeling effort in the new domain. [1]
ru: Адаптация предметной области
Dropout
Noun
Dropping out units (hidden and visible) in a neural network.
Example: Dropout is a technique that addresses both these issues. It prevents overfitting and provides a way of approximately combining exponentially many different neural network architectures efficiently. [7]
ru: Метод отключения случайных нейронов
Ensemble
Noun
A group of separate things that contribute to a coordinated whole.
Example: The second improvement made by Auto-sklearn was the automated ensemble construction of models evaluated during optimization. [6]
ru: Ансамбль
Features
Noun
An important or main item
Example: For training machine learning models, we implement the following steps: normalizing features (just for continuous data), randomly splitting the main dataset into train data and test data (30% of dataset was assigned to the test part), fitting the models and evaluating them by validation data (and ‘‘early stopping’’) to prevent overfitting, and using metrics for final evaluation with test data. [2]
ru: Особенности
Fine-tuning
noun
Improving forecasting accuracy
Example: In [69], iterative self-labeling and fine-tuning with spatial urban-scene location priors are used to perform the domain adaptation. [1]
ru: Тонкая настройка
Gradient
Noun
Iterative first-order optimisation algorithm used to find a local minimum/maximum of a given function
Example: The gradients for each parameter are averaged over the training cases in each mini-batch. [7]
ru: Градиент
Hidden layers
phrase
Layers in between input layers and output layers
Example: Deep neural networks contain multiple non-linear hidden layers and this makes them very expressive models that can learn very complicated relationships between their inputs and outputs. [7]
ru: Скрытые слои
Hyperparameters
A parameter whose value is set before the learning process begins.
Example: Every machine learning model has two types of parameters: hyperparameters that the model designer must manually set prior to training, and normal parameters that are optimized in the training of the model. [6]
ru: Гиперпараметр
Hypothesize
Verb
To assume or assert tentatively on uncertain grounds.
Example: We hypothesize that for each hidden unit, dropout prevents co-adaptation by making the presence of other hidden units unreliable. [7]
ru: Cтроить гипотезу
Input layer
phrase
The input of the whole CNN.
Example: The input layer takes in some numerical representation of the data. [6]
ru: Входной слой
K-Nearest Neighbour
phrase
Data classification method for estimating the likelihood that a data point will become a member of one group or another based on what group the data points nearest to it belong to.
Example: The paper provides an overview of the ML algorithm such as NB, SVM and KNN classifiers [3]
ru: Метод k-ближайших соседей
Linear Predictive Cepstral Coefficient (LPCC)
phrase
Cepstral coefficients derived from LPC calculated spectral envelope
Example: We have employed Linear Predictive Cepstral Coefficient (LPCC)-based features to characterize respiratory sounds. [4]
ru: Линейный прогнозируемый кепстральный коэффициент
Logistic regression
phrase
Process of modeling the probability of a discrete outcome given an input variable.
Example: Logistic regression is used to assign observations to a separated set of classes as a classifier. [2]
ru: Логистическая регрессия
Logistic regression
phrase
Statistical model that in its basic form uses a logistic function to model a binary dependent variable
Example: Logistic regression is used to assign observations to a separated set of classes as a classifier. [2]
ru: Логистическаая регрессия
Long short-term memory (LSTM)
phrase
LSTM is a special kind of recurrent neural network capable of handling long-term dependencies.
Example: LSTM is a particular type of RNN with an extensive range of uses such as document classification, time series analysis, voice and speech recognition. [2]
ru: Длинная цепь элементов краткосрочной памяти
Machine learning
phrase
The study of computer algorithms that can improve automatically through experience and by the use of data.
Example: Machine learning models have been utilized for multiple purposes in the field of computer science from resolving a network traffic issue to detecting a malware. [1]
ru: Машинное обучение
Model
Noun
Organizes data elements and standardizes how the data elements relate to one another
ru: Модель
Momentum
Noun
Accumulates an exponentially decaying moving average of past gradients and continues to move in their direction.
Example: Next, the momentum was varied from 0.1 to 0.5 with a step of 0.1, and results are provided in Table 4. [4]
ru: Момент
Multilayer Perceptron
phrase
Class of feedforward artificial neural network
Example: With Multilayer Perceptron (MLP)-based classifier, in our experiment, we have achieved the highest possible accuracy of 99.22% (AUC = 0.9993) on a publicly available dataset of size 6800+ clips. [4]
ru: Многослойный перцептрон
Naïve Bayes
phrase
Technique for constructing classifiers: models that assign class labels to problem instances, represented as vectors of feature values
Example: Naïve Bayes model is used to resolve classification problems by using probability techniques. [3]
ru: Наивный байесовский классификатор
Neural Network
Phrase
Network or circuit of neurons, composed of artificial neurons or nodes.
Example: We show that dropout improves the performance of neural networks on supervised learning tasks in vision, speech recognition, document classification and computational biology, obtaining state-of-the-art results on many benchmark data sets. [7]
ru: Нейронная сеть
Neural network architecture
The way the neural network is organized
Example: This has led to an increased interest among AutoML researchers to invest their time in the field of neural architecture search (NAS), which aims to find the best neural network architecture. [6]
ru: Архитектура нейронной сети
Node
Noun
A vertex or a leaf in a graph of a network, or other element in a data structure.
Example: There are three types of nodes. [5]
ru: Узел
Object recognition
phrase
Computer vision technique for identifying objects in images or videos.
Example: For object recognition, we also evaluate on the multiple source domain adaptation dataset PACS [29] dataset, which has 7 object categories and 4 domains (Photo, Art Paintings, Cartoon and Sketches). FIGURE 5 shows sample images from PACS dataset. [1]
ru: Распознавание объектов
Particle swarm optimization (PSO)
phrase
Computational method that optimizes a problem by iteratively trying to improve a candidate solution with regard to a given measure of quality. It solves a problem by having a population of candidate solutions, here dubbed particles, and moving these particles around in the search-space according to simple mathematical formula over the particle's position and velocity.
Example: Two popular examples of these types of methods are particle swarm optimization (PSO) and evolutionary algorithms [46], both of which are inspired by biological behaviors. [6]
ru: Метод роя частиц
Performance
Noun
The amount of useful work accomplished estimated in terms of time needed, resources used.
Example: The performance of several other classifiers was compared in order to establish the efficacy of MLP. [4]
ru: Производительность
Pipeline
Noun
Means of automating the machine learning workflow by enabling data to be transformed and correlated into a model that can then be analyzed to achieve outputs.
Example: The previous two sections only discussed methods that attempt to handle one component of the machine learning pipeline. [6]
ru: Пайплайн
Pipeline optimizer
Set of methods to maximize the rendering speed and allow stages that are not bottlenecks to consume as much time as the bottleneck
Example: Each pipeline optimizer performs one or more tasks in order to help automate the machine learning process. [6]
ru: Оптимизатор пайплайна
Potential input variables
phrase
The main component of decision tree
Example: Another potential problem is that strong correlation between different potential input variables may result in the selection of variables that improve the model statistics but are not causally related to the outcome of interest. [5]
ru: Потенциальные входные переменные
Precision
Noun
The ability of a measurement to be reproduced consistently
Example: The paper used the emails from Ling-Spam corpus and aimed to acquire an improvement in F1-score, Precision, Recall and Accuracy. [3]
ru: Точность
Prediction
Noun
A probability estimation based on statistical methods.
Example: We select image rotation prediction as pretext task due to its simplicity and superior performance on visual representation learning to other proposals. [1]
ru: Прозноз
Preprocessing
Noun
An executable task or program
Example: No preprocessing was performed to remove the heartbeat sounds. [4]
ru: Предварительная обработка
Pretext task learning
phrase
Self-supervised learning task solved to learn visual representations, with the aim of using the learned representations or model weights obtained in the process, for the downstream task.
Example: Thus, we believe that a proper cropping size is important to control the difficulty of learning pretext tasks. [1]
ru: Обучение по предварительным задач
Pretraining
Noun
To train in advance
Example: Pretraining followed by finetuning with backpropagation has been shown to give significant performance boosts over finetuning from random initializations in certain cases. [7]
ru: Предобучение
Random forest
Phrase
Random forest is a commonly-used machine learning algorithm trademarked by Leo Breiman and Adele Cutler, which combines the output of multiple decision trees to reach a single result. Its ease of use and flexibility have fueled its adoption, as it handles both classification and regression problems.
Example: Growing the 100 trees in random forests was considerably quicker than the 50 trees for Adaboost. [1]
ru: Случайный лес
Random forest
Recall
Noun
Determines the number of true positives among all class labels that have been defined as "positive"
Example: The paper used the emails from Ling-Spam corpus and aimed to acquire an improvement in F1-score, Precision, Recall and Accuracy. [3]
ru: Полнота, чувствительность
Receiver Operating Characteristics-Area Under the Curve (ROC-AUC)
phrase
An ROC curve (receiver operating characteristic curve) is a graph showing the performance of a classification model at all classification thresholds.
Example: F1-Score, Accuracy and Receiver Operating Characteristics-Area Under the Curve (ROC-AUC) metrics are employed to evaluate the performance of our models. [2]
ru: Площадь под кривой ошибок
Recurrent Neural Networks
phrase
Class of artificial neural networks where connections between nodes form a directed or undirected graph along a temporal sequence.
Example: As another important point, RNN has a specific memory, which stores previous computed information for a long time. [2]
ru: Рекуррентные нейронные сети
Regularizing
Verb
Process of reduction the error by fitting a function appropriately on the given training set and avoid overfitting.
Example: Dropout can be interpreted as a way of regularizing a neural network by adding noise to its hidden units. [7]
ru: Регуляризация
Selection
Noun
A variety of items taken from a larger collection.
Example: First, let us consider the model selection problem. [6]
ru: Выборка
Stochastic
Adjective
Random, randomly determined.
Example: Since dropout can be seen as a stochastic regularization technique, it is natural to consider its deterministic counterpart which is obtained by marginalizing out the noise. [7]
ru: Стохастический
Supervised learning
phrase
Machine learning task of learning a function that maps an input to an output based on example input-output pairs.
Example: It is a supervised learning algorithm trained on a dataset [4]
ru: Контролируемое обучение
Support Vector Machine
phrase
Discriminative classifier formally defined by a separating hyperplane.
Example: Support Vector Machine: 0.040 sec approx. [3]
ru: Метод опорных векторов
Tokenisation
Noun
Tokenization is the process of turning a meaningful piece of data, such as an account number, into a random string of characters called a token that has no meaningful value if breached.
Example: Tokenization was explored and the concept provided two stages: Training and Filtering. [3]
ru: Токенизация
Transfer learning
phrase
a research problem in machine learning (ML) that focuses on storing knowledge gained while solving one problem and applying it to a different but related problem.
Example: In transfer learning, a model is trained to perform a visual task (e.g. image classification) but aiming at reusing it to perform a new task (e.g. object detection) in a way that we minimize the amount of labeled data required to train for the new task (e.g. fine-tuning CNNs across tasks is a basic form of transfer learning). [1]
ru: Трансферное обучение
Vector
A directed quantity, one with both magnitude and direction; the signed difference between two points.
Example: This vector is sampled and multiplied element-wise with the outputs of that layer. [7]
ru: Вектор