A few useful things to know about machine learning

By Pedro M. Domingos. 2012

3,280 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistical Learning Theory

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Statistical Learning Theory

Citations in the corpus

Detecting Backdoor Attacks on Deep Neural Networks by Activation Clustering (2018, background): “In particular, as dimensionality increases distance metrics in general (and speciﬁcally the Euclidean metric used here), are less effective at distinguishing near and far points in high dimensional spaces Domingos (2012).”
A practical tutorial on autoencoders for nonlinear feature fusion: Taxonomy, models, software and guidelines (2018, background): “The most common are as follows: Feature engineering [19].”
How the machine ‘thinks’: Understanding opacity in machine learning algorithms (2016, background): “In an article on the folk knowledge of applying machine learning , Domingos (2012) notes that ‘intu-ition fails at high-dimensions.’”
Automatic Chemical Design Using a Data-Driven Continuous Representation of Molecules (2016, neutral): “34 Interpolating linearly between two points might pass by an area of low probability, to keep the sampling on the areas of high probability we utilize spherical interpolation 35 ( slerp ).”
Recurrent Neural Networks for Classifying Relations in Clinical Notes (2017, neutral): “Systems that use many human engineered features often do not generalize well to new datasets [21].”
Understanding disentangling in β-VAE (2018, background): “From the hand-crafted feature engineering prevalent in the past [11] to implicit representation learning of the modern deep learning approaches [23, 14, 38], it is a common theme that the performance of algorithms is critically dependent on the nature of their input representations.”
Meta-Learning in Neural Networks: A Survey (2020, background): “Historically, the success ofmachine learningwas driven by advancing hand-engineered features [11], [12].”
Problem Formulation and Fairness (2019, background): “Indeed, the openendedness that characterizes such forms of translation work is often described as the ‘art’ of data mining [12].”
The ethics of algorithms: Mapping the debate (2016, neutral): “A machine learning algorithm applied to classiﬁcation tasks, for example, typically consists of two components, a learner which produces a classiﬁer , with the intention to develop classes that can generalise beyond the training data (Domingos, 2012).”
A data-driven approach to predict the success of bank telemarketing (2014, methodology): “• We focus on feature engineering, which is a key aspect in DM [10], and propose generic social and economic indicators in addition to the more commonly used bank client and product attributes, in a total of 150 analyzed features.”
Deep Learning for Health Informatics (2017, background): “For more general background information related to the theory of machine learning, the reader can refer to the works in [23]–[25] where common training problems, such as overfitting, model interpretation and generalization, are explained in detail.”
Human Activity Recognition using Inertial, Physiological and Environmental Sensors: A Comprehensive Survey (2020, methodology): “Typically, one or more preprocessing steps will be also required, including feature extraction, vectorization/ segmentation, normalization or standardization, and projection [40].”
Automated Crack Detection on Concrete Bridges (2016, methodology): “Machine learning has been applied for visual recognition and classification and generally perform better than methods with hand-tuned parameters [16]–[21].”
On learning and branching: a survey (2017, background): “An interesting outlook of these and other concerns can be found in Domingos (2012).”
Machine-learning-guided directed evolution for protein engineering (2018, background): “How each protein sequence is vectorized determines what can be learned 63,64 .”