The NumPy Array: A Structure for Efficient Numerical Computation
By S. V. D. Walt, E. U. S. Chris Colbert, Gaël Varoquaux, Inria Saclay. 2011
7,422 citations recorded by the source corpus.
In the Python world, NumPy arrays are the standard representation for numerical data and enable efficient implementation of numerical computations in a high-level language. As this effort shows, NumPy performance can be improved through three techniques: vectorizing calculations, avoiding copying data in memory, and minimizing operation counts.
Subfields: Artificial Intelligence, Hardware and Architecture, ML Systems & Hardware
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
- Yoshua Bengio
- Mohit Bansal
- Quoc V. Le
- Kyunghyun Cho
- Percy Liang
- Sergey Levine
- Jianfeng Gao
- Ruslan Salakhutdinov
In Hardware and Architecture
- Yoshua Bengio
- Aaron C. Courville
- Kyunghyun Cho
- Colin Raffel
- Ian J. Goodfellow
- Miles Lubin
- Vincent Dumoulin
- Francesco Visin
In ML Systems & Hardware
- Yoshua Bengio
- Ian J. Goodfellow
- J. Vanderplas
- Andrew S. Rosen
- Colin Raffel
- Benjamin Goodrich
- Jonas Horsch
- Tom Brown
Citations in the corpus
- Array programming with NumPy (2020, methodology): “[18] S.”
- Scikit-learn: Machine Learning in Python (2011, methodology): “The central object is an estimator, that implements a fit method, accepting as arguments an input data array and, optionally, an array of labels for supervised problems.”
- Adaptive Stochastic Dual Coordinate Ascent for Conditional Random Fields (2017, methodology): “We used our own implementation 3 of SDCA coded in plain Python and Numpy (Walt et al., 2011).”
- Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context Learning (2022, methodology): “We use “broadcasting notation” [46] so that the (i, j) entry of l x is ljxi,j .”
- Theano: A Python framework for fast computation of mathematical expressions (2016, methodology): “Theano’s API mimics NumPy [6, 7], a widely adopted Python library that provides an n-dimensional array data type and many functions for indexing, reshaping, and performing elementary computations (exp, log, sin, etc.) on entire arrays at once.”
- Weight Agnostic Neural Networks (2019, methodology): “Our NumPy [111] implementation of NEAT [103] supports MPI [32] and OpenAI Gym [10] environments.”
- Albumentations: fast and flexible image augmentations (2018, methodology): “For example, NumPy [64] implementation of an image flip operation used in Albumentations is faster than OpenCV implementation.”
- Statistical Arbitrage in Cryptocurrency Markets (2019, background): “Data preparation mostly relies on the packages numpy (van der Walt et al. 2011) and pandas (McKinney 2010), which are powerful tools for handling large amounts of data.”
- Quantitative missense variant effect prediction using large-scale mutagenesis data (2017, positive): “REAGENT or RESOURCE SOURCE IDENTIFIER Software and Algorithms Python/2.7.3 Python https://www.python.org/ Numpy/1.13.1 van der Walt et al. (2011) http://www.numpy.org/ GraphLab/2.1 Low et al. (2012) https://turi.com/ Scipy/0.19.1 Jones et al. (2001) https://www.scipy.org/ scikit-learn/0.17.0…”
- Reinforcement Learning with Augmented Data (2020, neutral): “A key technical contribution of this work is the custom implementation of data augmentation modules that utilize vectorization [40] and GPU-acceleration [41] to add minimal overhead when processing stacked frames in the minibatch setting.”
- Julia: A Fresh Approach to Numerical Computing (2014, background): “This means that the range of Int values is bounded and wraps around at either end so that adding, subtracting and multiplying integers can overflow or underflow, leading to results that are familiar to C and Fortran programmers but which can be unsettling to users of systems like Mathematica or Pyth”
- Garbage in, garbage out?: do machine learning application papers in social computing report where human-labeled training data comes from? (2019, positive): “All computational analysis and scripting was conducted in Python 3.7 [65], using the following libraries: Pandas dataframes [36] for data parsing and transformation; SciPy [28] and NumPy [64] for quantitative computations; and Matplotlib [25] and Seaborn [67] for visualization.”
- Making AI Forget You: Data Deletion in Machine Learning (2019, methodology): “[82] S.”
- Essentials for Class Incremental Learning (2021, methodology): “Dataset: CIFAR-100 classes are shuffled using a fixed seed (Numpy [28] seed:1993) across all methods for fair comparison.”