convolutional neural network
In plain wordsA kind of neural network built for images and other grid-like data. It slides small pattern detectors across the input, finding edges first, then shapes, then whole objects.
A neural networkneural network, n.: A kind of computer program that learns a task from examples by tuning thousands to billions of connections between simple number-crunching units, an arrangement loosely inspired by nerve cells in the brain. in which some or all layers are convolutionalconvolution, n.: An operation that slides a small grid of weights across an image or signal, multiplying and adding at each spot, to measure how strongly a pattern appears at every location.: each output unit is computed from a small local patch of the input by a filterfilter, n.: In an image-recognition network, a small grid of learned numbers that acts as a pattern detector, for vertical edges, a patch of fur or a wheel, scanned across the whole picture. whose weights are reused at every position (weight sharingweight sharing, n.: Reusing the same learned numbers in many places in a network, for example one edge detector applied at every spot in an image, so the network needs fewer parameters and learns from less data.). This builds in two assumptions suited to images, audio and other grid-structured data, locality and translation equivarianceequivariance, n.: A property of a model whose output changes in step with its input: move the cat in a photo, and the network's internal 'cat' signal moves the same way., and cuts parameter counts by orders of magnitude compared with fully connectedfully connected layer, n.: A layer in which every input is connected to every output by its own adjustable weight. It is the most general and most common kind of neural-network layer. networks. A classic CNN alternates convolutions and nonlinearities with poolingpooling, n.: A step in an image network that shrinks its internal picture by summarizing each small patch with one number, usually the largest or the average, keeping what matters and discarding exact positions. or strided downsampling, so receptive fieldsreceptive field, n.: The patch of the original input, such as a region of an image, that a particular neuron inside a network can see and respond to. widen and feature mapsfeature map, n.: The grid of numbers produced when one learned pattern detector is scanned across an image, showing where, and how strongly, that pattern appears. become coarser but more abstract with depth, and ends in fully connected or global-pooling layers. Landmark CNNs include LeNetLeNet, prop. n.: A family of early image-recognition neural networks built at Bell Labs from 1989 that learned to read handwritten digits. A version read huge numbers of bank checks, and its design underlies today's image models., AlexNetAlexNet, prop. n.: The neural network that won a major image-recognition contest by a huge margin in 2012, convincing researchers and companies that deep learning trained on graphics chips was the way forward., VGGNet, GoogLeNet, ResNetResNet, prop. n.: A family of image-recognition networks from 2015 whose shortcut connections allowed far deeper networks than before, over 100 layers, and became a standard building block across AI., U-NetU-Net, prop. n.: An image network shaped like the letter U: one half shrinks the picture to understand it, the other enlarges it back, with shortcuts between the halves so fine detail is not lost. and EfficientNet. CNNs dominated computer visioncomputer vision, n.: The part of AI that lets computers make sense of images and video: recognizing faces, reading signs, spotting tumors on scans or judging how far away a car is. from 2012 until vision transformersvision transformer, n.: A way of applying the transformer, the design behind chatbots, to pictures: the image is cut into small square tiles that the model reads like words in a sentence. became competitive around 2020–21; hybrids are common, and CNNs remain standard on low-power devices.
We fine-tuned a pretrained CNN on 2,000 chest X-rays instead of training from scratch.
Origin
Kunihiko Fukushima's neocognitron (1980) introduced layers of shared-weight feature detectors modeled on David Hubel and Torsten Wiesel's simple and complex cells. Alex Waibel's time-delay neural network (1987), Wei Zhang's shift-invariant network (1988) and Yann LeCun's 1989 zip-code recognizer trained shared weights by backpropagation; LeCun's work led to LeNet. 'Convolutional network' became the usual name in the early 1990s, and AlexNet's 2012 ImageNet win made CNNs dominant.
Usage note
Despite the name, most CNN libraries compute cross-correlation (the filter is not flipped); because the filter is learned, the difference is immaterial.
Related
- convolutionconvolution, n.: An operation that slides a small grid of weights across an image or signal, multiplying and adding at each spot, to measure how strongly a pattern appears at every location.
- filterfilter, n.: In an image-recognition network, a small grid of learned numbers that acts as a pattern detector, for vertical edges, a patch of fur or a wheel, scanned across the whole picture.
- poolingpooling, n.: A step in an image network that shrinks its internal picture by summarizing each small patch with one number, usually the largest or the average, keeping what matters and discarding exact positions.
- receptive fieldreceptive field, n.: The patch of the original input, such as a region of an image, that a particular neuron inside a network can see and respond to.
- feature mapfeature map, n.: The grid of numbers produced when one learned pattern detector is scanned across an image, showing where, and how strongly, that pattern appears.
- LeNetLeNet, prop. n.: A family of early image-recognition neural networks built at Bell Labs from 1989 that learned to read handwritten digits. A version read huge numbers of bank checks, and its design underlies today's image models.
- AlexNetAlexNet, prop. n.: The neural network that won a major image-recognition contest by a huge margin in 2012, convincing researchers and companies that deep learning trained on graphics chips was the way forward.
- ResNetResNet, prop. n.: A family of image-recognition networks from 2015 whose shortcut connections allowed far deeper networks than before, over 100 layers, and became a standard building block across AI.
- vision transformervision transformer, n.: A way of applying the transformer, the design behind chatbots, to pictures: the image is cut into small square tiles that the model reads like words in a sentence.
- multilayer perceptronmultilayer perceptron, n.: A neural network made of several layers of artificial neurons, each connected to every neuron in the next layer. It is the simplest kind of deep network and a component of almost every modern model.
Sources
- LeCun, Y. et al. (1989). Backpropagation Applied to Handwritten Zip Code Recognition. Neural Computation, 1(4), 541–551.
- LeCun, Y., Bottou, L., Bengio, Y. & Haffner, P. (1998). Gradient-Based Learning Applied to Document Recognition. Proceedings of the IEEE, 86(11), 2278–2324.
- Fukushima, K. (1980). Neocognitron: A Self-Organizing Neural Network Model for a Mechanism of Pattern Recognition Unaffected by Shift in Position. Biological Cybernetics, 36, 193–202.
Mentioned in
AlphaGoAlphaGo, prop. n.: A DeepMind program that in 2016 beat Lee Sedol, one of the world's best players of the board game Go, a decade or more before many experts expected a computer to manage it., computer visioncomputer vision, n.: The part of AI that lets computers make sense of images and video: recognizing faces, reading signs, spotting tumors on scans or judging how far away a car is., deep learningdeep learning, n.: Machine learning done with neural networks built from many stacked layers: raw data such as pixels, sound or text goes in at one end, and each layer turns what it receives into something more abstract., image classificationimage classification, n.: Teaching a computer to say what a picture shows by choosing a label, such as 'cat', 'pneumonia' or 'stop sign', from a fixed list of possibilities., neural networkneural network, n.: A kind of computer program that learns a task from examples by tuning thousands to billions of connections between simple number-crunching units, an arrangement loosely inspired by nerve cells in the brain., optical character recognitionoptical character recognition, n.: Software that picks out the letters and words in a picture — a road sign, a handwritten note, a page of an old book — and turns them into real text a computer can copy, search or read aloud., transformertransformer, n.: The design behind today's chatbots and much of modern AI. It handles text by letting every word look at every other word and decide which of them matter for understanding it., Yann LeCunYann LeCun, prop. n.: A French-American computer scientist who pioneered the neural networks that recognize images, led Meta's AI research for over a decade, and argues that chatbots alone will not reach human-level AI., AlexNetAlexNet, prop. n.: The neural network that won a major image-recognition contest by a huge margin in 2012, convincing researchers and companies that deep learning trained on graphics chips was the way forward., Andrej KarpathyAndrej Karpathy, prop. n.: A Slovak-Canadian AI researcher and teacher who co-founded OpenAI, led Tesla's self-driving vision team, and coined popular phrases such as 'vibe coding'., architecturearchitecture, n.: The overall design of an AI model, meaning what kinds of parts it has and how they are connected, as opposed to the specific numbers it learns during training., batch normalizationbatch normalization, n.: A layer that rescales the numbers flowing through a network so that, across each batch of training examples, they keep a steady average and spread. It makes deep networks train faster and more reliably., biasbias, n.: A word with several AI meanings: an adjustable number added inside a model; a systematic error in estimates; or unfair treatment of some groups of people by an AI system., convolutionconvolution, n.: An operation that slides a small grid of weights across an image or signal, multiplying and adding at each spot, to measure how strongly a pattern appears at every location., CVPRCVPR, prop. n.: The leading yearly conference on computer vision — teaching computers to understand images and video — and one of the most cited publication venues in any field of science., deep Q-networkdeep Q-network, n.: A method, made famous by DeepMind in 2013–2015, in which a neural network learned to play dozens of Atari video games from screen pixels and the score alone, many at human level., featurefeature, n.: One piece of information about each example that a model uses, like a house's size or a pixel's color, or a pattern a model has learned to detect inside itself., feature extractionfeature extraction, n.: Turning raw data such as pixels or sound into a short list of telling numbers — computed by hand-designed rules or by an AI system trained earlier — that a new model can learn from or compare., feature mapfeature map, n.: The grid of numbers produced when one learned pattern detector is scanned across an image, showing where, and how strongly, that pattern appears., feedforward neural networkfeedforward neural network, n.: A neural network laid out like an assembly line: each stage hands its result only to later stages, never back, so every answer depends on the current input alone and nothing is remembered between inputs., filterfilter, n.: In an image-recognition network, a small grid of learned numbers that acts as a pattern detector, for vertical edges, a patch of fur or a wheel, scanned across the whole picture., fully connected layerfully connected layer, n.: A layer in which every input is connected to every output by its own adjustable weight. It is the most general and most common kind of neural-network layer., godfathers of AIgodfathers of AI, phr.: A media nickname for three scientists — Geoffrey Hinton, Yoshua Bengio and Yann LeCun — whose decades of work on neural networks underpins modern AI. They shared computing's top prize in 2019., Grad-CAMGrad-CAM, abbr.: A technique that shows, as a coarse heat map, which regions of an image a vision network relied on when choosing a particular label.