人工智慧

Randomly labeled data refers to a dataset where the labels assigned to the data points are completely random, not related to the underlying content or patterns in the data. For example, in an image classification task, pictures of cats might be assigned labels like "car" or "ant" without any connection to the actual image.

This type of labeling is often used in research to study how neural networks behave, especially to explore memorization versus generalization. Neural networks trained on randomly labeled data are unable to learn meaningful patterns, yet they may still memorize the input-label pairs, highlighting the model’s capacity to overfit when there is no true relationship between inputs and outputs .

Takoua Saadani, Understanding Data Labels and Data Labeling: Definition, Types, and How it Works for Machine Learning, Jun 29, 2023

Machine Learning Platform

TensorFlow: TensorFlow makes it easy to create ML models that can run in any environment. Learn how to use the intuitive APIs through interactive code samples. TensorFlow is a free and open-source software library for machine learning and artificial intelligence. It can be used across a range of tasks but has a particular focus on training and inference of deep neural networks. It is one of the two most popular deep learning libraries alongside PyTorch. (ref. https://en.wikipedia.org/wiki/TensorFlow)

PyTorch: PyTorch is a machine learning library based on the Torch library,[4][5][6] used for applications such as computer vision and natural language processing,[7] originally developed by Meta AI and now part of the Linux Foundation umbrella.[8][9][10][11] It is one of the two most popular deep learning libraries alongside TensorFlow, offering free and open-source software released under the modified BSD license. Although the Python interface is more polished and the primary focus of development, PyTorch also has a C++ interface.[12](ref. https://en.wikipedia.org/wiki/PyTorch )

Kaggle: "Your Machine Learning and Data Science Community". The world's largest data science community with powerful tools and resources to help you achieve your data science goals. Kaggle is a data science competition platform and online community for data scientists and machine learning practitioners under Google LLC. Kaggle enables users to find and publish datasets, explore and build models in a web-based data science environment, work with other data scientists and machine learning engineers, and enter competitions to solve data science challenges. (ref. https://en.wikipedia.org/wiki/Kaggle)

Tools

scikit-learn (Machine Learning in Python)

Simple and efficient tools for predictive data analysis
Accessible to everybody, and reusable in various contexts
Built on NumPy, SciPy, and matplotlib
Open source, commercially usable - BSD license

Image processing:

Template Matching

Using Normalized Cross Correlation:

image

(450X280)

template

(40X52)

Convolution Neural Network

Examples

Picachuy.jpg

Ganesh.jpg

How does the gradient estimation work?

Back-propagation: Back-propagation(BP)是目前深度學習大多數NN(Neural Network)模型更新梯度的方式，In machine learning, backpropagation is a gradient estimation method commonly used for training neural networks to compute the network parameter updates. (ref. https://en.wikipedia.org/wiki/Backpropagation)
Hebbian theory: Hebbian theory is a neuropsychological theory claiming that an increase in synaptic efficacy arises from a presynaptic cell's repeated and persistent stimulation of a postsynaptic cell. It is an attempt to explain synaptic plasticity, the adaptation of brain neurons during the learning process. It was introduced by Donald Hebb in his 1949 book The Organization of Behavior. The theory is also called Hebb's rule, Hebb's postulate, and cell assembly theory. (ref. https://en.wikipedia.org/wiki/Backpropagation)