Interface administrators, Administrators (Semantic MediaWiki), Curators (Semantic MediaWiki), Editors (Semantic MediaWiki), Suppressors, Administrators
7,785
edits
Training: Difference between revisions
no edit summary
No edit summary |
No edit summary |
||
| Line 10: | Line 10: | ||
===Supervised Learning=== | ===Supervised Learning=== | ||
[[Supervised learning]] is the most common type of machine learning training. This method involves providing the algorithm with a [[labeled dataset]], in which each [[data point]] has an assigned [[class]] [[label]]. The algorithm then utilizes this labeled data to learn relationships between [[features]] in the data and their corresponding labels. Ultimately, supervised learning aims to build a model capable of accurately predicting labels for new, unseen [[data sets]]. | [[Supervised learning]] is the most common type of machine learning training. This method involves providing the [[algorithm]] with a [[labeled dataset]], in which each [[data point]] has an assigned [[class]] [[label]]. The algorithm then utilizes this labeled data to learn relationships between [[features]] in the data and their corresponding labels. Ultimately, supervised learning aims to build a model capable of accurately predicting labels for new, unseen [[data sets]]. | ||
===Unsupervised Learning=== | ===Unsupervised Learning=== | ||
Unsupervised learning is a type of machine-learning training in which an algorithm is given data that has not yet been labeled. The objective is for this | [[Unsupervised learning]] is a type of machine-learning training in which an algorithm is given data that has not yet been labeled. The objective is for this algorithm to discover patterns and relationships without the assistance of labeled information. Unsupervised learning is often employed for [[clustering]] and [[dimensionality reduction]] tasks, as it allows the computer to recognize structure within data that may otherwise go undetected. | ||
===Semi-Supervised Learning=== | ===Semi-Supervised Learning=== | ||
Semi-supervised learning is a type of training that incorporates elements from both supervised and unsupervised learning methods. In this type of scenario, an algorithm receives some labeled data but most of it remains unlabeled. The algorithm then utilizes this labeled information to direct its learning process and make predictions about unlabeled data. | Semi-supervised learning is a type of training that incorporates elements from both supervised and unsupervised learning methods. In this type of scenario, an algorithm receives some [[labeled data]] but most of it remains [[unlabeled]]. The algorithm then utilizes this labeled information to direct its learning process and make predictions about unlabeled data. | ||
===Reinforcement Learning=== | ===Reinforcement Learning=== | ||
Reinforcement learning is a type of training in which an algorithm learns from its own actions and the outcomes. In reinforcement learning, an algorithm is given an environment and set of actions it can take within that environment; then it learns which actions produce the most favorable outcomes based on feedback from its environment. | [[Reinforcement learning]] is a type of training in which an algorithm learns from its own actions and the outcomes. In reinforcement learning, an algorithm is given an environment and set of actions it can take within that environment; then it learns which actions produce the most favorable outcomes based on feedback from its environment. | ||
==Role of Training in the Machine Learning Process== | ==Role of Training in the Machine Learning Process== | ||
| Line 28: | Line 28: | ||
==Explain Like I'm 5 (ELI5)== | ==Explain Like I'm 5 (ELI5)== | ||
Picture having a big box of crayons and wanting to teach a robot how to color. Since the robot doesn't know how yet, you must demonstrate for it. | |||
Take out one crayon at a time and show the robot what color it is. Let them try to color with it - it may not match exactly, but that's okay; with | Take out one crayon at a time and show the robot what color it is. Let them try to color with it - it may not match exactly, but that's okay; with your help each time, it becomes better at coloring. | ||
Training a machine learning model is kind of like training a robot. You give it some [[examples]], and it tries to learn from them. The more examples you give it, the better it gets at coloring on its own. | |||
[[Category:Terms]] [[Category:Machine learning terms]] | [[Category:Terms]] [[Category:Machine learning terms]] | ||