This chapter discusses continuous latent variable models including principal component analysis (PCA), probabilistic PCA, and factor analysis. PCA finds projections of data that maximize variance or minimize error through eigenvectors of the covariance matrix. Probabilistic PCA places a probabilistic treatment on PCA by modeling the data and latent variables as Gaussian distributions. Factor analysis similarly models the data as a linear combination of latent factors plus noise.
The document discusses reinforcement learning and its application to training an AI agent to play Super Mario Bros. It begins by explaining how animals and humans learn through reinforcement and punishment in their environments based on the rewards and consequences of their actions. It then provides an overview of reinforcement learning, including key concepts like the Markov decision process, exploration versus exploitation, and using a replay memory buffer to train a deep learning model. It concludes by describing how the Super Mario Bros environment can be set up and used with a reinforcement learning agent, including defining the state and action spaces, rewards and penalties, and the process of minimizing the loss to optimize the agent's behavior.
This document discusses support vector machines (SVMs) for classification. It explains that SVMs find the optimal separating hyperplane that maximizes the margin between positive and negative examples. This is formulated as a convex optimization problem. Both primal and dual formulations are presented, with the dual having fewer variables that scale with the number of examples rather than dimensions. Methods for handling non-separable data using soft margins and kernels for nonlinear classification are also summarized. Popular kernel functions like polynomial and Gaussian kernels are mentioned.
猟彌B初Learning From Noisy Labels With Deep Neural Networks: A SurveyToru Tamaki
?
H. Song, M. Kim, D. Park, Y. Shin and J. -G. Lee, "Learning From Noisy Labels With Deep Neural Networks: A Survey", in IEEE Transactions on Neural Networks and Learning Systems, doi: 10.1109/TNNLS.2022.3152527. IEEE TNNLS 2022
https://ieeexplore.ieee.org/document/9729424
https://arxiv.org/abs/2007.08199
A Support Vector Machine (SVM) is a discriminative classifier formally defined by a separating hyperplane. In other words, given labeled training data (supervised learning), the algorithm outputs an optimal hyperplane which categorizes new examples. In two dimentional space this hyperplane is a line dividing a plane in two parts where in each class lay in either side.
This document discusses the perceptron algorithm for linear classification. It begins by introducing feature representations and linear classifiers. It then describes the perceptron algorithm, which attempts to learn a weight vector that separates the training data into classes with some margin. The document proves that for any separable training set, the perceptron will converge after a finite number of mistakes, where the number depends on the margin size and properties of the data. However, it notes that while the perceptron finds weights perfectly classifying the training data, these weights may not generalize well to new examples.
The document discusses fuzzy sets and rough sets. It explains that fuzzy sets allow partial membership, where an element can belong to a set to a degree between 0 and 1, while in classical sets an element either fully belongs or does not belong. Rough sets address vagueness through boundary regions rather than partial membership. Rough sets are defined using topological operations on lower and upper approximations. Indiscernibility relations are also discussed, where indiscernible elements cannot be distinguished based on available attributes.
猟彌B初Learning From Noisy Labels With Deep Neural Networks: A SurveyToru Tamaki
?
H. Song, M. Kim, D. Park, Y. Shin and J. -G. Lee, "Learning From Noisy Labels With Deep Neural Networks: A Survey", in IEEE Transactions on Neural Networks and Learning Systems, doi: 10.1109/TNNLS.2022.3152527. IEEE TNNLS 2022
https://ieeexplore.ieee.org/document/9729424
https://arxiv.org/abs/2007.08199
A Support Vector Machine (SVM) is a discriminative classifier formally defined by a separating hyperplane. In other words, given labeled training data (supervised learning), the algorithm outputs an optimal hyperplane which categorizes new examples. In two dimentional space this hyperplane is a line dividing a plane in two parts where in each class lay in either side.
This document discusses the perceptron algorithm for linear classification. It begins by introducing feature representations and linear classifiers. It then describes the perceptron algorithm, which attempts to learn a weight vector that separates the training data into classes with some margin. The document proves that for any separable training set, the perceptron will converge after a finite number of mistakes, where the number depends on the margin size and properties of the data. However, it notes that while the perceptron finds weights perfectly classifying the training data, these weights may not generalize well to new examples.
The document discusses fuzzy sets and rough sets. It explains that fuzzy sets allow partial membership, where an element can belong to a set to a degree between 0 and 1, while in classical sets an element either fully belongs or does not belong. Rough sets address vagueness through boundary regions rather than partial membership. Rough sets are defined using topological operations on lower and upper approximations. Indiscernibility relations are also discussed, where indiscernible elements cannot be distinguished based on available attributes.
Dsh data sensitive hashing for high dimensional k-nn searchWooSung Choi
?
Gao, Jinyang, et al. "Dsh: data sensitive hashing for high-dimensional k-nnsearch." Proceedings of the 2014 ACM SIGMOD international conference on Management of data. ACM, 2014.
10. ? Affine set
$ ? ? ? ?
is affine ★ ? ?? 2?? ?? ?? point? ??? line? ?? ?? ?.
$ ?? ??, ?? or hyperplane ?? ??.
? Convex set
$ ? ? ? ?
is convex ★ ? ?? 2?? ?? ?? point? ??? line segment? ?? ??
?.
$ ?? ??, ? or ??? ?? ??.
Affine / Convex
convex set (???)convex set (??)
line
line segment
11. ? Affine function
$ a function composed of a linear function and constant (translation)
$ in 1-dim : ? = ?? + ?
$ in 2-dim : ? ?, ? = ?? + ?? + ?
$ in 3-dim : ? ?, ?, ? = ?? + ?? + ?? + ?
translation : a transformation consisting of a constant offset with no rotation or distortion
Affine function
15. ? Lagrange dual function ?: ? ? 〜 ? ? ★ ?
? Lagrange dual problem
$ ?(Λ, ?) is concave and constraints are convex, so this is convex optimization
problem.
$ Lagrange dual problem? ?(Λ?
, ??
)? primal problem(??
) ?? lower bound? ??
??.
inf ?(??
, Λ?
, ??
) + ?0(??
)
? Λ, ? = inf
?(?
?(?, Λ, ? ) = inf
?(?
?0 ? + ?
?=1
?
?? ?? ? + ?
?=1
?
???? ?
Lagrange dual function /
Lagrange dual problem
maximize ?(Λ, ?)
subject to Λ ? 0
How to make it coincide?
16. ? If
$ ?? are convex (????, inequality constraint)
$ ?? are affine (equality constraint)
$ ??
, Λ?
, ??
satisfy KKT condition
? Then
$ ??
is primal optimal and (Λ?
, ??
) is dual optimal with zero duality gap.
$ ? ?? ?? ????.
KKT condition and
zero duality
53. ? ???? - ??? ?
? http://cs229.stanford.edu/materials/smo.pdf
? http://cs229.stanford.edu/notes/cs229-notes3.pdf - Andrew Ng
? https://web.stanford.edu/~boyd/cvxbook/bv_cvxbook.pdf
? Fast Training of Support Vector Machine using Sequential Minimal Optimization,
in Advances in Kernel Methods - Platt John
? Machine Learning in Action - ?? ???
? Machine Learning-A Probabilistic Perspective - Kevin P. Murphy