When you type'machine learning' into Google News, the first link you see is a Forbes Magazine piece called "What's The Difference Between Machine Learning And Artificial Intelligence?" This article contained so many flowery, grandiose descriptions about ML and AI technology that I couldn't help but laugh. With all the nonsense the media uses to describe machine learning (ML) and artificial intelligence (AI), it's time we do a deep dive into what these technologies actually do. First, we need to learn the difference between AI and ML. Fortunately, a fellow writer has already written an excellent explanation here.

One of the most amazing things about Python's scikit-learn library is that is has a 4-step modeling pattern that makes it easy to code a machine learning classifier. While this tutorial uses a classifier called Logistic Regression, the coding process in this tutorial applies to other classifiers in sklearn (Decision Tree, K-Nearest Neighbors etc). In this tutorial, we use Logistic Regression to predict digit labels based on images. The image above shows a bunch of training digits (observations) from the MNIST dataset whose category membership is known (labels 0–9). After training a model with logistic regression, it can be used to predict an image label (labels 0–9) given an image. The first part of this tutorial post goes over a toy dataset (digits dataset) to show quickly illustrate scikit-learn's 4 step modeling pattern and show the behavior of the logistic regression algorthm.

Get the code: To follow along, all the code is also available as an iPython notebook on Github. In this post we will implement a simple 3-layer neural network from scratch. We won't derive all the math that's required, but I will try to give an intuitive explanation of what we are doing. I will also point to resources for you read up on the details. Here I'm assuming that you are familiar with basic Calculus and Machine Learning concepts, e.g.

Several multiscale methods account for sub-grid scale features using coarse scale basis functions. For example, in the Multiscale Finite Volume method the coarse scale basis functions are obtained by solving a set of local problems over dual-grid cells. We introduce a data-driven approach for the estimation of these coarse scale basis functions. Specifically, we employ a neural network predictor fitted using a set of solution samples from which it learns to generate subsequent basis functions at a lower computational cost than solving the local problems. The computational advantage of this approach is realized for uncertainty quantification tasks where a large number of realizations has to be evaluated. We attribute the ability to learn these basis functions to the modularity of the local problems and the redundancy of the permeability patches between samples. The proposed method is evaluated on elliptic problems yielding very promising results.

This composition between two finite-sum structures 1 n n i 1 F i ( 1 m m j 1 G j (x)) arises in many machine learning applications such as reinforcement learning [1, 2, 3] and nonlinear embedding [4]. For example, stochastic neighbor embedding (SNE) [4] is a powerful approach to map data from a high dimensional space to a low dimensional space. Let{ z i} n i 1 and { x i} n i 1 denote the representation ofn data points in the high dimensional space and the low dimensional space, respectively. The objective is to pursue a low dimensional representation{ x i} n i 1, such that the distribution in the low dimensional space is as close to the distribution in the high dimensional space as possible. This problem is essentially a composition optimization problem: min x t i p i t log p i t q i t, (2) where p i t exp( ‖ z t z i‖ 2 / 2σ 2 i) j 6 t exp( ‖ z t z j ‖ 2 /2σ 2 i), q i t exp( ‖ x t x i‖ 2) j 6 t exp( ‖ x t x j ‖ 2), lliu8101@uni.sydney.edu.au

The worldwide surge of multiresistant microbial strains has propelled the search for alternative treatment options. The study of Protein-Protein Interactions (PPIs) has been a cornerstone in the clarification of complex physiological and pathogenic processes, thus being a priority for the identification of vital components and mechanisms in pathogens. Despite the advances of laboratorial techniques, computational models allow the screening of protein interactions between entire proteomes in a fast and inexpensive manner. Here, we present a supervised machine learning model for the prediction of PPIs based on the protein sequence. We cluster amino acids regarding their physicochemical properties, and use the discrete cosine transform to represent protein sequences. A mesh of classifiers was constructed to create hyper-specialised classifiers dedicated to the most relevant pairs of molecular function annotations from Gene Ontology. Based on an exhaustive evaluation that includes datasets with different configurations, cross-validation and out-of-sampling validation, the obtained results outscore the state-of-the-art for sequence-based methods. For the final mesh model using SVM with RBF, a consistent average AUC of 0.84 was attained.

Spectral clustering has found extensive use in many areas. Most traditional spectral clustering algorithms work in three separate steps: similarity graph construction; continuous labels learning; discretizing the learned labels by k-means clustering. Such common practice has two potential flaws, which may lead to severe information loss and performance degradation. First, predefined similarity graph might not be optimal for subsequent clustering. It is well-accepted that similarity graph highly affects the clustering results. To this end, we propose to automatically learn similarity information from data and simultaneously consider the constraint that the similarity matrix has exact c connected components if there are c clusters. Second, the discrete solution may deviate from the spectral solution since k-means method is well-known as sensitive to the initialization of cluster centers. In this work, we transform the candidate solution into a new one that better approximates the discrete one. Finally, those three subtasks are integrated into a unified framework, with each subtask iteratively boosted by using the results of the others towards an overall optimal solution. It is known that the performance of a kernel method is largely determined by the choice of kernels. To tackle this practical problem of how to select the most suitable kernel for a particular data set, we further extend our model to incorporate multiple kernel learning ability. Extensive experiments demonstrate the superiority of our proposed method as compared to existing clustering approaches.

Image understanding using deep convolutional network has reached human-level performance, yet a closely related problem of video understanding especially, action recognition has not reached the requisite level of maturity. We combine multi-kernels based support-vector-machines (SVM) with a multi-stream deep convolutional neural network to achieve close to state-of-the-art performance on a 51-class activity recognition problem (HMDB-51 dataset); this specific dataset has proved to be particularly challenging for deep neural networks due to the heterogeneity in camera viewpoints, video quality, etc. The resulting architecture is named pillar networks as each (very) deep neural network acts as a pillar for the hierarchical classifiers. In addition, we illustrate that hand-crafted features such as improved dense trajectories (iDT) and Multi-skip Feature Stacking (MIFS), as additional pillars, can further supplement the performance.

When used as a surrogate objective for maximum likelihood estimation in latent variable models, the evidence lower bound (ELBO) produces state-of-the-art results. Inspired by this, we consider the extension of the ELBO to a family of lower bounds defined by a particle filter's estimator of the marginal likelihood, the filtering variational objectives (FIVOs). FIVOs take the same arguments as the ELBO, but can exploit a model's sequential structure to form tighter bounds. We present results that relate the tightness of FIVO's bound to the variance of the particle filter's estimator by considering the generic case of bounds defined as log-transformed likelihood estimators. Experimentally, we show that training with FIVO results in substantial improvements over training the same model architecture with the ELBO on sequential data.

Collecting labeled data is costly and thus a critical bottleneck in real-world classification tasks. To mitigate this problem, we propose a novel setting, namely learning from complementary labels for multi-class classification. A complementary label specifies a class that a pattern does not belong to. Collecting complementary labels would be less laborious than collecting ordinary labels, since users do not have to carefully choose the correct class from a long list of candidate classes. However, complementary labels are less informative than ordinary labels and thus a suitable approach is needed to better learn from them. In this paper, we show that an unbiased estimator to the classification risk can be obtained only from complementarily labeled data, if a loss function satisfies a particular symmetric condition. We derive estimation error bounds for the proposed method and prove that the optimal parametric convergence rate is achieved. We further show that learning from complementary labels can be easily combined with learning from ordinary labels (i.e., ordinary supervised learning), providing a highly practical implementation of the proposed method. Finally, we experimentally demonstrate the usefulness of the proposed methods.