Deep learning, machine learning, artificial intelligence -- all buzzwords that represent the future of analytics. In this post, we will explain what machine learning and deep learning are at a high level with some real-world examples. In future posts, we will explore vertical use cases. The goal of this is not to turn you into a data scientist but to give you a better understanding of what you can do with machine learning. Machine learning is becoming more accessible to developers, and data scientists work with domain experts, architects, developers, and data engineers, so it is important for everyone to have a good understanding of the possibilities.

There were many advances in Deep Learning and AI in 2017, but few generated as much publicity and interest as DeepMind's AlphaGo Zero. This program was truly a shocking breakthrough: not only did it beat the prior version of AlphaGo -- the program that beat 17 time world champion Lee Sedol just a year and a half earlier -- 100–0, it was trained without any data from real human games. Xavier Amatrain called it "more [significant] than anything…in the last 5 years" in Machine Learning. So how did DeepMind do it? In this essay, I'll try to give an intuitive idea of the techniques AlphaGo Zero used, what made them work, and what the implications for future AI research are.

The rise of deep learning in recent years has brought with it increasingly clever optimization methods to deal with complex, non-linear loss functions. These methods are often designed with convex optimization in mind, but have been shown to work well in practice even for the highly non-convex optimization associated with neural networks. However, one significant drawback of these methods when they are applied to deep learning is that the magnitude of the update step is sometimes disproportionate to the magnitude of the weights (much smaller or larger), leading to training instabilities such as vanishing and exploding gradients. An idea to combat this issue is gradient descent with proportional updates. Gradient descent with proportional updates was introduced in 2017. It was independently developed by You et al (Layer-wise Adaptive Rate Scaling (LARS) algorithm) and by Abu-El-Haija (PercentDelta algorithm). The basic idea of both of these algorithms is to make each step of the gradient descent proportional to the current weight norm and independent of the gradient magnitude. It is common in the context of new optimization methods to prove convergence or derive regret bounds under the assumption of Lipschitz continuity and convexity. However, even though LARS and PercentDelta were shown to work well in practice, there is no theoretical analysis of the convergence properties of these algorithms. Thus it is not clear if the idea of gradient descent with proportional updates is used in the optimal way, or if it could be improved by using a different norm or specific learning rate schedule, for example. Moreover, it is not clear if these algorithms can be extended to other problems, besides neural networks. We attempt to answer these questions by establishing the theoretical analysis of gradient descent with proportional updates, and verifying this analysis with empirical examples.

Graph Convolutional Neural Networks (Graph CNNs) are generalizations of classical CNNs to handle graph data such as molecular data, point could and social networks. Current filters in graph CNNs are built for fixed and shared graph structure. However, for most real data, the graph structures varies in both size and connectivity. The paper proposes a generalized and flexible graph CNN taking data of arbitrary graph structure as input. In that way a task-driven adaptive graph is learned for each graph data while training. To efficiently learn the graph, a distance metric learning is proposed. Extensive experiments on nine graph-structured datasets have demonstrated the superior performance improvement on both convergence speed and predictive accuracy.

We discuss its design and provide brief experimental results demonstrating its usefulness in improving the classification correctness of LSTM-AD, a state-of-the-art anomaly detection model.

To compare entities of differing types and structural components, the artificial neural network paradigm was used to cross-compare structural components between heterogeneous documents. Trainable weighted structural components were input into machine-learned activation functions of the neurons. The model was used for matching news articles and videos, where the inputs and activation functions respectively consisted of term vectors and cosine similarity measures between the weighted structural components. The model was tested with different weights, achieving as high as 59.2% accuracy for matching videos to news articles. A mobile application user interface for recommending related videos for news articles was developed to demonstrate consumer value, including its potential usefulness for cross-selling products from unrelated categories.

Deep neural networks are vulnerable to adversarial examples. Prior defenses attempted to make deep networks more robust by either improving the network architecture or adding adversarial examples into the training set, with their respective limitations. We propose a new direction. Motivated by recent research that shows that outliers in the training set have a high negative influence on the trained model, our approach makes the model more robust by detecting and removing outliers in the training set without modifying the network architecture or requiring adversarial examples. We propose two methods for detecting outliers based on canonical examples and on training errors, respectively. After removing the outliers, we train the classifier with the remaining examples to obtain a sanitized model. Our evaluation shows that the sanitized model improves classification accuracy and forces the attacks to generate adversarial examples with higher distortions. Moreover, the Kullback-Leibler divergence from the output of the original model to that of the sanitized model allows us to distinguish between normal and adversarial examples reliably.

Recent studies show that widely used deep neural networks (DNNs) are vulnerable to carefully crafted adversarial examples. Many advanced algorithms have been proposed to generate adversarial examples by leveraging the $\mathcal{L}_p$ distance for penalizing perturbations. Researchers have explored different defense methods to defend against such adversarial attacks. While the effectiveness of $\mathcal{L}_p$ distance as a metric of perceptual quality remains an active research area, in this paper we will instead focus on a different type of perturbation, namely spatial transformation, as opposed to manipulating the pixel values directly as in prior works. Perturbations generated through spatial transformation could result in large $\mathcal{L}_p$ distance measures, but our extensive experiments show that such spatially transformed adversarial examples are perceptually realistic and more difficult to defend against with existing defense systems. This potentially provides a new direction in adversarial example generation and the design of corresponding defenses. We visualize the spatial transformation based perturbation for different examples and show that our technique can produce realistic adversarial examples with smooth image deformation. Finally, we visualize the attention of deep networks with different types of adversarial examples to better understand how these examples are interpreted.

Recurrent neural networks are nowadays successfully used in an abundance of applications, going from text, speech and image processing to recommender systems. Backpropagation through time is the algorithm that is commonly used to train these networks on specific tasks. Many deep learning frameworks have their own implementation of training and sampling procedures for recurrent neural networks, while there are in fact multiple other possibilities to choose from and other parameters to tune. In existing literature this is very often overlooked or ignored. In this paper we therefore give an overview of possible training and sampling schemes for character-level recurrent neural networks to solve the task of predicting the next token in a given sequence. We test these different schemes on a variety of datasets, neural network architectures and parameter settings, and formulate a number of take-home recommendations. The choice of training and sampling scheme turns out to be subject to a number of trade-offs, such as training stability, sampling time, model performance and implementation effort, but is largely independent of the data. Perhaps the most surprising result is that transferring hidden states for correctly initializing the model on subsequences often leads to unstable training behavior depending on the dataset.

We consider the use of Deep Learning methods for modeling complex phenomena like those occurring in natural physical processes. With the large amount of data gathered on these phenomena the data intensive paradigm could begin to challenge more traditional approaches elaborated over the years in fields like maths or physics. However, despite considerable successes in a variety of application domains, the machine learning field is not yet ready to handle the level of complexity required by such problems. Using an example application, namely Sea Surface Temperature Prediction, we show how general background knowledge gained from physics could be used as a guideline for designing efficient Deep Learning models. In order to motivate the approach and to assess its generality we demonstrate a formal link between the solution of a class of differential equations underlying a large family of physical phenomena and the proposed model. Experiments and comparison with series of baselines including a state of the art numerical approach is then provided.