A prominent approach is to generate music symbolically in the form of a piano roll, which specifies the timing, pitch, velocity, and instrument of each note to be played. This has led to impressive results like producing Bach chorals, polyphonic music with multiple instruments, as well as minute long musical pieces. But symbolic generators have limitations--they cannot capture human voices or many of the more subtle timbres, dynamics, and expressivity that are essential to music. A different approach[1] is to model music directly as raw audio. Generating music at the audio level is challenging since the sequences are very long.

In November 2007, Google laid the basis to dominate the mobile market by launching Android, an open source operating system for smartphones. After eight years to the month, Android has an 80% market share and Google is utilizing a similar stunt, this time with artificial intelligence. Google later announced TensorFlow, its open source platform for machine learning, giving anybody a computer, internet connection and casual background in deep learning algorithms access to one of the most powerful machine learning platforms ever made. More than 50 Google products have embraced TensorFlow to harness deep learning (machine learning utilizing deep neural networks) as a tool, from distinguishing you and your companions in the Photos application to refining its core search engine. Google has become a machine learning organization.

Morse S. How health insurance companies use AI to make consumers healthier.

The fragility of deep neural networks to adversarially-chosen inputs has motivated the need to revisit deep learning algorithms. Including adversarial examples during training is a popular defense mechanism against adversarial attacks. This mechanism can be formulated as a min-max optimization problem, where the adversary seeks to maximize the loss function using an iterative first-order algorithm while the learner attempts to minimize it. However, finding adversarial examples in this way causes excessive computational overhead during training. By interpreting the min-max problem as an optimal control problem, it has recently been shown that one can exploit the compositional structure of neural networks in the optimization problem to improve the training time significantly. In this paper, we provide the first convergence analysis of this adversarial training algorithm by combining techniques from robust optimal control and inexact oracle methods in optimization. Our analysis sheds light on how the hyperparameters of the algorithm affect the its stability and convergence. We support our insights with experiments on a robust classification problem.

Data sparsity is a common issue to train machine learning tools such as neural networks for engineering and scientific applications, where experiments and simulations are expensive. Recently physics-constrained neural networks (PCNNs) were developed to reduce the required amount of training data. However, the weights of different losses from data and physical constraints are adjusted empirically in PCNNs. In this paper, a new physics-constrained neural network with the minimax architecture (PCNNMM) is proposed so that the weights of different losses can be adjusted systematically. The training of the PCNN-MM is searching the high-order saddle points of the objective function. A novel saddle point search algorithm called Dual-Dimer method is developed. It is demonstrated that the Dual-Dimer method is computationally more efficient than the gradient descent ascent method for nonconvex-nonconcave functions and provides additional eigenvalue information to verify search results. A heat transfer example also shows that the convergence of PCNN-MMs is faster than that of traditional PCNNs.

Multi-label image classification has generated significant interest in recent years and the performance of such systems often suffers from the not so infrequent occurrence of incorrect or missing labels in the training data. In this paper, we extend the state-of the-art of training classifiers to jointly deal with both forms of errorful data. We accomplish this by modeling noisy and missing labels in multi-label images with a new Noise Modeling Network (NMN) that follows our convolutional neural network (CNN), integrates with it, forming an end-to-end deep learning system, which can jointly learn the noise distribution and CNN parameters. The NMN learns the distribution of noise patterns directly from the noisy data without the need for any clean training data. The NMN can model label noise that depends only on the true label or is also dependent on the image features. We show that the integrated NMN/CNN learning system consistently improves the classification performance, for different levels of label noise, on the MSR-COCO dataset and MSR-VTT dataset. We also show that noise performance improvements are obtained when multiple instance learning methods are used.

While typical named entity recognition (NER) models require the training set to be annotated with all target types, each available datasets may only cover a part of them. Instead of relying on fully-typed NER datasets, many efforts have been made to leverage multiple partially-typed ones for training and allow the resulting model to cover a full type set. However, there is neither guarantee on the quality of integrated datasets, nor guidance on the design of training algorithms. Here, we conduct a systematic analysis and comparison between partially-typed NER datasets and fully-typed ones, in both theoretical and empirical manner. Firstly, we derive a bound to establish that models trained with partially-typed annotations can reach a similar performance with the ones trained with fully-typed annotations, which also provides guidance on the algorithm design. Moreover, we conduct controlled experiments, which shows partially-typed datasets leads to similar performance with the model trained with the same amount of fully-typed annotations.

Federated learning aims to protect data privacy by collaboratively learning a model without sharing private data among users. However, an adversary may still be able to infer the private training data by attacking the released model. Differential privacy(DP) provides a statistical guarantee against such attacks, at a privacy of possibly degenerating the accuracy or utility of the trained models. In this paper, we apply a utility enhancement scheme based on Laplacian smoothing for differentially-private federated learning (DP-Fed-LS), where the parameter aggregation with injected Gaussian noise is improved in statistical precision. We provide tight closed-form privacy bounds for both uniform and Poisson subsampling and derive corresponding DP guarantees for differential private federated learning, with or without Laplacian smoothing. Experiments over MNIST, SVHN and Shakespeare datasets show that the proposed method can improve model accuracy with DP-guarantee under both subsampling mechanisms.

Inference capabilities of machine learning (ML) systems skyrocketed in recent years, now playing a pivotal role in various aspect of society. The goal in statistical learning is to use data to obtain simple algorithms for predicting a random variable $Y$ from a correlated observation $X$. Since the dimension of $X$ is typically huge, computationally feasible solutions should summarize it into a lower-dimensional feature vector $T$, from which $Y$ is predicted. The algorithm will successfully make the prediction if $T$ is a good proxy of $Y$, despite the said dimensionality-reduction. A myriad of ML algorithms (mostly employing deep learning (DL)) for finding such representations $T$ based on real-world data are now available. While these methods are often effective in practice, their success is hindered by the lack of a comprehensive theory to explain it. The information bottleneck (IB) theory recently emerged as a bold information-theoretic paradigm for analyzing DL systems. Adopting mutual information as the figure of merit, it suggests that the best representation $T$ should be maximally informative about $Y$ while minimizing the mutual information with $X$. In this tutorial we survey the information-theoretic origins of this abstract principle, and its recent impact on DL. For the latter, we cover implications of the IB problem on DL theory, as well as practical algorithms inspired by it. Our goal is to provide a unified and cohesive description. A clear view of current knowledge is particularly important for further leveraging IB and other information-theoretic ideas to study DL models.