The recently proposed Thermodynamic Variational Objective (TVO) leverages thermodynamic integration to provide a family of variational inference objectives, which both tighten and generalize the ubiquitous Evidence Lower Bound (ELBO). However, the tightness of TVO bounds was not previously known, an expensive grid search was used to choose a "schedule" of intermediate distributions, and model learning suffered with ostensibly tighter bounds. In this work, we propose an exponential family interpretation of the geometric mixture curve underlying the TVO and various path sampling methods, which allows us to characterize the gap in TVO likelihood bounds as a sum of KL divergences. We propose to choose intermediate distributions using equal spacing in the moment parameters of our exponential family, which matches grid search performance and allows the schedule to adaptively update over the course of training. Finally, we derive a doubly reparameterized gradient estimator which improves model learning and allows the TVO to benefit from more refined bounds. To further contextualize our contributions, we provide a unified framework for understanding thermodynamic integration and the TVO using Taylor series remainders.

Accurate and scalable hydrologic models are essential building blocks of several important applications, from water resource management to timely flood warnings. However, as the climate changes, precipitation and rainfall-runoff pattern variations become more extreme, and accurate training data that can account for the resulting distributional shifts become more scarce. In this work we present a novel family of hydrologic models, called HydroNets, which leverages river network structure. HydroNets are deep neural network models designed to exploit both basin specific rainfall-runoff signals, and upstream network dynamics, which can lead to improved predictions at longer horizons. The injection of the river structure prior knowledge reduces sample complexity and allows for scalable and more accurate hydrologic modeling even with only a few years of data. We present an empirical study over two large basins in India that convincingly support the proposed model and its advantages.

We introduce a method to speed up training by 2x and inference by 3x in deep neural networks using structured pruning applied before training. Unlike previous works on pruning before training which prune individual weights, our work develops a methodology to remove entire channels and hidden units with the explicit aim of speeding up training and inference. We introduce a compute-aware scoring mechanism which enables pruning in units of sensitivity per FLOP removed, allowing even greater speed ups. Our method is fast, easy to implement, and needs just one forward/backward pass on a single batch of data to complete pruning before training begins.

Conventional predictive Artificial Neural Networks (ANNs) commonly employ deterministic weight matrices; therefore, their prediction is a point estimate. Such a deterministic nature in ANNs causes the limitations of using ANNs for medical diagnosis, law problems, and portfolio management, in which discovering not only the prediction but also the uncertainty of the prediction is essentially required. To address such a problem, we propose a predictive probabilistic neural network model, which corresponds to a different manner of using the generator in conditional Generative Adversarial Network (cGAN) that has been routinely used for conditional sample generation. By reversing the input and output of ordinary cGAN, the model can be successfully used as a predictive model; besides, the model is robust against noises since adversarial training is employed. In addition, to measure the uncertainty of predictions, we introduce the entropy and relative entropy for regression problems and classification problems, respectively. The proposed framework is applied to stock market data and an image classification task. As a result, the proposed framework shows superior estimation performance, especially on noisy data; moreover, it is demonstrated that the proposed framework can properly estimate the uncertainty of predictions.

Any forecasting or an estimation technique used for the prediction of a signal always leads to some level of random variations in the values of the prediction at each time point. A statistical way to address such variation is by providing a confidence band of that predicted values. All statistical signal processing techniques address the issue while making their prediction. Machine learning approaches do not have any well-studied framework to address this critical problem. One of the reasons is such construction of confidence interval requires appropriate modeling of signal noise or asymptotic properties of the estimated parameters.

Deep neural networks (DNNs) exhibit great success on many tasks with the help of large-scale well annotated datasets. However, labeling large-scale data can be very costly and error-prone so that it is difficult to guarantee the annotation quality (i.e., having noisy labels). Training on these noisy labeled datasets may adversely deteriorate their generalization performance. Existing methods either rely on complex training stage division or bring too much computation for marginal performance improvement. In this paper, we propose a Temporal Calibrated Regularization (TCR), in which we utilize the original labels and the predictions in the previous epoch together to make DNN inherit the simple pattern it has learned with little overhead. We conduct extensive experiments on various neural network architectures and datasets, and find that it consistently enhances the robustness of DNNs to label noise.

This technical report describes the system participating to the Detection and Classification of Acoustic Scenes and Events (DCASE) 2020 Challenge, Task 6: automated audio captioning. Our submission focuses on solving two indeterminacy problems in automated audio captioning: word selection indeterminacy and sentence length indeterminacy. We simultaneously solve the main caption generation and sub indeterminacy problems by estimating keywords and sentence length through multi-task learning. We tested a simplified model of our submission using the development-testing dataset. Our model achieved 20.7 SPIDEr score where that of the baseline system was 5.4.

I started looking at deep learning. I've focused mainly on recurrent neural networks (RNNs), specifically LSTM because of their "unreasonable effectiveness" in the domain of Guess.js. In the same time, I started playing with convolutional neural networks (CNNs), which although less traditionally, are also often used for time series. CNNs are usually used for image classification, recognition, and detection. After playing around with CNNs, I remembered an experiment I did a few years ago, when the browser vendors introduced the getUserMedia API. In this experiment, I used the user's camera as a controller for playing a small JavaScript clone of Mortal Kombat 3. You can find the game at my GitHub account. The algorithm is so simple that I'd be able to explain it in a few sentences: The algorithm takes a snapshot of the background behind the user.

Machine learning has continued to evolve since its beginnings some seven decades ago. Learn how deep learning has catalyzed a new phase in the evolution of machine learning. Machine learning had a rich history long before deep learning reached fever pitch. Researchers and vendors were using machine learning algorithms to develop a variety of models for improving statistics, recognizing speech, predicting risk and other applications. While many of the machine learning algorithms developed over the decades are still in use today, deep learning -- a form of machine learning based on multilayered neural networks -- catalyzed a renewed interest in AI and inspired the development of better tools, processes and infrastructure for all types of machine learning.

A machine learning system is an automated process that ingests data continuously — or, at least, regularly – and that passes through a problem-solving …