A fundamental component of neural network verification is the computation of bounds on the values their outputs can take. Previous methods have either used off-the-shelf solvers, discarding the problem structure, or relaxed the problem even further, making the bounds unnecessarily loose. We propose a novel approach based on Lagrangian Decomposition. Our formulation admits an efficient supergradient ascent algorithm, as well as an improved proximal algorithm. Both the algorithms offer three advantages: (i) they yield bounds that are provably at least as tight as previous dual algorithms relying on Lagrangian relaxations; (ii) they are based on operations analogous to forward/backward pass of neural networks layers and are therefore easily parallelizable, amenable to GPU implementation and able to take advantage of the convolutional structure of problems; and (iii) they allow for anytime stopping while still providing valid bounds. Empirically, we show that we obtain bounds comparable with off-the-shelf solvers in a fraction of their running time, and obtain tighter bounds in the same time as previous dual algorithms. This results in an overall speed-up when employing the bounds for formal verification.

Neural architecture search (NAS) relies on a good controller to generate better architectures or predict the accuracy of given architectures. However, training the controller requires both abundant and high-quality pairs of architectures and their accuracy, while it is costly to evaluate an architecture and obtain its accuracy. In this paper, we propose \emph{SemiNAS}, a semi-supervised NAS approach that leverages numerous unlabeled architectures~(without evaluation and thus nearly no cost) to improve the controller. Specifically, SemiNAS 1) trains an initial controller with a small set of architecture-accuracy data pairs; 2) uses the trained controller to predict the accuracy of large amount of architectures~(without evaluation); and 3) adds the generated data pairs to the original data to further improve the controller. SemiNAS has two advantages: 1) It reduces the computational cost under the same accuracy guarantee. 2) It achieves higher accuracy under the same computational cost. On NASBench-101 benchmark dataset, it discovers a top 0.01% architecture after evaluating roughly 300 architectures, with only 1/7 computational cost compared with regularized evolution and gradient-based methods. On ImageNet, it achieves 24.2% top-1 error rate (under the mobile setting) using 4 GPU-days for search. We further apply it to LJSpeech text to speech task and it achieves 97% intelligibility rate in the low-resource setting and 15% test error rate in the robustness setting, with 9%, 7% improvements over the baseline respectively. Our code is available at https://github.com/renqianluo/SemiNAS.

Standard methods and theories in finance can be ill-equipped to capture highly non-linear interactions in financial prediction problems based on large-scale datasets, with deep learning offering a way to gain insights into correlations in markets as complex systems. In this paper, we apply deep learning to econometrically constructed gradients to learn and exploit lagged correlations among S&P 500 stocks to compare model behaviour in stable and volatile market environments, and under the exclusion of target stock information for predictions. In order to measure the effect of time horizons, we predict intraday and daily stock price movements in varying interval lengths and gauge the complexity of the problem at hand with a modification of our model architecture. Our findings show that accuracies, while remaining significant and demonstrating the exploitability of lagged correlations in stock markets, decrease with shorter prediction horizons. We discuss implications for modern finance theory and our work's applicability as an investigative tool for portfolio managers. Lastly, we show that our model's performance is consistent in volatile markets by exposing it to the environment of the recent financial crisis of 2007/2008.

We propose a novel biologically-plausible solution to the credit assignment problem, being motivated by observations in the ventral visual pathway and trained deep neural networks. In both, representations of objects in the same category become progressively more similar, while objects belonging to different categories becomes less similar. We use this observation to motivate a layer-specific learning goal in a deep network: each layer aims to learn a representational similarity matrix that interpolates between previous and later layers. We formulate this idea using a supervised deep similarity matching cost function and derive from it deep neural networks with feedforward, lateral and feedback connections, and neurons that exhibit biologically-plausible Hebbian and anti-Hebbian plasticity. Supervised deep similarity matching can be interpreted as an energy-based learning algorithm, but with significant differences from others in how a contrastive function is constructed.

Learning rate schedule has a major impact on the performance of deep learning models. Still, the choice of a schedule is often heuristical. We aim to develop a precise understanding of the effects of different learning rate schedules and the appropriate way to select them. To this end, we isolate two distinct phases of training, the first, which we refer to as the "large-step" regime, exhibits a rather poor performance from an optimization point of view but is the primary contributor to model generalization; the latter, "small-step" regime exhibits much more "convex-like" optimization behavior but used in isolation produces models that generalize poorly. We find that by treating these regimes separately-and em specializing our training algorithm to each one of them, we can significantly simplify learning rate schedules.

We propose self-adaptive training---a new training algorithm that dynamically corrects problematic training labels by model predictions without incurring extra computational cost---to improve generalization of deep learning for potentially corrupted training data. This problem is crucial towards robustly learning from data that are corrupted by, e.g., label noises and out-of-distribution samples. The standard empirical risk minimization (ERM) for such data, however, may easily overfit noises and thus suffers from sub-optimal performance. In this paper, we observe that model predictions can substantially benefit the training process: self-adaptive training significantly improves generalization over ERM under various levels of noises, and mitigates the overfitting issue in both natural and adversarial training. We evaluate the error-capacity curve of self-adaptive training: the test error is monotonously decreasing w.r.t. model capacity. This is in sharp contrast to the recently-discovered double-descent phenomenon in ERM which might be a result of overfitting of noises. Experiments on CIFAR and ImageNet datasets verify the effectiveness of our approach in two applications: classification with label noise and selective classification. We release our code at \url{https://github.com/LayneH/self-adaptive-training}.

Positive-unlabeled (PU) learning trains a binary classifier using only positive and unlabeled data. A common simplifying assumption is that the positive data is representative of the target positive class. This assumption is often violated in practice due to time variation, domain shift, or adversarial concept drift. This paper shows that PU learning is possible even with arbitrarily non-representative positive data when provided unlabeled datasets from the source and target distributions. Our key insight is that only the negative class's distribution need be fixed. We propose two methods to learn under such arbitrary positive bias. The first couples negative-unlabeled (NU) learning with unlabeled-unlabeled (UU) learning while the other uses a novel recursive risk estimator robust to positive shift. Experimental results demonstrate our methods' effectiveness across numerous real-world datasets and forms of positive data bias, including disjoint positive class-conditional supports.

Deep learning image classifiers usually rely on huge training sets and their training process can be described as learning the similarities and differences among training images. But, images in large training sets are not usually studied from this perspective and fine-level similarities and differences among images is usually overlooked. Some studies aim to identify the influential and redundant training images, but such methods require a model that is already trained on the entire training set. Here, we show that analyzing the contents of large training sets can provide valuable insights about the classification task at hand, prior to training a model on them. We use wavelet decomposition of images and other image processing tools to perform such analysis, with no need for a pre-trained model. This makes the analysis of training sets, straightforward and fast. We show that similar images in standard datasets (such as CIFAR) can be identified in a few seconds, a significant speed-up compared to alternative methods in the literature. We also show that similarities between training and testing images may explain the generalization of models and their mistakes. Finally, we investigate the similarities between images in relation to decision boundaries of a trained model.

Representing shapes as level sets of neural networks has been recently proved to be useful for different shape analysis and reconstruction tasks So far, such representations were computed using either: (i) pre-computed implicit shape representations; or (ii) loss functions explicitly defined over the neural level sets. In this paper we offer a new paradigm for computing high fidelity implicit neural representations directly from raw data (i.e., point clouds, with or without normal information). We observe that a rather simple loss function, encouraging the neural network to vanish on the input point cloud and to have a unit norm gradient, possesses an implicit geometric regularization property that favors smooth and natural zero level set surfaces, avoiding bad zero-loss solutions. We provide a theoretical analysis of this property for the linear case, and show that, in practice, our method leads to state of the art implicit neural representations with higher level-of-details and fidelity compared to previous methods.

The complexity of large-scale neural networks can lead to poor understanding of their internal details. We show that this opaqueness provides an opportunity for adversaries to embed unintended functionalities into the network in the form of Trojan horses. Our novel framework hides the existence of a Trojan network with arbitrary desired functionality within a benign transport network. We prove theoretically that the Trojan network's detection is computationally infeasible and demonstrate empirically that the transport network does not compromise its disguise. Our paper exposes an important, previously unknown loophole that could potentially undermine the security and trustworthiness of machine learning.