ICML 2019 Papers

Nowadays, many problems require learning a model from data owned by different participants who are restricted to share their examples due to privacy concerns, which is referred to as multiparty learning in the literature. In conventional multiparty learning, a global model is usually trained from scratch via a communication protocol, ignoring the fact that each party may already have a local model trained on her own dataset. In this paper, we define a multiparty multiclass margin to measure the global behavior of a set of heterogeneous local models, and propose a general learning method called HMR (Heterogeneous Model Reuse) to optimize the margin. Our method reuses local models to approximate a global model, even when data are non-i.i.d distributed among parties, by exchanging few examples under predefined budget. Experiments on synthetic and real-world data covering different multiparty scenarios show the effectiveness of our proposal.

Reducing the variance of estimators for policy gradient has long been the focus of reinforcement learning research. While classic algorithms like REINFORCE find an $\epsilon$-approximate first-order stationary point in $\OM({1}/{\epsilon^4})$ random trajectory simulations, no provable improvement on the complexity has been made so far. This paper presents a Hessian aided policy gradient method with the first improved sample complexity of $\OM({1}/{\epsilon^3})$. While our method exploits information from the policy Hessian, it can be implemented in linear time with respect to the parameter dimension and is hence applicable to sophisticated DNN parameterization. Simulations on standard tasks validate the efficiency of our method.

We present an environment, benchmark, and deep learning driven automated theorem prover for higher-order logic. Higher-order interactive theorem provers enable the formalization of arbitrary mathematical theories and thereby present an interesting challenge for deep learning. We provide an open-source framework based on the HOL Light theorem prover that can be used as a reinforcement learning environment. HOL Light comes with a broad coverage of basic mathematical theorems on calculus and the formal proof of the Kepler conjecture, from which we derive a challenging benchmark for automated reasoning approaches. We also present a deep reinforcement learning driven automated theorem prover, DeepHOL, that gives strong initial results on this benchmark.

Many applications in machine learning require optimizing a function whose true gradient is unknown or computationally expensive, but where surrogate gradient information, directions that may be correlated with the true gradient, is cheaply available. For example, this occurs when an approximate gradient is easier to compute than the full gradient (e.g. in meta-learning or unrolled optimization), or when a true gradient is intractable and is replaced with a surrogate (e.g. in reinforcement learning or training networks with discrete variables). We propose Guided Evolutionary Strategies (GES), a method for optimally using surrogate gradient directions to accelerate random search. GES defines a search distribution for evolutionary strategies that is elongated along a subspace spanned by the surrogate gradients and estimates a descent direction which can then be passed to a first-order optimizer. We analytically and numerically characterize the tradeoffs that result from tuning how strongly the search distribution is stretched along the guiding subspace and use this to derive a setting of the hyperparameters that works well across problems. We evaluate GES on several example problems, demonstrating an improvement over both standard evolutionary strategies and first-order methods that directly follow the surrogate gradient.

Given the widespread popularity of spectral clustering (SC) for partitioning graph data, we study a version of constrained SC in which we try to incorporate the fairness notion proposed by Chierichetti et al. (2017). According to this notion, a clustering is fair if every demographic group is approximately proportionally represented in each cluster. To this end, we develop variants of both normalized and unnormalized constrained SC and show that they help find fairer clusterings on both synthetic and real data. We also provide a rigorous theoretical analysis of our algorithms on a natural variant of the stochastic block model, where $h$ groups have strong inter-group connectivity, but also exhibit a “natural” clustering structure which is fair. We prove that our algorithms can recover this fair clustering with high probability.

A novel Gromov-Wasserstein learning framework is proposed to jointly match (align) graphs and learn embedding vectors for the associated graph nodes. Using Gromov-Wasserstein discrepancy, we measure the dissimilarity between two graphs and find their correspondence, according to the learned optimal transport. The node embeddings associated with the two graphs are learned under the guidance of the optimal transport, the distance of which not only reflects the topological structure of each graph but also yields the correspondence across the graphs. These two learning steps are mutually-beneficial, and are unified here by minimizing the Gromov-Wasserstein discrepancy with structural regularizers. This framework leads to an optimization problem that is solved by a proximal point method. We apply the proposed method to matching problems in real-world networks, and demonstrate its superior performance compared to alternative approaches.

We consider the problem of multi-agent reinforcement learning (MARL) in video game AI, where the agents are located in a spatial grid-world environment and the number of agents varies both within and across episodes. The challenge is to flexibly control an arbitrary number of agents while achieving effective collaboration. Existing MARL methods usually suffer from the trade-off between these two considerations. To address the issue, we propose a novel architecture that learns a spatial joint representation of all the agents and outputs grid-wise actions. Each agent will be controlled independently by taking the action from the grid it occupies. By viewing the state information as a grid feature map, we employ a convolutional encoder-decoder as the policy network. This architecture naturally promotes agent communication because of the large receptive field provided by the stacked convolutional layers. Moreover, the spatially shared convolutional parameters enable fast parallel exploration that the experiences discovered by one agent can be immediately transferred to others. The proposed method can be conveniently integrated with general reinforcement learning algorithms, e.g., PPO and Q-learning. We demonstrate the effectiveness of the proposed method in extensive challenging multi-agent tasks in StarCraft II.

Non-negative matrix factorization is a powerful tool for learning useful representations in the data and has been widely applied in many problems such as data mining and signal processing. Orthogonal NMF, which can improve the locality of decomposition, has drawn considerable interest in solving clustering problems in recent years. However, imposing simultaneous non-negative and orthogonal structure can be quite difficult, and so existing algorithms can only solve it approximately. To address this challenge, we propose an innovative procedure called Greedy Orthogonal Pivoting Algorithm (GOPA). The GOPA algorithm fully exploits the sparsity of non-negative orthogonal solutions to break the global problem into a series of local optimizations, in which an adaptive subset of coordinates are updated in a greedy, closed-form manner. The biggest advantage of GOPA is that it promotes exact orthogonality and provides solid empirical evidence that stronger orthogonality does contribute favorably to better clustering performance. On the other hand, we further design randomized and parallel version of GOPA, which can further reduce the computational cost and improve accuracy, making it suitable for large data.

Shallow supervised 1-hidden layer neural networks have a number of favorable properties that make them easier to interpret, analyze, and optimize than their deep counterparts, but lack their representational power. Here we use 1-hidden layer learning problems to sequentially build deep networks layer by layer, which can inherit properties from shallow networks. Contrary to previous approaches using shallow networks, we focus on problems where deep learning is reported as critical for success. We thus study CNNs on image classification tasks using the large-scale ImageNet dataset and the CIFAR-10 dataset. Using a simple set of ideas for architecture and training we find that solving sequential 1-hidden-layer auxiliary problems lead to a CNN that exceeds AlexNet performance on ImageNet. Extending this training methodology to construct individual layers by solving 2-and-3-hidden layer auxiliary problems, we obtain an 11-layer network that exceeds several members of the VGG model family on ImageNet, and can train a VGG-11 model to the same accuracy as end-to-end learning. To our knowledge, this is the first competitive alternative to end-to-end training of CNNs that can scale to ImageNet. We illustrate several interesting properties of these models and conduct a range of experiments to study the properties this training induces on the intermediate layers.

Graphs are a fundamental abstraction for modeling relational data. However, graphs are discrete and combinatorial in nature, and learning representations suitable for machine learning tasks poses statistical and computational challenges. In this work, we propose Graphite, an algorithmic framework for unsupervised learning of representations over nodes in large graphs using deep latent variable generative models. Our model parameterizes variational autoencoders (VAE) with graph neural networks, and uses a novel iterative graph refinement strategy inspired by low-rank approximations for decoding. On a wide variety of synthetic and benchmark datasets, Graphite outperforms competing approaches for the tasks of density estimation, link prediction, and node classification. Finally, we derive a theoretical connection between message passing in graph neural networks and mean-field variational inference.

Many privacy mechanisms reveal high-level information about a data distribution through noisy measurements. It is common to use this information to estimate the answers to new queries. In this work, we provide an approach to solve this estimation problem efficiently using graphical models, which is particularly effective when the distribution is high-dimensional but the measurements are over low-dimensional marginals. We show that our approach is far more efficient than existing estimation techniques from the privacy literature and that it can improve the accuracy and scalability of many state-of-the-art mechanisms.

We consider the problem of representation learning for graph data. Convolutional neural networks can naturally operate on images, but have significant challenges in dealing with graph data. Given images are special cases of graphs with nodes lie on 2D lattices, graph embedding tasks have a natural correspondence with image pixel-wise prediction tasks such as segmentation. While encoder-decoder architectures like U-Nets have been successfully applied on many image pixel-wise prediction tasks, similar methods are lacking for graph data. This is due to the fact that pooling and up-sampling operations are not natural on graph data. To address these challenges, we propose novel graph pooling (gPool) and unpooling (gUnpool) operations in this work. The gPool layer adaptively selects some nodes to form a smaller graph based on their scalar projection values on a trainable projection vector. We further propose the gUnpool layer as the inverse operation of the gPool layer. The gUnpool layer restores the graph into its original structure using the position information of nodes selected in the corresponding gPool layer. Based on our proposed gPool and gUnpool layers, we develop an encoder-decoder model on graph, known as the graph U-Nets. Our experimental results on node classification and graph classification tasks demonstrate that our methods achieve consistently better performance than previous models.

We consider the problem of learning the qualities of a collection of items by performing noisy comparisons among them. Following the standard paradigm, we assume there is a fixed “comparison graph” and every neighboring pair of items in this graph is compared k times according to the Bradley-Terry-Luce model (where the probability than an item wins a comparison is proportional the item quality). We are interested in how the relative error in quality estimation scales with the comparison graph in the regime where k is large. We show that, asymptotically, the relevant graph-theoretic quantity is the square root of the resistance of the comparison graph. Specifically, we provide an algorithm with relative error decay that scales with the square root of the graph resistance, and provide a lower bound showing that (up to log factors) a better scaling is impossible. The performance guarantee of our algorithm, both in terms of the graph and the skewness of the item quality distribution, significantly outperforms earlier results.

Music score is often handled as one-dimensional sequential data. Unlike words in a text document, notes in music score can be played simultaneously by the polyphonic nature and each of them has its own duration. In this paper, we represent the unique form of musical score using graph neural network and apply it for rendering expressive piano performance from the music score. Specifically, we design the model using note-level gated graph neural network and measure-level hierarchical attention network with bidirectional long short-term memory with an iterative feedback method. In addition, to model different styles of performance for a given input score, we employ a variational auto-encoder. The result of the listening test shows that our proposed model generated more human-like performances compared to a baseline model and a hierarchical attention network model that handles music score as a word-like sequence.

This paper addresses the challenging problem of retrieval and matching of graph structured objects, and makes two key contributions. First, we demonstrate how Graph Neural Networks (GNN), which have emerged as an effective model for various supervised prediction problems defined on structured data, can be trained to produce embedding of graphs in vector spaces that enables efficient similarity reasoning. Second, we propose a novel Graph Matching Network model that, given a pair of graphs as input, computes a similarity score between them by jointly reasoning on the pair through a new cross-graph attention-based matching mechanism. We demonstrate the effectiveness of our models on different domains including the challenging problem of control-flow graph based function similarity search that plays an important role in the detection of vulnerabilities in software systems. The experimental analysis demonstrates that our models are not only able to exploit structure in the context of similarity learning but they can also outperform domain specific baseline systems that have been carefully hand-engineered for these problems.

We explore the use of graph neural networks (GNNs) to model spatial processes in which there is no a priori graphical structure. Similar to finite element analysis, we assign nodes of a GNN to spatial locations and use a computational process defined on the graph to model the relationship between an initial function defined over a space and a resulting function in the same space. We use GNNs as a computational substrate, and show that the locations of the nodes in space as well as their connectivity can be optimized to focus on the most complex parts of the space. Moreover, this representational strategy allows the learned input-output relationship to generalize over the size of the underlying space and run the same model at different levels of precision, trading computation for accuracy. We demonstrate this method on a traditional PDE problem, a physical prediction problem from robotics, and learning to predict scene images from novel viewpoints.

We propose a novel Bayesian nonparametric method to learn translation-invariant relationships on non-Euclidean domains. The resulting graph convolutional Gaussian processes can be applied to problems in machine learning for which the input observations are functions with domains on general graphs. The structure of these models allows for high dimensional inputs while retaining expressibility, as is the case with convolutional neural networks. We present applications of graph convolutional Gaussian processes to images and triangular meshes, demonstrating their versatility and effectiveness, comparing favorably to existing methods, despite being relatively simple models.

Gradient descent finds a global minimum in training deep neural networks despite the objective function being non-convex. The current paper proves gradient descent achieves zero training loss in polynomial time for a deep over-parameterized neural network with residual connections (ResNet). Our analysis relies on the particular structure of the Gram matrix induced by the neural network architecture. This structure allows us to show the Gram matrix is stable throughout the training process and this stability implies the global optimality of the gradient descent algorithm. We further extend our analysis to deep residual convolutional neural networks and obtain a similar convergence result.

Stochastic variational inference is an established way to carry out approximate Bayesian inference for deep models flexibly and at scale. While there have been effective proposals for good initializations for loss minimization in deep learning, far less attention has been devoted to the issue of initialization of stochastic variational inference. We address this by proposing a novel layer-wise initialization strategy based on Bayesian linear models. The proposed method is extensively validated on regression and classification tasks, including Bayesian Deep Nets and Conv Nets, showing faster and better convergence compared to alternatives inspired by the literature on initializations for loss minimization.