Bryan Perozzi

Oleksandr Ferludin, Arno Eigenwillig, Martin Blais et al. · deepmind

TensorFlow-GNN (TF-GNN) is a scalable library for Graph Neural Networks in TensorFlow. It is designed from the bottom up to support the kinds of rich heterogeneous graph data that occurs in today's information ecosystems. In addition to enabling machine learning researchers and advanced developers, TF-GNN offers low-code solutions to empower the broader developer community in graph learning. Many production models at Google use TF-GNN, and it has been recently released as an open source project. In this paper we describe the TF-GNN data model, its Keras message passing API, and relevant capabilities such as graph sampling and distributed training.

Phitchaya Mangpo Phothilimthana, Sami Abu-El-Haija, Kaidi Cao et al. · stanford

Precise hardware performance models play a crucial role in code optimizations. They can assist compilers in making heuristic decisions or aid autotuners in identifying the optimal configuration for a given program. For example, the autotuner for XLA, a machine learning compiler, discovered 10-20% speedup on state-of-the-art models serving substantial production traffic at Google. Although there exist a few datasets for program performance prediction, they target small sub-programs such as basic blocks or kernels. This paper introduces TpuGraphs, a performance prediction dataset on full tensor programs, represented as computational graphs, running on Tensor Processing Units (TPUs). Each graph in the dataset represents the main computation of a machine learning workload, e.g., a training epoch or an inference step. Each data sample contains a computational graph, a compilation configuration, and the execution time of the graph when compiled with the configuration. The graphs in the dataset are collected from open-source machine learning programs, featuring popular model architectures, e.g., ResNet, EfficientNet, Mask R-CNN, and Transformer. TpuGraphs provides 25x more graphs than the largest graph property prediction dataset (with comparable graph sizes), and 770x larger graphs on average compared to existing performance prediction datasets on machine learning programs. This graph-level prediction task on large graphs introduces new challenges in learning, ranging from scalability, training efficiency, to model quality.

Bahare Fatemi, Sami Abu-El-Haija, Anton Tsitsulin et al.

Graph neural networks (GNNs) demonstrate outstanding performance in a broad range of applications. While the majority of GNN applications assume that a graph structure is given, some recent methods substantially expanded the applicability of GNNs by showing that they may be effective even when no graph structure is explicitly provided. The GNN parameters and a graph structure are jointly learned. Previous studies adopt different experimentation setups, making it difficult to compare their merits. In this paper, we propose a benchmarking strategy for graph structure learning using a unified framework. Our framework, called Unified Graph Structure Learning (UGSL), reformulates existing models into a single model. We implement a wide range of existing models in our framework and conduct extensive analyses of the effectiveness of different components in the framework. Our results provide a clear and concise understanding of the different methods in this area as well as their strengths and weaknesses. The benchmark code is available at https://github.com/google-research/google-research/tree/master/ugsl.

Minji Yoon, John Palowitch, Dustin Zelle et al.

Data continuously emitted from industrial ecosystems such as social or e-commerce platforms are commonly represented as heterogeneous graphs (HG) composed of multiple node/edge types. State-of-the-art graph learning methods for HGs known as heterogeneous graph neural networks (HGNNs) are applied to learn deep context-informed node representations. However, many HG datasets from industrial applications suffer from label imbalance between node types. As there is no direct way to learn using labels rooted at different node types, HGNNs have been applied to only a few node types with abundant labels. We propose a zero-shot transfer learning module for HGNNs called a Knowledge Transfer Network (KTN) that transfers knowledge from label-abundant node types to zero-labeled node types through rich relational information given in the HG. KTN is derived from the theoretical relationship, which we introduce in this work, between distinct feature extractors for each node type given in an HGNN model. KTN improves performance of 6 different types of HGNN models by up to 960% for inference on zero-labeled node types and outperforms state-of-the-art transfer learning baselines by up to 73% across 18 different transfer learning tasks on HGs.

Bahare Fatemi, Jonathan Halcrow, Bryan Perozzi

Graphs are a powerful tool for representing and analyzing complex relationships in real-world applications such as social networks, recommender systems, and computational finance. Reasoning on graphs is essential for drawing inferences about the relationships between entities in a complex system, and to identify hidden patterns and trends. Despite the remarkable progress in automated reasoning with natural text, reasoning on graphs with large language models (LLMs) remains an understudied problem. In this work, we perform the first comprehensive study of encoding graph-structured data as text for consumption by LLMs. We show that LLM performance on graph reasoning tasks varies on three fundamental levels: (1) the graph encoding method, (2) the nature of the graph task itself, and (3) interestingly, the very structure of the graph considered. These novel results provide valuable insight on strategies for encoding graphs as text. Using these insights we illustrate how the correct choice of encoders can boost performance on graph reasoning tasks inside LLMs by 4.8% to 61.8%, depending on the task.

Minji Yoon, Yue Wu, John Palowitch et al.

As the field of Graph Neural Networks (GNN) continues to grow, it experiences a corresponding increase in the need for large, real-world datasets to train and test new GNN models on challenging, realistic problems. Unfortunately, such graph datasets are often generated from online, highly privacy-restricted ecosystems, which makes research and development on these datasets hard, if not impossible. This greatly reduces the amount of benchmark graphs available to researchers, causing the field to rely only on a handful of publicly-available datasets. To address this problem, we introduce a novel graph generative model, Computation Graph Transformer (CGT) that learns and reproduces the distribution of real-world graphs in a privacy-controlled way. More specifically, CGT (1) generates effective benchmark graphs on which GNNs show similar task performance as on the source graphs, (2) scales to process large-scale graphs, (3) incorporates off-the-shelf privacy modules to guarantee end-user privacy of the generated graph. Extensive experiments across a vast body of graph generative models show that only our model can successfully generate privacy-controlled, synthetic substitutes of large-scale real-world graphs that can be effectively used to benchmark GNN models.

Alessandro Epasto, Vahab Mirrokni, Bryan Perozzi et al.

Personalized PageRank (PPR) is a fundamental tool in unsupervised learning of graph representations such as node ranking, labeling, and graph embedding. However, while data privacy is one of the most important recent concerns, existing PPR algorithms are not designed to protect user privacy. PPR is highly sensitive to the input graph edges: the difference of only one edge may cause a big change in the PPR vector, potentially leaking private user data. In this work, we propose an algorithm which outputs an approximate PPR and has provably bounded sensitivity to input edges. In addition, we prove that our algorithm achieves similar accuracy to non-private algorithms when the input graph has large degrees. Our sensitivity-bounded PPR directly implies private algorithms for several tools of graph learning, such as, differentially private (DP) PPR ranking, DP node classification, and DP node embedding. To complement our theoretical analysis, we also empirically verify the practical performances of our algorithms.

Anton Tsitsulin, Benedek Rozemberczki, John Palowitch et al.

Graph learning algorithms have attained state-of-the-art performance on many graph analysis tasks such as node classification, link prediction, and clustering. It has, however, become hard to track the field's burgeoning progress. One reason is due to the very small number of datasets used in practice to benchmark the performance of graph learning algorithms. This shockingly small sample size (~10) allows for only limited scientific insight into the problem. In this work, we aim to address this deficiency. We propose to generate synthetic graphs, and study the behaviour of graph learning algorithms in a controlled scenario. We develop a fully-featured synthetic graph generator that allows deep inspection of different models. We argue that synthetic graph generations allows for thorough investigation of algorithms and provides more insights than overfitting on three citation datasets. In the case study, we show how our framework provides insight into unsupervised and supervised graph neural network models.

Mehran Kazemi, Anton Tsitsulin, Hossein Esfandiari et al.

Representative Selection (RS) is the problem of finding a small subset of exemplars from a dataset that is representative of the dataset. In this paper, we study RS for attributed graphs, and focus on finding representative nodes that optimize the accuracy of a model trained on the selected representatives. Theoretically, we establish a new hardness result forRS (in the absence of a graph structure) by proving that a particular, highly practical variant of it (RS for Learning) is hard to approximate in polynomial time within any reasonable factor, which implies a significant potential gap between the optimum solution of widely-used surrogate functions and the actual accuracy of the model. We then study the setting where a (homophilous) graph structure is available, or can be constructed, between the data points.We show that with an appropriate modeling approach, the presence of such a structure can turn a hard RS (for learning) problem into one that can be effectively solved. To this end, we develop RS-GNN, a representation learning-based RS model based on Graph Neural Networks. Empirically, we demonstrate the effectiveness of RS-GNN on problems with predefined graph structures as well as problems with graphs induced from node feature similarities, by showing that RS-GNN achieves significant improvements over established baselines on a suite of eight benchmarks.

Qi Zhu, Yizhu Jiao, Natalia Ponomareva et al.

Graph Neural Networks (GNNs) have shown remarkable performance on graph-structured data. However, recent empirical studies suggest that GNNs are very susceptible to distribution shift. There is still significant ambiguity about why graph-based models seem more vulnerable to these shifts. In this work we provide a thorough theoretical analysis on it by quantifying the magnitude of conditional shift between the input features and the output label. Our findings show that both graph heterophily and model architecture exacerbate conditional shifts, leading to performance degradation. To address this, we propose an approach that involves estimating and minimizing the conditional shift for unsupervised domain adaptation on graphs. In our controlled synthetic experiments, our algorithm demonstrates robustness towards distribution shift, resulting in up to 10% absolute ROC AUC improvement versus the second-best algorithm. Furthermore, comprehensive experiments on both node classification and graph classification show its robust performance under various distribution shifts.

Mustafa Yasir, John Palowitch, Anton Tsitsulin et al.

Despite a surge in interest in GNN development, homogeneity in benchmarking datasets still presents a fundamental issue to GNN research. GraphWorld is a recent solution which uses the Stochastic Block Model (SBM) to generate diverse populations of synthetic graphs for benchmarking any GNN task. Despite its success, the SBM imposed fundamental limitations on the kinds of graph structure GraphWorld could create. In this work we examine how two additional synthetic graph generators can improve GraphWorld's evaluation; LFR, a well-established model in the graph clustering literature and CABAM, a recent adaptation of the Barabasi-Albert model tailored for GNN benchmarking. By integrating these generators, we significantly expand the coverage of graph space within the GraphWorld framework while preserving key graph properties observed in real-world networks. To demonstrate their effectiveness, we generate 300,000 graphs to benchmark 11 GNN models on a node classification task. We find GNN performance variations in response to homophily, degree distribution and feature signal. Based on these findings, we classify models by their sensitivity to the new generators under these properties. Additionally, we release the extensions made to GraphWorld on the GitHub repository, offering further evaluation of GNN performance on new graphs.

Kimon Fountoulakis, Dake He, Silvio Lattanzi et al.

The recent years we have seen the rise of graph neural networks for prediction tasks on graphs. One of the dominant architectures is graph attention due to its ability to make predictions using weighted edge features and not only node features. In this paper we analyze, theoretically and empirically, graph attention networks and their ability of correctly labelling nodes in a classic classification task. More specifically, we study the performance of graph attention on the classic contextual stochastic block model (CSBM). In CSBM the nodes and edge features are obtained from a mixture of Gaussians and the edges from a stochastic block model. We consider a general graph attention mechanism that takes random edge features as input to determine the attention coefficients. We study two cases, in the first one, when the edge features are noisy, we prove that the majority of the attention coefficients are up to a constant uniform. This allows us to prove that graph attention with edge features is not better than simple graph convolution for achieving perfect node classification. Second, we prove that when the edge features are clean graph attention can distinguish intra- from inter-edges and this makes graph attention better than classic graph convolution.

Timo Stoll, Chendi Qian, Ben Finkelshtein et al.

Machine learning on graphs has recently achieved impressive progress in various domains, including molecular property prediction and chip design. However, benchmarking practices remain fragmented, often relying on narrow, task-specific datasets and inconsistent evaluation protocols, which hampers reproducibility and broader progress. To address this, we introduce GraphBench, a comprehensive benchmarking suite that spans diverse domains and prediction tasks, including node-level, edge-level, graph-level, and generative settings. GraphBench provides standardized evaluation protocols -- with consistent dataset splits and performance metrics that account for out-of-distribution generalization -- as well as a unified hyperparameter tuning framework. Additionally, we benchmark GraphBench using message-passing neural networks and graph transformer models, providing principled baselines and establishing a reference performance. See www.graphbench.io for further details.

Brandon Mayer, Anton Tsitsulin, Hendrik Fichtenberger et al.

Graphs are a representation of structured data that captures the relationships between sets of objects. With the ubiquity of available network data, there is increasing industrial and academic need to quickly analyze graphs with billions of nodes and trillions of edges. A common first step for network understanding is Graph Embedding, the process of creating a continuous representation of nodes in a graph. A continuous representation is often more amenable, especially at scale, for solving downstream machine learning tasks such as classification, link prediction, and clustering. A high-performance graph embedding architecture leveraging Tensor Processing Units (TPUs) with configurable amounts of high-bandwidth memory is presented that simplifies the graph embedding problem and can scale to graphs with billions of nodes and trillions of edges. We verify the embedding space quality on real and synthetic large-scale datasets.

Gheorghe Comanici, Eric Bieber, Mike Schaekermann et al. · amazon-science, baidu

In this report, we introduce the Gemini 2.X model family: Gemini 2.5 Pro and Gemini 2.5 Flash, as well as our earlier Gemini 2.0 Flash and Flash-Lite models. Gemini 2.5 Pro is our most capable model yet, achieving SoTA performance on frontier coding and reasoning benchmarks. In addition to its incredible coding and reasoning skills, Gemini 2.5 Pro is a thinking model that excels at multimodal understanding and it is now able to process up to 3 hours of video content. Its unique combination of long context, multimodal and reasoning capabilities can be combined to unlock new agentic workflows. Gemini 2.5 Flash provides excellent reasoning abilities at a fraction of the compute and latency requirements and Gemini 2.0 Flash and Flash-Lite provide high performance at low latency and cost. Taken together, the Gemini 2.X model generation spans the full Pareto frontier of model capability vs cost, allowing users to explore the boundaries of what is possible with complex agentic problem solving.

Zhikai Chen, Haitao Mao, Jingzhe Liu et al.

Given the ubiquity of graph data and its applications in diverse domains, building a Graph Foundation Model (GFM) that can work well across different graphs and tasks with a unified backbone has recently garnered significant interests. A major obstacle to achieving this goal stems from the fact that graphs from different domains often exhibit diverse node features. Inspired by multi-modal models that align different modalities with natural language, the text has recently been adopted to provide a unified feature space for diverse graphs. Despite the great potential of these text-space GFMs, current research in this field is hampered by two problems. First, the absence of a comprehensive benchmark with unified problem settings hinders a clear understanding of the comparative effectiveness and practical value of different text-space GFMs. Second, there is a lack of sufficient datasets to thoroughly explore the methods' full potential and verify their effectiveness across diverse settings. To address these issues, we conduct a comprehensive benchmark providing novel text-space datasets and comprehensive evaluation under unified problem settings. Empirical results provide new insights and inspire future research directions. Our code and data are publicly available from \url{https://github.com/CurryTang/TSGFM}.

Bahare Fatemi, Mehran Kazemi, Anton Tsitsulin et al.

Large language models (LLMs) have showcased remarkable reasoning capabilities, yet they remain susceptible to errors, particularly in temporal reasoning tasks involving complex temporal logic. Existing research has explored LLM performance on temporal reasoning using diverse datasets and benchmarks. However, these studies often rely on real-world data that LLMs may have encountered during pre-training or employ anonymization techniques that can inadvertently introduce factual inconsistencies. In this work, we address these limitations by introducing novel synthetic datasets specifically designed to assess LLM temporal reasoning abilities in various scenarios. The diversity of question types across these datasets enables systematic investigation into the impact of the problem structure, size, question type, fact order, and other factors on LLM performance. Our findings provide valuable insights into the strengths and weaknesses of current LLMs in temporal reasoning tasks. To foster further research in this area, we are open-sourcing the datasets and evaluation framework used in our experiments: https://huggingface.co/datasets/baharef/ToT.

John Palowitch, Anton Tsitsulin, Brandon Mayer et al.

Despite advances in the field of Graph Neural Networks (GNNs), only a small number (~5) of datasets are currently used to evaluate new models. This continued reliance on a handful of datasets provides minimal insight into the performance differences between models, and is especially challenging for industrial practitioners who are likely to have datasets which look very different from those used as academic benchmarks. In the course of our work on GNN infrastructure and open-source software at Google, we have sought to develop improved benchmarks that are robust, tunable, scalable,and generalizable. In this work we introduce GraphWorld, a novel methodology and system for benchmarking GNN models on an arbitrarily-large population of synthetic graphs for any conceivable GNN task. GraphWorld allows a user to efficiently generate a world with millions of statistically diverse datasets. It is accessible, scalable, and easy to use. GraphWorld can be run on a single machine without specialized hardware, or it can be easily scaled up to run on arbitrary clusters or cloud frameworks. Using GraphWorld, a user has fine-grained control over graph generator parameters, and can benchmark arbitrary GNN models with built-in hyperparameter tuning. We present insights from GraphWorld experiments regarding the performance characteristics of tens of thousands of GNN models over millions of benchmark datasets. We further show that GraphWorld efficiently explores regions of benchmark dataset space uncovered by standard benchmarks, revealing comparisons between models that have not been historically obtainable. Using GraphWorld, we also are able to study in-detail the relationship between graph properties and task performance metrics, which is nearly impossible with the classic collection of real-world benchmarks.

Elan Markowitz, Keshav Balasubramanian, Mehrnoosh Mirtaheri et al.

Graph Representation Learning (GRL) methods have impacted fields from chemistry to social science. However, their algorithmic implementations are specialized to specific use-cases e.g.message passing methods are run differently from node embedding ones. Despite their apparent differences, all these methods utilize the graph structure, and therefore, their learning can be approximated with stochastic graph traversals. We propose Graph Traversal via Tensor Functionals(GTTF), a unifying meta-algorithm framework for easing the implementation of diverse graph algorithms and enabling transparent and efficient scaling to large graphs. GTTF is founded upon a data structure (stored as a sparse tensor) and a stochastic graph traversal algorithm (described using tensor operations). The algorithm is a functional that accept two functions, and can be specialized to obtain a variety of GRL models and objectives, simply by changing those two functions. We show for a wide class of methods, our algorithm learns in an unbiased fashion and, in expectation, approximates the learning as if the specialized implementations were run directly. With these capabilities, we scale otherwise non-scalable methods to set state-of-the-art on large graph datasets while being more efficient than existing GRL libraries - with only a handful of lines of code for each method specialization. GTTF and its various GRL implementations are on: https://github.com/isi-usc-edu/gttf.

Bryan Perozzi, Bahare Fatemi, Dustin Zelle et al.

How can we best encode structured data into sequential form for use in large language models (LLMs)? In this work, we introduce a parameter-efficient method to explicitly represent structured data for LLMs. Our method, GraphToken, learns an encoding function to extend prompts with explicit structured information. Unlike other work which focuses on limited domains (e.g. knowledge graph representation), our work is the first effort focused on the general encoding of structured data to be used for various reasoning tasks. We show that explicitly representing the graph structure allows significant improvements to graph reasoning tasks. Specifically, we see across the board improvements - up to 73% points - on node, edge and, graph-level tasks from the GraphQA benchmark.

Maya Bechler-Speicher, Ben Finkelshtein, Fabrizio Frasca et al. · deepmind

While machine learning on graphs has demonstrated promise in drug design and molecular property prediction, significant benchmarking challenges hinder its further progress and relevance. Current benchmarking practices often lack focus on transformative, real-world applications, favoring narrow domains like two-dimensional molecular graphs over broader, impactful areas such as combinatorial optimization, relational databases, or chip design. Additionally, many benchmark datasets poorly represent the underlying data, leading to inadequate abstractions and misaligned use cases. Fragmented evaluations and an excessive focus on accuracy further exacerbate these issues, incentivizing overfitting rather than fostering generalizable insights. These limitations have prevented the development of truly useful graph foundation models. This position paper calls for a paradigm shift toward more meaningful benchmarks, rigorous evaluation protocols, and stronger collaboration with domain experts to drive impactful and reliable advances in graph learning research, unlocking the potential of graph learning.

Mohit Agarwal, Mimi Sun, Chaitanya Kamath et al.

Supporting the health and well-being of dynamic populations around the world requires governmental agencies, organizations and researchers to understand and reason over complex relationships between human behavior and local contexts in order to identify high-risk groups and strategically allocate limited resources. Traditional approaches to these classes of problems often entail developing manually curated, task-specific features and models to represent human behavior and the natural and built environment, which can be challenging to adapt to new, or even, related tasks. To address this, we introduce a Population Dynamics Foundation Model (PDFM) that aims to capture the relationships between diverse data modalities and is applicable to a broad range of geospatial tasks. We first construct a geo-indexed dataset for postal codes and counties across the United States, capturing rich aggregated information on human behavior from maps, busyness, and aggregated search trends, and environmental factors such as weather and air quality. We then model this data and the complex relationships between locations using a graph neural network, producing embeddings that can be adapted to a wide range of downstream tasks using relatively simple models. We evaluate the effectiveness of our approach by benchmarking it on 27 downstream tasks spanning three distinct domains: health indicators, socioeconomic factors, and environmental measurements. The approach achieves state-of-the-art performance on all 27 geospatial interpolation tasks, and on 25 out of the 27 extrapolation and super-resolution tasks. We combined the PDFM with a state-of-the-art forecasting foundation model, TimesFM, to predict unemployment and poverty, achieving performance that surpasses fully supervised forecasting. The full set of embeddings and sample code are publicly available for researchers.

Florian Grötschla, Luis Müller, Jan Tönshoff et al.

Large-language models (LLMs) have demonstrated powerful problem-solving capabilities, in particular when organized in multi-agent systems. However, the advent of such systems also raises several questions on the ability of a complex network of agents to effectively self-organize and collaborate. While measuring performance on standard reasoning benchmarks indicates how well multi-agent systems can solve reasoning tasks, it is unclear whether these systems are able to leverage their topology effectively. Here, we propose AgentsNet, a new benchmark for multi-agent reasoning. By drawing inspiration from classical problems in distributed systems and graph theory, AgentsNet measures the ability of multi-agent systems to collaboratively form strategies for problem-solving, self-organization, and effective communication given a network topology. We evaluate a variety of baseline methods on AgentsNet including homogeneous networks of agents which first have to agree on basic protocols for organization and communication. We find that some frontier LLMs are already demonstrating strong performance for small networks but begin to fall off once the size of the network scales. While existing multi-agent benchmarks cover at most 2-5 agents, AgentsNet is practically unlimited in size and can scale with new generations of LLMs. As such, we also probe frontier models in a setup with up to 100 agents.

Anton Tsitsulin, Bryan Perozzi

Graph learning methods help utilize implicit relationships among data items, thereby reducing training label requirements and improving task performance. However, determining the optimal graph structure for a particular learning task remains a challenging research problem. In this work, we introduce the Graph Lottery Ticket (GLT) Hypothesis - that there is an extremely sparse backbone for every graph, and that graph learning algorithms attain comparable performance when trained on that subgraph as on the full graph. We identify and systematically study 8 key metrics of interest that directly influence the performance of graph learning algorithms. Subsequently, we define the notion of a "winning ticket" for graph structure - an extremely sparse subset of edges that can deliver a robust approximation of the entire graph's performance. We propose a straightforward and efficient algorithm for finding these GLTs in arbitrary graphs. Empirically, we observe that performance of different graph learning algorithms can be matched or even exceeded on graphs with the average degree as low as 5.

Ali Behrouz, Ali Parviz, Mahdi Karami et al.

Modern sequence models (e.g., Transformers, linear RNNs, etc.) emerged as dominant backbones of recent deep learning frameworks, mainly due to their efficiency, representational power, and/or ability to capture long-range dependencies. Adopting these sequence models for graph-structured data has recently gained popularity as the alternative to Message Passing Neural Networks (MPNNs). There is, however, a lack of a common foundation about what constitutes a good graph sequence model, and a mathematical description of the benefits and deficiencies in adopting different sequence models for learning on graphs. To this end, we first present Graph Sequence Model (GSM), a unifying framework for adopting sequence models for graphs, consisting of three main steps: (1) Tokenization, which translates the graph into a set of sequences; (2) Local Encoding, which encodes local neighborhoods around each node; and (3) Global Encoding, which employs a scalable sequence model to capture long-range dependencies within the sequences. This framework allows us to understand, evaluate, and compare the power of different sequence model backbones in graph tasks. Our theoretical evaluations of the representation power of Transformers and modern recurrent models through the lens of global and local graph tasks show that there are both negative and positive sides for both types of models. Building on this observation, we present GSM++, a fast hybrid model that uses the Hierarchical Affinity Clustering (HAC) algorithm to tokenize the graph into hierarchical sequences, and then employs a hybrid architecture of Transformer to encode these sequences. Our theoretical and experimental results support the design of GSM++, showing that GSM++ outperforms baselines in most benchmark evaluations.

Filipe Miguel Gonçalves de Almeida, CJ Carey, Hendrik Fichtenberger et al.

Learning and constructing large-scale graphs has attracted attention in recent decades, resulting in a rich literature that introduced various systems, tools, and algorithms. Grale is one of such tools that is designed for offline environments and is deployed in more than 50 different industrial settings at Google. Grale is widely applicable because of its ability to efficiently learn and construct a graph on datasets with multiple types of features. However, it is often the case that applications require the underlying data to evolve continuously and rapidly and the updated graph needs to be available with low latency. Such setting make the use of Grale prohibitive. While there are Approximate Nearest Neighbor (ANN) systems that handle dynamic updates with low latency, they are mostly limited to similarities over a single embedding. In this work, we introduce a system that inherits the advantages and the quality of Grale, and maintains a graph construction in a dynamic setting with tens of milliseconds of latency per request. We call the system Dynamic Grale Using ScaNN (Dynamic GUS). Our system has a wide range of applications with over 10 deployments at Google. One of the applications is in Android Security and Privacy, where Dynamic Grale Using ScaNN enables capturing harmful applications 4 times faster, before they can reach users.

Zixuan Yi, Sami Abu-el-Haija, Yawen Wang et al.

DB engines produce efficient query execution plans by relying on cost models. Practical implementations estimate cardinality of queries using heuristics, with magic numbers tuned to improve average performance on benchmarks. Empirically, estimation error significantly grows with query complexity. Alternatively, learning-based estimators offer improved accuracy, but add operational complexity preventing their adoption in-practice. Recognizing that query workloads contain highly repetitive subquery patterns, we learn many simple regressors online, each localized to a pattern. The regressor corresponding to a pattern can be randomly-accessed using hash of graph structure of the subquery. Our method has negligible overhead and competes with SoTA learning-based approaches on error metrics. Further, amending PostgreSQL with our method achieves notable accuracy and runtime improvements over traditional methods and drastically reduces operational costs compared to other learned cardinality estimators, thereby offering the most practical and efficient solution on the Pareto frontier. Concretely, simulating JOB-lite workload on IMDb speeds-up execution by 7.5 minutes (>30%) while incurring only 37 seconds overhead for online learning.

Anton Tsitsulin, Marina Munkhoeva, Bryan Perozzi

Unsupervised learning has recently significantly gained in popularity, especially with deep learning-based approaches. Despite numerous successes and approaching supervised-level performance on a variety of academic benchmarks, it is still hard to train and evaluate SSL models in practice due to the unsupervised nature of the problem. Even with networks trained in a supervised fashion, it is often unclear whether they will perform well when transferred to another domain. Past works are generally limited to assessing the amount of information contained in embeddings, which is most relevant for self-supervised learning of deep neural networks. This works chooses to follow a different approach: can we quantify how easy it is to linearly separate the data in a stable way? We survey the literature and uncover three methods that could be potentially used for evaluating quality of representations. We also introduce one novel method based on recent advances in understanding the high-dimensional geometric structure of self-supervised learning. We conduct extensive experiments and study the properties of these metrics and ones introduced in the previous work. Our results suggest that while there is no free lunch, there are metrics that can robustly estimate embedding quality in an unsupervised way.

Kaidi Cao, Phitchaya Mangpo Phothilimthana, Sami Abu-El-Haija et al.

Learning to predict properties of large graphs is challenging because each prediction requires the knowledge of an entire graph, while the amount of memory available during training is bounded. Here we propose Graph Segment Training (GST), a general framework that utilizes a divide-and-conquer approach to allow learning large graph property prediction with a constant memory footprint. GST first divides a large graph into segments and then backpropagates through only a few segments sampled per training iteration. We refine the GST paradigm by introducing a historical embedding table to efficiently obtain embeddings for segments not sampled for backpropagation. To mitigate the staleness of historical embeddings, we design two novel techniques. First, we finetune the prediction head to fix the input distribution shift. Second, we introduce Stale Embedding Dropout to drop some stale embeddings during training to reduce bias. We evaluate our complete method GST-EFD (with all the techniques together) on two large graph property prediction benchmarks: MalNet and TpuGraphs. Our experiments show that GST-EFD is both memory-efficient and fast, while offering a slight boost on test accuracy over a typical full graph training regime.

Qi Zhu, Natalia Ponomareva, Jiawei Han et al.

There has been a recent surge of interest in designing Graph Neural Networks (GNNs) for semi-supervised learning tasks. Unfortunately this work has assumed that the nodes labeled for use in training were selected uniformly at random (i.e. are an IID sample). However in many real world scenarios gathering labels for graph nodes is both expensive and inherently biased -- so this assumption can not be met. GNNs can suffer poor generalization when this occurs, by overfitting to superfluous regularities present in the training data. In this work we present a method, Shift-Robust GNN (SR-GNN), designed to account for distributional differences between biased training data and the graph's true inference distribution. SR-GNN adapts GNN models for the presence of distributional shifts between the nodes which have had labels provided for training and the rest of the dataset. We illustrate the effectiveness of SR-GNN in a variety of experiments with biased training datasets on common GNN benchmark datasets for semi-supervised learning, where we see that SR-GNN outperforms other GNN baselines by accuracy, eliminating at least (~40%) of the negative effects introduced by biased training data. On the largest dataset we consider, ogb-arxiv, we observe an 2% absolute improvement over the baseline and reduce 30% of the negative effects.

Benedek Rozemberczki, Peter Englert, Amol Kapoor et al.

In this work we propose Pathfinder Discovery Networks (PDNs), a method for jointly learning a message passing graph over a multiplex network with a downstream semi-supervised model. PDNs inductively learn an aggregated weight for each edge, optimized to produce the best outcome for the downstream learning task. PDNs are a generalization of attention mechanisms on graphs which allow flexible construction of similarity functions between nodes, edge convolutions, and cheap multiscale mixing layers. We show that PDNs overcome weaknesses of existing methods for graph attention (e.g. Graph Attention Networks), such as the diminishing weight problem. Our experimental results demonstrate competitive predictive performance on academic node classification tasks. Additional results from a challenging suite of node classification experiments show how PDNs can learn a wider class of functions than existing baselines. We analyze the relative computational complexity of PDNs, and show that PDN runtime is not considerably higher than static-graph models. Finally, we discuss how PDNs can be used to construct an easily interpretable attention mechanism that allows users to understand information propagation in the graph.

Ştefan Postăvaru, Anton Tsitsulin, Filipe Miguel Gonçalves de Almeida et al.

In this paper, we introduce InstantEmbedding, an efficient method for generating single-node representations using local PageRank computations. We theoretically prove that our approach produces globally consistent representations in sublinear time. We demonstrate this empirically by conducting extensive experiments on real-world datasets with over a billion edges. Our experiments confirm that InstantEmbedding requires drastically less computation time (over 9,000 times faster) and less memory (by over 8,000 times) to produce a single node's embedding than traditional methods including DeepWalk, node2vec, VERSE, and FastRP. We also show that our method produces high quality representations, demonstrating results that meet or exceed the state of the art for unsupervised representation learning on tasks like node classification and link prediction.

Jonathan Halcrow, Alexandru Moşoi, Sam Ruth et al.

How can we find the right graph for semi-supervised learning? In real world applications, the choice of which edges to use for computation is the first step in any graph learning process. Interestingly, there are often many types of similarity available to choose as the edges between nodes, and the choice of edges can drastically affect the performance of downstream semi-supervised learning systems. However, despite the importance of graph design, most of the literature assumes that the graph is static. In this work, we present Grale, a scalable method we have developed to address the problem of graph design for graphs with billions of nodes. Grale operates by fusing together different measures of(potentially weak) similarity to create a graph which exhibits high task-specific homophily between its nodes. Grale is designed for running on large datasets. We have deployed Grale in more than 20 different industrial settings at Google, including datasets which have tens of billions of nodes, and hundreds of trillions of potential edges to score. By employing locality sensitive hashing techniques,we greatly reduce the number of pairs that need to be scored, allowing us to learn a task specific model and build the associated nearest neighbor graph for such datasets in hours, rather than the days or even weeks that might be required otherwise. We illustrate this through a case study where we examine the application of Grale to an abuse classification problem on YouTube with hundreds of million of items. In this application, we find that Grale detects a large number of malicious actors on top of hard-coded rules and content classifiers, increasing the total recall by 89% over those approaches alone.

Amol Kapoor, Xue Ben, Luyang Liu et al.

In this work, we examine a novel forecasting approach for COVID-19 case prediction that uses Graph Neural Networks and mobility data. In contrast to existing time series forecasting models, the proposed approach learns from a single large-scale spatio-temporal graph, where nodes represent the region-level human mobility, spatial edges represent the human mobility based inter-region connectivity, and temporal edges represent node features through time. We evaluate this approach on the US county level COVID-19 dataset, and demonstrate that the rich spatial and temporal information leveraged by the graph neural network allows the model to learn complex dynamics. We show a 6% reduction of RMSLE and an absolute Pearson Correlation improvement from 0.9978 to 0.998 compared to the best performing baseline models. This novel source of information combined with graph based deep learning approaches can be a powerful tool to understand the spread and evolution of COVID-19. We encourage others to further develop a novel modeling paradigm for infectious disease based on GNNs and high resolution mobility data.

Aleksandar Bojchevski, Johannes Gasteiger, Bryan Perozzi et al.

Graph neural networks (GNNs) have emerged as a powerful approach for solving many network mining tasks. However, learning on large graphs remains a challenge - many recently proposed scalable GNN approaches rely on an expensive message-passing procedure to propagate information through the graph. We present the PPRGo model which utilizes an efficient approximation of information diffusion in GNNs resulting in significant speed gains while maintaining state-of-the-art prediction performance. In addition to being faster, PPRGo is inherently scalable, and can be trivially parallelized for large datasets like those found in industry settings. We demonstrate that PPRGo outperforms baselines in both distributed and single-machine training environments on a number of commonly used academic graphs. To better analyze the scalability of large-scale graph learning methods, we introduce a novel benchmark graph with 12.4 million nodes, 173 million edges, and 2.8 million node features. We show that training PPRGo from scratch and predicting labels for all nodes in this graph takes under 2 minutes on a single machine, far outpacing other baselines on the same graph. We discuss the practical application of PPRGo to solve large-scale node classification problems at Google.

Anton Tsitsulin, John Palowitch, Bryan Perozzi et al.

Graph Neural Networks (GNNs) have achieved state-of-the-art results on many graph analysis tasks such as node classification and link prediction. However, important unsupervised problems on graphs, such as graph clustering, have proved more resistant to advances in GNNs. Graph clustering has the same overall goal as node pooling in GNNs - does this mean that GNN pooling methods do a good job at clustering graphs? Surprisingly, the answer is no - current GNN pooling methods often fail to recover the cluster structure in cases where simple baselines, such as k-means applied on learned representations, work well. We investigate further by carefully designing a set of experiments to study different signal-to-noise scenarios both in graph structure and attribute data. To address these methods' poor performance in clustering, we introduce Deep Modularity Networks (DMoN), an unsupervised pooling method inspired by the modularity measure of clustering quality, and show how it tackles recovery of the challenging clustering structure of real-world graphs. Similarly, on real-world data, we show that DMoN produces high quality clusters which correlate strongly with ground truth labels, achieving state-of-the-art results with over 40% improvement over other pooling methods across different metrics.

Ines Chami, Sami Abu-El-Haija, Bryan Perozzi et al.

There has been a surge of recent interest in learning representations for graph-structured data. Graph representation learning methods have generally fallen into three main categories, based on the availability of labeled data. The first, network embedding (such as shallow graph embedding or graph auto-encoders), focuses on learning unsupervised representations of relational structure. The second, graph regularized neural networks, leverages graphs to augment neural network losses with a regularization objective for semi-supervised learning. The third, graph neural networks, aims to learn differentiable functions over discrete topologies with arbitrary structure. However, despite the popularity of these areas there has been surprisingly little work on unifying the three paradigms. Here, we aim to bridge the gap between graph neural networks, network embedding and graph regularization models. We propose a comprehensive taxonomy of representation learning methods for graph-structured data, aiming to unify several disparate bodies of work. Specifically, we propose a Graph Encoder Decoder Model (GRAPHEDM), which generalizes popular algorithms for semi-supervised learning on graphs (e.g. GraphSage, Graph Convolutional Networks, Graph Attention Networks), and unsupervised learning of graph representations (e.g. DeepWalk, node2vec, etc) into a single consistent approach. To illustrate the generality of this approach, we fit over thirty existing methods into this framework. We believe that this unifying view both provides a solid foundation for understanding the intuition behind these methods, and enables future research in the area.

Anton Tsitsulin, Marina Munkhoeva, Bryan Perozzi

Graph comparison is a fundamental operation in data mining and information retrieval. Due to the combinatorial nature of graphs, it is hard to balance the expressiveness of the similarity measure and its scalability. Spectral analysis provides quintessential tools for studying the multi-scale structure of graphs and is a well-suited foundation for reasoning about differences between graphs. However, computing full spectrum of large graphs is computationally prohibitive; thus, spectral graph comparison methods often rely on rough approximation techniques with weak error guarantees. In this work, we propose SLaQ, an efficient and effective approximation technique for computing spectral distances between graphs with billions of nodes and edges. We derive the corresponding error bounds and demonstrate that accurate computation is possible in time linear in the number of graph edges. In a thorough experimental evaluation, we show that SLaQ outperforms existing methods, oftentimes by several orders of magnitude in approximation accuracy, and maintains comparable performance, allowing to compare million-scale graphs in a matter of minutes on a single machine.

John Palowitch, Bryan Perozzi

Are Graph Neural Networks (GNNs) fair? In many real world graphs, the formation of edges is related to certain node attributes (e.g. gender, community, reputation). In this case, standard GNNs using these edges will be biased by this information, as it is encoded in the structure of the adjacency matrix itself. In this paper, we show that when metadata is correlated with the formation of node neighborhoods, unsupervised node embedding dimensions learn this metadata. This bias implies an inability to control for important covariates in real-world applications, such as recommendation systems. To solve these issues, we introduce the Metadata-Orthogonal Node Embedding Training (MONET) unit, a general model for debiasing embeddings of nodes in a graph. MONET achieves this by ensuring that the node embeddings are trained on a hyperplane orthogonal to that of the node metadata. This effectively organizes unstructured embedding dimensions into an interpretable topology-only, metadata-only division with no linear interactions. We illustrate the effectiveness of MONET though our experiments on a variety of real world graphs, which shows that our method can learn and remove the effect of arbitrary covariates in tasks such as preventing the leakage of political party affiliation in a blog network, and thwarting the gaming of embedding-based recommendation systems.

Alessandro Epasto, Bryan Perozzi

Recent interest in graph embedding methods has focused on learning a single representation for each node in the graph. But can nodes really be best described by a single vector representation? In this work, we propose a method for learning multiple representations of the nodes in a graph (e.g., the users of a social network). Based on a principled decomposition of the ego-network, each representation encodes the role of the node in a different local community in which the nodes participate. These representations allow for improved reconstruction of the nuanced relationships that occur in the graph -- a phenomenon that we illustrate through state-of-the-art results on link prediction tasks on a variety of graphs, reducing the error by up to $90\%$. In addition, we show that these embeddings allow for effective visual analysis of the learned community structure.

Sami Abu-El-Haija, Bryan Perozzi, Amol Kapoor et al.

Existing popular methods for semi-supervised learning with Graph Neural Networks (such as the Graph Convolutional Network) provably cannot learn a general class of neighborhood mixing relationships. To address this weakness, we propose a new model, MixHop, that can learn these relationships, including difference operators, by repeatedly mixing feature representations of neighbors at various distances. Mixhop requires no additional memory or computational complexity, and outperforms on challenging baselines. In addition, we propose sparsity regularization that allows us to visualize how the network prioritizes neighborhood information across different graph datasets. Our analysis of the learned architectures reveals that neighborhood mixing varies per datasets.

Rami Al-Rfou, Dustin Zelle, Bryan Perozzi

Can neural networks learn to compare graphs without feature engineering? In this paper, we show that it is possible to learn representations for graph similarity with neither domain knowledge nor supervision (i.e.\ feature engineering or labeled graphs). We propose Deep Divergence Graph Kernels, an unsupervised method for learning representations over graphs that encodes a relaxed notion of graph isomorphism. Our method consists of three parts. First, we learn an encoder for each anchor graph to capture its structure. Second, for each pair of graphs, we train a cross-graph attention network which uses the node representations of an anchor graph to reconstruct another graph. This approach, which we call isomorphism attention, captures how well the representations of one graph can encode another. We use the attention-augmented encoder's predictions to define a divergence score for each pair of graphs. Finally, we construct an embedding space for all graphs using these pair-wise divergence scores. Unlike previous work, much of which relies on 1) supervision, 2) domain specific knowledge (e.g. a reliance on Weisfeiler-Lehman kernels), and 3) known node alignment, our unsupervised method jointly learns node representations, graph representations, and an attention-based alignment between graphs. Our experimental results show that Deep Divergence Graph Kernels can learn an unsupervised alignment between graphs, and that the learned representations achieve competitive results when used as features on a number of challenging graph classification tasks. Furthermore, we illustrate how the learned attention allows insight into the the alignment of sub-structures across graphs.

Sami Abu-El-Haija, Amol Kapoor, Bryan Perozzi et al.

Graph Convolutional Networks (GCNs) have shown significant improvements in semi-supervised learning on graph-structured data. Concurrently, unsupervised learning of graph embeddings has benefited from the information contained in random walks. In this paper, we propose a model: Network of GCNs (N-GCN), which marries these two lines of work. At its core, N-GCN trains multiple instances of GCNs over node pairs discovered at different distances in random walks, and learns a combination of the instance outputs which optimizes the classification objective. Our experiments show that our proposed N-GCN model improves state-of-the-art baselines on all of the challenging node classification tasks we consider: Cora, Citeseer, Pubmed, and PPI. In addition, our proposed method has other desirable properties, including generalization to recently proposed semi-supervised learning methods such as GraphSAGE, allowing us to propose N-SAGE, and resilience to adversarial input perturbations.

Sami Abu-El-Haija, Bryan Perozzi, Rami Al-Rfou et al.

Graph embedding methods represent nodes in a continuous vector space, preserving information from the graph (e.g. by sampling random walks). There are many hyper-parameters to these methods (such as random walk length) which have to be manually tuned for every graph. In this paper, we replace random walk hyper-parameters with trainable parameters that we automatically learn via backpropagation. In particular, we learn a novel attention model on the power series of the transition matrix, which guides the random walk to optimize an upstream objective. Unlike previous approaches to attention models, the method that we propose utilizes attention parameters exclusively on the data (e.g. on the random walk), and not used by the model for inference. We experiment on link prediction tasks, as we aim to produce embeddings that best-preserve the graph structure, generalizing to unseen information. We improve state-of-the-art on a comprehensive suite of real world datasets including social, collaboration, and biological networks. Adding attention to random walks can reduce the error by 20% to 45% on datasets we attempted. Further, our learned attention parameters are different for every graph, and our automatically-found values agree with the optimal choice of hyper-parameter if we manually tune existing methods.

Sami Abu-El-Haija, Bryan Perozzi, Rami Al-Rfou

We propose a new method for embedding graphs while preserving directed edge information. Learning such continuous-space vector representations (or embeddings) of nodes in a graph is an important first step for using network information (from social networks, user-item graphs, knowledge bases, etc.) in many machine learning tasks. Unlike previous work, we (1) explicitly model an edge as a function of node embeddings, and we (2) propose a novel objective, the "graph likelihood", which contrasts information from sampled random walks with non-existent edges. Individually, both of these contributions improve the learned representations, especially when there are memory constraints on the total size of the embeddings. When combined, our contributions enable us to significantly improve the state-of-the-art by learning more concise representations that better preserve the graph structure. We evaluate our method on a variety of link-prediction task including social networks, collaboration networks, and protein interactions, showing that our proposed method learn representations with error reductions of up to 76% and 55%, on directed and undirected graphs. In addition, we show that the representations learned by our method are quite space efficient, producing embeddings which have higher structure-preserving accuracy but are 10 times smaller.

Aria Rezaei, Bryan Perozzi, Leman Akoglu

Given a set of attributed subgraphs known to be from different classes, how can we discover their differences? There are many cases where collections of subgraphs may be contrasted against each other. For example, they may be assigned ground truth labels (spam/not-spam), or it may be desired to directly compare the biological networks of different species or compound networks of different chemicals. In this work we introduce the problem of characterizing the differences between attributed subgraphs that belong to different classes. We define this characterization problem as one of partitioning the attributes into as many groups as the number of classes, while maximizing the total attributed quality score of all the given subgraphs. We show that our attribute-to-class assignment problem is NP-hard and an optimal $(1 - 1/e)$-approximation algorithm exists. We also propose two different faster heuristics that are linear-time in the number of attributes and subgraphs. Unlike previous work where only attributes were taken into account for characterization, here we exploit both attributes and social ties (i.e. graph structure). Through extensive experiments, we compare our proposed algorithms, show findings that agree with human intuition on datasets from Amazon co-purchases, Congressional bill sponsorships, and DBLP co-authorships. We also show that our approach of characterizing subgraphs is better suited for sense-making than discriminating classification approaches.

Yingtao Tian, Vivek Kulkarni, Bryan Perozzi et al.

Do word embeddings converge to learn similar things over different initializations? How repeatable are experiments with word embeddings? Are all word embedding techniques equally reliable? In this paper we propose evaluating methods for learning word representations by their consistency across initializations. We propose a measure to quantify the similarity of the learned word representations under this setting (where they are subject to different random initializations). Our preliminary results illustrate that our metric not only measures a intrinsic property of word embedding methods but also correlates well with other evaluation metrics on downstream tasks. We believe our methods are is useful in characterizing robustness -- an important property to consider when developing new word embedding methods.

Vivek Kulkarni, Bryan Perozzi, Steven Skiena

We present a new computational technique to detect and analyze statistically significant geographic variation in language. Our meta-analysis approach captures statistical properties of word usage across geographical regions and uses statistical methods to identify significant changes specific to regions. While previous approaches have primarily focused on lexical variation between regions, our method identifies words that demonstrate semantic and syntactic variation as well. We extend recently developed techniques for neural language models to learn word representations which capture differing semantics across geographical regions. In order to quantify this variation and ensure robust detection of true regional differences, we formulate a null model to determine whether observed changes are statistically significant. Our method is the first such approach to explicitly account for random variation due to chance while detecting regional variation in word meaning. To validate our model, we study and analyze two different massive online data sets: millions of tweets from Twitter spanning not only four different countries but also fifty states, as well as millions of phrases contained in the Google Book Ngrams. Our analysis reveals interesting facets of language change at multiple scales of geographic resolution -- from neighboring states to distant continents. Finally, using our model, we propose a measure of semantic distance between languages. Our analysis of British and American English over a period of 100 years reveals that semantic variation between these dialects is shrinking.

Vivek Kulkarni, Rami Al-Rfou, Bryan Perozzi et al.

We propose a new computational approach for tracking and detecting statistically significant linguistic shifts in the meaning and usage of words. Such linguistic shifts are especially prevalent on the Internet, where the rapid exchange of ideas can quickly change a word's meaning. Our meta-analysis approach constructs property time series of word usage, and then uses statistically sound change point detection algorithms to identify significant linguistic shifts. We consider and analyze three approaches of increasing complexity to generate such linguistic property time series, the culmination of which uses distributional characteristics inferred from word co-occurrences. Using recently proposed deep neural language models, we first train vector representations of words for each time period. Second, we warp the vector spaces into one unified coordinate system. Finally, we construct a distance-based distributional time series for each word to track it's linguistic displacement over time. We demonstrate that our approach is scalable by tracking linguistic change across years of micro-blogging using Twitter, a decade of product reviews using a corpus of movie reviews from Amazon, and a century of written books using the Google Book-ngrams. Our analysis reveals interesting patterns of language usage change commensurate with each medium.

Rami Al-Rfou, Vivek Kulkarni, Bryan Perozzi et al.

The increasing diversity of languages used on the web introduces a new level of complexity to Information Retrieval (IR) systems. We can no longer assume that textual content is written in one language or even the same language family. In this paper, we demonstrate how to build massive multilingual annotators with minimal human expertise and intervention. We describe a system that builds Named Entity Recognition (NER) annotators for 40 major languages using Wikipedia and Freebase. Our approach does not require NER human annotated datasets or language specific resources like treebanks, parallel corpora, and orthographic rules. The novelty of approach lies therein - using only language agnostic techniques, while achieving competitive performance. Our method learns distributed word representations (word embeddings) which encode semantic and syntactic features of words in each language. Then, we automatically generate datasets from Wikipedia link structure and Freebase attributes. Finally, we apply two preprocessing stages (oversampling and exact surface form matching) which do not require any linguistic expertise. Our evaluation is two fold: First, we demonstrate the system performance on human annotated datasets. Second, for languages where no gold-standard benchmarks are available, we propose a new method, distant evaluation, based on statistical machine translation.