A5091272738- Complex Network Analysis Techniques
- Advanced Graph Neural Networks
- Opinion Dynamics and Social Influence
- Topic Modeling
- Bioinformatics and Genomic Networks
- Graph Theory and Algorithms
- Social Media and Politics
- Recommender Systems and Techniques
- Computational Drug Discovery Methods
- Machine Learning in Materials Science
- Spam and Phishing Detection
- Single-cell and spatial transcriptomics
- Misinformation and Its Impacts
- Domain Adaptation and Few-Shot Learning
- Sentiment Analysis and Opinion Mining
- Data Visualization and Analytics
- Natural Language Processing Techniques
- Machine Learning and Data Classification
- Human Mobility and Location-Based Analysis
- Peer-to-Peer Network Technologies
- Advanced Text Analysis Techniques
- Cell Image Analysis Techniques
- Wikis in Education and Collaboration
- Multimodal Machine Learning Applications
- Biomedical Text Mining and Ontologies
Stanford University
2016-2025
Chan Zuckerberg Initiative (United States)
2017-2024
Palo Alto University
2010-2024
Stanford Medicine
2011-2022
Pinterest (United States)
2019-2020
UC San Diego Health System
2019
National Bureau of Economic Research
2017
Cornell University
2009-2017
Carnegie Mellon University
2005-2015
PayPal (United States)
2015
Prediction tasks over nodes and edges in networks require careful effort engineering features used by learning algorithms. Recent research the broader field of representation has led to significant progress automating prediction themselves. However, present feature approaches are not expressive enough capture diversity connectivity patterns observed networks. Here we propose node2vec, an algorithmic framework for continuous representations In learn a mapping low-dimensional space that...
Low-dimensional embeddings of nodes in large graphs have proved extremely useful a variety prediction tasks, from content recommendation to identifying protein functions. However, most existing approaches require that all the graph are present during training embeddings; these previous inherently transductive and do not naturally generalize unseen nodes. Here we GraphSAGE, general, inductive framework leverages node feature information (e.g., text attributes) efficiently generate for...
Graph Neural Networks (GNNs) are an effective framework for representation learning of graphs. GNNs follow a neighborhood aggregation scheme, where the vector node is computed by recursively aggregating and transforming vectors its neighboring nodes. Many GNN variants have been proposed achieved state-of-the-art results on both graph classification tasks. However, despite revolutionizing learning, there limited understanding their representational properties limitations. Here, we present...
Recent advancements in deep neural networks for graph-structured data have led to state-of-the-art performance on recommender system benchmarks. However, making these methods practical and scalable web-scale recommendation tasks with billions of items hundreds millions users remains a challenge. Here we describe large-scale engine that developed deployed at Pinterest. We develop data-efficient Graph Convolutional Network (GCN) algorithm PinSage, which combines efficient random walks graph...
Even though human movement and mobility patterns have a high degree of freedom variation, they also exhibit structural due to geographic social constraints. Using cell phone location data, as well data from two online location-based networks, we aim understand what basic laws govern motion dynamics. We find that humans experience combination periodic is geographically limited seemingly random jumps correlated with their networks. Short-ranged travel both spatially temporally not effected by...
Given a water distribution network, where should we place sensors toquickly detect contaminants? Or, which blogs read to avoid missing important stories?.
How do real graphs evolve over time? What are normal growth patterns in social, technological, and information networks? Many studies have discovered static , identifying properties a single snapshot of large network or very small number snapshots; these include heavy tails for in- out-degree distributions, communities, small-world phenomena, others. However, given the lack about evolution long periods, it has been hard to convert findings into statements trends time. Here we study wide...
How do real graphs evolve over time? What are "normal" growth patterns in social, technological, and information networks? Many studies have discovered static graphs, identifying properties a single snapshot of large network, or very small number snapshots; these include heavy tails for in- out-degree distributions, communities, small-world phenomena, others. However, given the lack about network evolution long periods, it has been hard to convert findings into statements trends time.Here we...
We present an analysis of a person-to-person recommendation network, consisting 4 million people who made 16 recommendations on half products. observe the propagation and cascade sizes, which we explain by simple stochastic model. analyze how user behavior varies within communities defined network. Product purchases follow ‘long tail’ where significant share belongs to rarely sold items. establish network grows over time effective it is from viewpoint sender receiver recommendations. While...
A large body of work has been devoted to defining and identifying clusters or communities in social information networks, i.e., graphs which the nodes represent underlying entities edges some sort interaction between pairs nodes. Most such research begins with premise that a community cluster should be thought as set more and/or better connections its members than remainder network. In this paper, we explore from novel perspective several questions related meaningful come striking...
Tracking new topics, ideas, and "memes" across the Web has been an issue of considerable interest. Recent work developed methods for tracking topic shifts over long time scales, as well abrupt spikes in appearance particular named entities. However, these approaches are less suited to identification content that spreads widely then fades scales on order days - scale at which we perceive news events.
AI is undergoing a paradigm shift with the rise of models (e.g., BERT, DALL-E, GPT-3) that are trained on broad data at scale and adaptable to wide range downstream tasks. We call these foundation underscore their critically central yet incomplete character. This report provides thorough account opportunities risks models, ranging from capabilities language, vision, robotics, reasoning, human interaction) technical principles(e.g., model architectures, training procedures, data, systems,...
In order to recommend products users we must ultimately predict how a user will respond new product. To do so uncover the implicit tastes of each as well properties For example, in whether enjoy Harry Potter, it helps identify that book is about wizards, user's level interest wizardry. User feedback required discover these latent product and dimensions. Such often comes form numeric rating accompanied by review text. However, traditional methods discard text, which makes dimensions difficult...
We study online social networks in which relationships can be either positive (indicating relations such as friendship) or negative opposition antagonism). Such a mix of and links arise variety settings; we datasets from Epinions, Slashdot Wikipedia. find that the signs underlying predicted with high accuracy, using models generalize across this diverse range sites. These provide insight into some fundamental principles drive formation signed networks, shedding light on theories balance...
Prediction tasks over nodes and edges in networks require careful effort engineering features used by learning algorithms. Recent research the broader field of representation has led to significant progress automating prediction themselves. However, present feature approaches are not expressive enough capture diversity connectivity patterns observed networks. Here we propose node2vec, an algorithmic framework for continuous representations In learn a mapping low-dimensional space that...
Relations between users on social media sites often reflect a mixture of positive (friendly) and negative (antagonistic) interactions. In contrast to the bulk research networks that has focused almost exclusively interpretations links people, we study how interplay relationships affects structure on-line networks. We connect our analyses theories signed from psychology. find classical theory structural balance tends capture certain common patterns interaction, but it is also at odds with...
Given a huge real graph, how can we derive representative sample? There are many known algorithms to compute interesting measures (shortest paths, centrality, betweenness, etc.), but several of them become impractical for large graphs. Thus graph sampling is essential.The natural questions ask (a) which method use, (b) small the sample size be, and (c) scale up measurements (e.g., diameter), get estimates graph. The deeper, underlying question subtle: do measure success?.We answer above...
Networks are a fundamental tool for understanding and modeling complex systems in physics, biology, neuroscience, engineering, social science. Many networks known to exhibit rich, lower-order connectivity patterns that can be captured at the level of individual nodes edges. However, higher-order organization networks---at small network subgraphs---remains largely unknown. Here we develop generalized framework clustering based on patterns. This provides mathematical guarantees optimality...
The use of drug combinations, termed polypharmacy, is common to treat patients with complex diseases or co-existing conditions. However, a major consequence polypharmacy much higher risk adverse side effects for the patient. Polypharmacy emerge because drug-drug interactions, in which activity one may change, favorably unfavorably, if taken another drug. knowledge interactions often limited these relationships are rare, and usually not observed relatively small clinical testing. Discovering...