GraphNews

What’s advancing the world of graph databases, and what’s slowing us down?

My engineering friends often ask me: deep learning on graphs sounds great, but are there any real applications? While Graph Neural Networks are used in recommendation systems at Pinterest, Alibaba and Twitter, a more subtle success story is the Transformer architecture, which has taken the NLP world by storm. Through

The metaphactory platform enables knowledge workers to create and gain meaningful insight into their data with one comprehensive software solution.

Learn how DeepWalk treats creating graph embeddings as equivalent to creating word embeddings

Cross-modal Knowledge Reasoning for Knowledge-based Visual Question Answering https://t.co/AdDke9AtZ0 pic.twitter.com/QXPucOK1Xo— Aaron Bradley (@aaranged) September 3, 2020

Great job to Prasun Gera for presenting our joint research on Traversing Large #Graphs on #GPUs with Unified Memory, with Hyojong Kim, @piyusch, & Hyesoon Kim, in virtual Tokyo @VLDB2020 #DataScience @NJIT @NJITYingWu https://t.co/J7V4K94VSF pic.twitter.com/LjsrKrqHJY— David Bader (@Prof_DavidBader) September 7, 2020

We got this year's @AmazonScience AWS ML Award for our work with @befcorreia on #protein design using #geometricdeeplearning Will help to take #masif to the next level https://t.co/Ec7t2g7nqV pic.twitter.com/OrHGwHp1cE— Michael Bronstein (@mmbronstein) September 9, 2020

"PNEL: Pointer Network based End-To-End Entity Linking over Knowledge Graphs." with an evaluation over three datasets on the #Wikidata Knowledge Graph.(@debayan Banerjee et al, 2020)https://t.co/cTCq4EhrGP pic.twitter.com/Ci1yn2CPDp— WikiResearch (@WikiResearch) September 8, 2020

Some good progress happening on the RDF* mailing list, towards a de-facto standard for representing statements about statements in a user-friendly syntax. See whole thread for context if interested https://t.co/S6N9w3XLH7— Holger Knublauch (@HolgerKnublauch) September 7, 2020

(1/5) Thank you everyone who came to Graph-n-Code livestreams with @SonicDMG and I. 🙏This thread has all the links you need for FREE access to:📌 The code📌 The Images📌 The bookWe are cooking up more livestreams; stay tuned! pic.twitter.com/QFbvdSiAO9— Denise Gosnell, PhD (@DeniseKGosnell) September 8, 2020

"We propose Factual News Graph (FANG), a novel graphical social context representation and learning framework for fake news detection" > FANG: Leveraging Social Context for Fake News Detection Using Graph Representation @ngnvnhng et al. https://t.co/mTJDQW9bYo pic.twitter.com/vVeYFToMN9— Aaron Bradley (@aaranged) September 8, 2020

SPEX introspects knowledge graphs in SPARQL endpoints, using RDF's self-describing nature to give a better understanding of its schema. Once the schema is available, SPEX can be used to browse instances of this data and follow links to other data. https://t.co/B0WTz6zsQT— Tim Finin (@timFinin) August 24, 2020

Very cool: a new Knowledge Graph Search tool by @maxxeight & @MerkleAllows you to:• View entities associated with a query• Extract Knowledge Graph IDs• See scoring for different results+ preview the SERP (favourite feature)Test + bookmark here: https://t.co/WcCENJNG11 pic.twitter.com/KokprslthX— Brodie Clark (@brodieseo) September 10, 2020

Can we use #graphneuralnetworks when the graph is not given? In a new blog post I show that a new type of "latent graph learning" architectures can be thought of as a modern take on #manifoldlearninghttps://t.co/p40Sod9EOr pic.twitter.com/jw7RsKuiMi— Michael Bronstein (@mmbronstein) September 10, 2020

Just added >11M #OpenCitations to #COCI, for an overall amount of >733M citations currently available in our dataset – it can be queried via #REST API & #SPARQL endpoint and can be fully downloaded as a dump (available on #Figshare)+info at https://t.co/nxSlZGkb3G #OpenScience pic.twitter.com/FZhHYN782y— OpenCitations (@opencitations) September 7, 2020

Water networks are graphs. And, if poorly designed, vulnerable. Storing their topology in a #graphdatabase helps identifying components & weaknesses. This @graphileon demo was built on top of #neo4j, without writing any code. #lowcoding #YourAppIsAGraphhttps://t.co/2WW3YCRQwF pic.twitter.com/bJ9pYjRTkt— Graphileon (@graphileon) September 12, 2020

In this week's #twin4j, @adamcowley hows us how to build a Knowledge Graph from our Slack archiveshttps://t.co/isibvvCAZ0#neo4j pic.twitter.com/BCm6o0fcNd— Neo4j (@neo4j) September 12, 2020

Rule-Guided Graph Neural Networks for Recommender Systems https://t.co/Rt2TRzVllt pic.twitter.com/52A8tvKNz5— Aaron Bradley (@aaranged) September 10, 2020

"Pixie is one of Pinterest’s major recommendation systems used for fetching relevant Pins. Pixie is composed of a bipartite graph of all Pins and boards on Pinterest." https://t.co/Ng5PszF07x— Aaron Bradley (@aaranged) September 11, 2020

"In RDF, properties cannot be directly associated with edges. How would we represent something like [an LPG labeled edge] in RDF? In fact there are multiple ways of modeling this. A common approach is reification." @chrismungall https://t.co/0MQxFomSvX— Aaron Bradley (@aaranged) September 11, 2020

GeoSPARQL+: Syntax, Semantics and System for Integrated Querying of Graph, Raster and Vector Data - Technical Report / @situxxx, @ststaab, Daniel Janke https://t.co/qrsPikb9U0 pic.twitter.com/dQ7qdzP26O— Aaron Bradley (@aaranged) September 11, 2020

Using Graph Convolutional Networks and TD(λ) to play the game of Risk https://t.co/Js0tUctECl pic.twitter.com/9EgI7hZa3d— Aaron Bradley (@aaranged) September 15, 2020

Speaking of personalized knowledge graphs....Knowledge Graphs to Empower Humanity-inspired AI Systems @hemant_pt, Valerie Shalin, @amit_p https://t.co/7gYiAxNHhU pic.twitter.com/bVccm8IxCK— Aaron Bradley (@aaranged) September 16, 2020

Tax Knowledge Graph for a Smarter and More Personalized TurboTax @jiebingyu et al. of @Intuit https://t.co/XZP9afAU34 pic.twitter.com/iYAcCVmmVn— Aaron Bradley (@aaranged) September 15, 2020

@davidebus is delivering a brilliant keynote at DeepOntoNLP (https://t.co/XYE99QzEgB) about of role of #NLP and #DeepLearning in the generation of the #Artificial #Intelligence #KnowledgeGraph (AI-KG, https://t.co/rH7MgH7e89) from research publications. pic.twitter.com/SkWhpx6IUg— Francesco Osborne (@FraOsborne) September 16, 2020

Traditional KGs are based on triples, whereas new KGs like #wikidata use statements and qualifiers to instantiate each edge further making the graph hyper-relational (img1). We incorporate these qualifiers by modifying existing multi-relational GNN (CompGCN) in the StarE (img 2). pic.twitter.com/Qs2xYZEDQy— Michael Galkin (@michael_galkin) September 15, 2020

#kdd2020 Very excited to share our work on Continuous-time Graph Neural Network: "Neural Dynamics on Complex Networks" soon in the KDD2020 research track with @feiwang03 @WCMPopHealthSci @WeillCornell @kdd_news Arxiv: https://t.co/nfaLr94aPkA learned network dynamics of genes pic.twitter.com/I9FwP7wkqz— Chengxi Zang (@calvin_zcx) August 27, 2020