GraphNews

Co-authors: Sofus Macskássy, Carol Jin, Shiyong Lin, Xiaomin Wei, and Michael O’Neill

Two experts go head-to-head – then you decide

We present the Graph Forward-Forward (GFF) algorithm, an extension of the Forward-Forward procedure to graphs, able to handle features distributed over a graph's nodes. This allows training graph neural networks with forward passes only, without backpropagation. Our method is agnostic to the message-passing scheme, and provides a more biologically plausible learning scheme than backpropagation, while also carrying computational advantages. With GFF, graph neural networks are trained greedily layer by layer, using both positive and negative samples. We run experiments on 11 standard graph property prediction tasks, showing how GFF provides an effective alternative to backpropagation for training graph neural networks. This shows in particular that this procedure is remarkably efficient in spite of combining the per-layer training with the locality of the processing in a GNN.

TigerGraph, provider of an advanced analytics and ML platform for connected data, is releasing the latest version (3.9) of TigerGraph Cloud, the native parallel graph database-as-a-service. TigerGraph Cloud 3.9 includes new security, advanced AI, and machine learning capabilities that meet the demands of its rapidly growing customer base and streamline the adoption, deployment, and management of the most scalable graph database platform, according to the company. The underlying parallel native graph database engine is also available for on-prem or self-managed cloud installation.

AI-chatbot software for complex requirements

If you're thinking about building Data Products in your organisation then you need to know about JSON-LD! JSON-LD, short for JavaScript Object Notation for… | 27 comments on LinkedIn

Geometric Deep Learning is a term for approaches considering ML problems from the perspectives of symmetry and invariance. It provides a common blueprint for CNNs, GNNs, and Transformers. Here, we study the history of GDL from ancient Greek geometry to Graph Neural Networks.

As promised by many analysts, the automatic creation of knowledge graphs should have allowed us to reach the Holy Grail of knowledge…

It can be hard to work out from Wikipedia if an Ontologist is a philosopher, a physicist or something completely different

Graph databases are finding new use cases in sales, ecommerce, healthcare, financial services, fraud detection and much more.

This paper provides an extensive overview of the use of knowledge graphs in the context of Explainable Machine Learning. As of late, explainable AI ha…

Topics related to information management taxonomies posted by the author of the book, The Accidental Taxonomist.

Digital Science has completed the acquisition of metaphacts, which has become the newest member of the Digital Science family.

SeeNews talks to Ontotext’s CEO Atanas Kiryakov about the latest developments in semantic technology and graph technology as well as GraphDB

In many ways, graphs are the main modality of data we receive from nature. This is due to the fact that most of the patterns we see, both in natural and artificial systems, are elegantly...

Proper modeling of a graph database can be challenging. Let’s check how ChatGPT will perform with a semi-real case.

Hot trends and major advancements

The story so far

We’re on a journey to advance and democratize artificial intelligence through open source and open science.

Using open data to make your own graph…

🔎🔬🔍 Improve your knowledge graph with GRAPE! 🦟🦠🪲 You may not be aware, but KGs can have topological oddities affecting your analysis and modelling…

Critical steps in improving the productivity of enterprise note-taking

Knowledge Graph Costs I just finished my primary research for a new paper on the costs and obstacles of adopting knowledge graph. The three themes that… | 10 comments on LinkedIn

To use this feature, you just need to add 'with("Neptune#ml.inductiveInference")' to your Gremlin query with Amazon Neptune. Happy graphing, folks!...

RDFox v6.0 is here and it’s a big one! This release is landmarked by ground-breaking features including the new high availability setup and a total reimagining of the handling of named graphs. These features alone offer the opportunity for truly cutting-edge solutions, but when combined with the myriad of other improvements and the blistering capabilities of RDFox already in place, they allow you to step ahead of the curve. | 6 min read | Nov 29, 2022

Is Denoising Diffusion all you need?

Our most recent edition of the G-Research Distinguished Speaker Series took place at the Royal College of Physicians in Central London. Our featured speaker, Amy Hodler, is a graph evangelist, network science expert, and co-author of 'Graph Algorithms'. In her talk 'Average is a Lie - Using Graph Analytics to Improve Predictions', Amy discusses a range of topics, including: Graph queries Graph algorithms Relationships and structures in data Predicting behavioural change Machine learning Graph embedding Link prediction Data lineage Financial contagion Use of graph analytics in security Learn more about G-Research (https://www.gresearch.co.uk/) and view talks from speakers such as Professor Sir Martin Hairer and Wes McKinney in our Distinguished Speaker Series playlist. Interested in attending a future Distinguished Speaker Series event? Register your interest here now: https://events.beamery.com/gresearch/all-dss-events-mntauiaxr