Graph databases to map AI in massive exercise in meta-understanding
- Reference: 1621504929
- News link: https://www.theregister.co.uk/2021/05/20/graph_databases_to_map_ai/
- Source link:
No longer the last on the shopping list of new database trends, graph processing will grow 100 per cent annually to 2023, the IT analyst giant forecast.
Graph databases are being used to help analyse network relationships like those in social media or company ownership. But they are not about to stop there, according to Pieter den Hamer, Gartner senior research director.
[1]
[2]
[3]
"The key thing to keep in mind is that graphs are, indeed, everywhere, they are in our brains, for example," Den Hamer said.
Speaking to the Gartner Data & Analytics Summit, he elaborated that graph technology would go on to underpin a new trend: composite AI.
"It is one of the biggest trends that we're seeing today in AI," Den Hamer said. "Because of this growing pervasiveness of this fundamental role of graph, we see that this will lead to composite AI, which is about the notion that graphs provide a common ground for the culmination, or if you like the composition of notable existing and new AI techniques together, they'll go well beyond the current generation of fully data-driven machine learning."
[4]
Roughly speaking, graph databases work by storing a thing in a node – say, a person or a company – and then describing its relationship to other nodes using an edge, to which a variety of parameters can be attached. They are not just being used by data scientists to solve business problems, they are also useful in the data science process, to help understand ontology and augment data integration, Den Hamer said.
Meanwhile, graph databases often come in handy for data scientists, data engineers and subject matter experts trying to quickly understand how the data is structured, using graph visualisation techniques to start "identifying the likely most relevant features and input variables that are needed for the prediction or the categorisation that they're working on," he added.
Graph may also be employed to help in feature engineering, that tricky business of figuring out what is important in the dataset. They could also be used as a basis for new types of neural networks, to help explain the outputs of AI and to uncover the business rules behind data, he said.
[5]
The status of graph has increased such that Gartner plugged it as one of the top four analytics technologies that would enable businesses to "adapt to a changing world."
"Whether it's building recommendation engines for detection systems or infrastructure monitoring, graph will become key to understand increasingly complex inter-relationships," keynote presenter and Gartner research director Gareth Herschel said.
Others on Gartner's data analytics hype list include the "data fabric", an abstraction layer that lets the user get to the right data quickly, while also controlling security and governance, and generative adversarial networks, the technique of playing one ML model off against another to generate new data.
Herschel also backed OpenAI's GPT-3, the ML language model that uses deep learning to produce human-like text.
[6]Jaguar Land Rover reaches for graph database in search of supply chain knowledge during chip shortage
[7]SQL now a dirty word for Oracle, at least in cloudy data warehouses
[8]Apache foundation ousts TinkerPop project co-founder for tweeting 'offensive humor that borders on hate speech'
[9]Life after proprietary wares: German support biz flees IBM Db2 databases for something more Postgres-shaped
"These natural language generation techniques will enable machines to tell us data stories. Instead of us becoming data literate, they are becoming human literate," he claimed.
But there are other perspectives on GPT-3, especially when dealing with important topics apparently marginal to mainstream western culture. As one [10]cognitive science PhD student pointed out , the language generator has produced "factually wrong and grossly racist text" on the subject of Ethiopia, for example.
The need for graph databases to help understand AI techniques, or GPT-3 to communicate the stories within data, raise questions about whether we always need more tools to understand or manage the tools we already have. Maybe it will be turtles all the way up, [11]as well as down . ®
Get our [12]Tech Resources
[1] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/databases&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2YKaHm@cV@iefDawMCcmVowAAAMo&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/databases&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YKaHm@cV@iefDawMCcmVowAAAMo&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/databases&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33YKaHm@cV@iefDawMCcmVowAAAMo&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/databases&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44YKaHm@cV@iefDawMCcmVowAAAMo&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/databases&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33YKaHm@cV@iefDawMCcmVowAAAMo&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[6] https://www.theregister.com/2021/05/10/jaguar_land_rover_tigergraph/
[7] https://www.theregister.com/2021/03/18/oracle_cloud_data_warehouse/
[8] https://www.theregister.com/2021/02/23/apache_tinkerpop_speech/
[9] https://www.theregister.com/2020/12/08/leaving_ibm_db2_for_postgres/
[10] https://twitter.com/abebab/status/1309137018404958215?lang=en
[11] https://en.wikipedia.org/wiki/Turtles_all_the_way_down
[12] https://whitepapers.theregister.com/
Almost there
> the language generator has produced "factually wrong and grossly racist text"
In short, almost human...
What time is it, you say?
If there's one company I wouldn't even trust to tell time it's got to be Gartner...
... analyse network relationships
.. at which point you might decide to start with some basics, like here:
https://arxiv.org/abs/2101.00863
The Atlas for the Aspiring Network Scientist
Michele Coscia
Network science is the field dedicated to the investigation and analysis of complex systems via their representations as networks. We normally model such networks as graphs: sets of nodes connected by sets of edges and a number of node and edge attributes. This deceptively simple object is the starting point of never-ending complexity, due to its ability to represent almost every facet of reality: chemical interactions, protein pathways inside cells, neural connections inside the brain, scientific collaborations, financial relations, citations in art history, just to name a few examples. If we hope to make sense of complex networks, we need to master a large analytic toolbox: graph and probability theory, linear algebra, statistical physics, machine learning, combinatorics, and more.
This book aims at providing the first access to all these tools. It is intended as an "Atlas", because its interest is not in making you a specialist in using any of these techniques. Rather, after reading this book, you will have a general understanding about the existence and the mechanics of all these approaches. You can use such an understanding as the starting point of your own career in the field of network science. This has been, so far, an interdisciplinary endeavor. The founding fathers of this field come from many different backgrounds: mathematics, sociology, computer science, physics, history, digital humanities, and more. This Atlas is charting your path to be something different from all of that: a pure network scientist.
The next graph database is not a graph database
Graph theory can be applied to any form of information (just like relational theory), but that does not mean you need necessarily to structure it as a network of nodes connected by edges.
If you want to find out if Vladimir Putin is connected to Donald Trump on LinkedIn, if Boris Johnson is related to Joseph Stalin on a DNA tracking service, or if Taliban are financed by heroin; a graph database is an excellent solution because the “graph” is unbounded. Pandemic contact tracing is also an unbounded graph, but a graph database is not a good solution because of the rate of change.
Language parsing and Bill-of-materials are also graph problems, but the best solutions is to assemble them in memory because they are bounded and finite in scale and apply complex constraints. Constraints are the weak point of graph database because “structure” is not included in meta-data – you can either apply constrains outside the database or suffer the performance problems of interpreting predicate logic.
The next “graph database” will take advantage of increased computer memory and GPGPU to traverse graphs in parallel.
"graph processing will grow 100 per cent annually to 2023"
Easy prediction.
When you're nowhere, doubling your market penetration isn't difficult.
When the market is saturated, you don't increase by 100%. You'll be lucky to increase by 5%.
Gartner remains Gartner.