Artificial intelligence (AI) – already in use across a wide range of platforms and services, including Google and Facebook – is also starting to play a bigger role in scholarly communications. While there is understandable concern about the use of systems that rely entirely on algorithms, there’s no doubt that AI can also deliver real benefits to researchers and their organizations alike.
Yewno, a startup launched in 2014 by Ruggero Gramatica, is one of several new organizations seeking to harness the power of AI in the scholarly communications space. It’s based on technology that began as academic research in the field of complex systems, while investigating the foundation of econophysics — a discipline aimed at describing economic and financial cycles utilizing mathematical structures derived from physics. The mathematical framework was then applied to another complex field, that of biology and drug discovery, where the results were dramatically higher than those achieved through the manual work of a team of scientists. This success — and the recognition that this technology could be applied across many different verticals — led to the founding of Yewno.
Following the recent launch of several new features for Yewno Unearth at last week’s SSP Annual Meeting, the company’s co-founder and Chief Business Development & Strategy Officer, Ruth Pickering, answers some questions about Yewno – past, present, and future.
On your website you describe Yewno as “Technology and products that extract meaning at the atomic level, to help you understand deeper” — what exactly do you mean by that? What problem(s) are you trying to solve, and why?
Yewno’s mission is ‘Knowledge Singularity’ and by that we mean the day when knowledge, not information, is at everyone’s fingertips. In the search and discovery space the problems that people face today are the overwhelming volume of information and the fact that sources are fragmented and dispersed. There’s a great T.S. Eliot quote ‘Where’s the knowledge we lost in information’ and that sums up the problem perfectly.
What is your business model, and who are your customers?
We have a multi-vertical approach and our technology can be applied to many different disciplines. Our initial focus has been in the Education and Publishing spaces. We have customers of Yewno Discover, our Education product, in the US, UK and Germany, at both teaching focused and research institutions of all sizes. When working with publishers contributing content to Discover, many provided insights into their own discovery needs, and as a result we developed ‘Yewno Unearth’ specifically for publishers, which we we were very excited to launch last week!
Do you have any competitors? If so, how does Yewno differ from their products?
I’ve not seen another product that is able to extract concepts and correlate them in a knowledge map or provide an intelligent topic model that articulates information into knowledge by referring to specific content sets. Yewno Unearth leverages leading-edge computational semantics and machine learning algorithms that process every single part of a publication, and uses this data to construct a hierarchical fine grain topic model hugely augmenting existing categorization and taxonomy.
In the broader ecosystem our competitors are anyone in the next generation search and discovery space. There is also some competition from ‘in house’ technology teams at companies that prefer to build their own products rather than buy from vendors. The advantage of using Unearth over in-house technology is that it is compatible with the Discover product, making it possible to accurately compare your own portfolio within any given subject with a wider body of resources.
You have several products — Yewno Discover, Yewno Finance, Yewno Life Sciences, and Yewno Unearth. Can you briefly describe explain what each one is and why it’s needed?
All our products share the same technology and use a blend of computational linguistics, graph theory, and machine learning to extract meaning from information and provide knowledge, but each vertical uses a specific set of ad-hoc machine learning based algorithms and content. The Yewno Unearth product sits across all verticals and can be applied to any content set in any domain of information.
You recently launched some new features for Yewno Unearth at SSP — what are they and how will they help your mission to “transform information into knowledge”?
For the launch of Yewno Unearth we’ve added a new feature called ‘topic filter’, which allows a user to search across a portfolio of information, at topic and sub-topic level, for a particular blend of information. For example, you may chose to explore books with a blend of 75% mathematics and 25% jazz. In terms of transforming information into knowledge, the first thing you need to understand, at a granular level, is what type of information a document or a larger corpus of documents contains, and this is where Yewno can help. Providing details at the topic, sub-topic, concept and connection level, Yewno Unearth provides a new depth of granularity, accuracy and consistency about what a content set contains and from here both publishers and end users can take their journey forward based on knowledge rather than ‘guess work’.
You describe Yewno Unearth as ‘relevant to multiple publishing functions.’ Which functions do you see Unearth being well-positioned to help?
Yewno Unearth addresses portfolio categorization challenges, but also can help inform acquisitions editors about their own content, helping them to spot gaps in their list and align to courses where relevant. The categorization tool can help marketers understand the content within a single title with more clarity and detail, and also help them to select relevant content for cluster promotions targeted to specific audiences. It can be challenging for sales representatives to find an accurate list of titles pertaining to a particular course or topic that a customer is interested in, but by using Unearth they are able to quickly search all content to find the relevant content regardless of a publisher’s static taxonomy. This is particularly helpful when describing the content of a backlist, where metadata may be missing or flawed. As Unearth reads the full-text of all content, it provides the same level of detail about backlist as frontlist titles. Unearth can also help advance communication between publishing departments, allowing editors to more accurately describe their content to other teams. Journal editors may find it helpful to relay the concept and topical structure of a journal volume to society clients, and using the date filter, compare the concepts within different volumes. Overall, Unearth can empower publishers by providing essential insights into their own content across an organization, helping them to be efficient and maintain relevance.
Do you have any other new features in the pipeline for Unearth or any of your other products?
Yes, we’re exploring workflows for journal publishers enabling the ingestion of preprints which will help direct articles accurately and efficiently to the right journals and reviewers for consideration — saving time and cost in doing so.
Where do you hope Yewno will be in five years time and what are the main barriers to getting there?
In five years time I hope that, across the product range, Yewno continues to provide an environment in which knowledge is available to people so that they can work and study effectively. The AI market is in its infancy and it’s just beginning to take off, therefore our ability to remain successful rests on continuing to provide a best in class, unique technical solution alongside maintaining partnerships with our customers and communities in each vertical, to ensure that the solutions we provide continue to address their big challenges.
How would you describe Yewno to someone who has never heard of it — in one sentence?
We are a next-generation technology company focused on transforming information into knowledge, helping you to understand at a deeper level.