In an exciting collaboration with the extraordinary AI and Machine Learning team at
Microsoft Academic, the Lens has augmented its holdings dramatically and now provides vastly better searching, metadata, filtering and visualization options. This is the beginning of a major public good activity to render global innovation metadata fully open, and ultimately a comprehensive, definitive and curated resource.
As a first step, we've spent a great deal of time harmonizing three distinctly different metadata resources, namely PubMed, CrossRef and now Microsoft Academic Graph. This will be a work in progress, so watch this space as the quality control marches on, with your help.
The Lens' scholarly dataset now includes 188M scholarly records sourced from
- Scholarly records from PubMed (28M, PM)
- Scholarly records from Crossref (94M, CR)
- Scholarly records from Microsoft Academic (158M, MAG)
For the first time, anyone can now easily visualise the entire work product of almost any institution in the world, filter it by subject, by keywords, by date, by type of publication. And using the powerful linking to global patents, you can explore the Influence Mapping (In4M) of any scholarly work.
For instance, here's the
life sciences published work product of University of California at Davis. To our delight, it shows the exceptional work product of Cambia's old pal, fine scientist and former sabbatical, Bruce Hammock, as their most published scholar. Good on you Bruce!
Determining the nature of a web document is not trivial. Microsoft on the average crawls and indexes each 'item' about 5.5 times, so its clear that the many forms can make assignment to a 'Type' challenging. As more than 80M of the MAG records have no "publication type" yet (these can be anything from lab protocols, to strangely formatted books or journals to chemical datasheets!), we offer Lens users a default scholarly search within a scholarly set of 104M records that excludes the "unknown" publication types and includes known types such as conference papers, book chapters, journal articles, dissertations, and a few others.
If you are keen on exploring all the set (188M records) which could be great for Prior Art Searching as it includes standards documents, data sheets, conference papers, theses, journal articles, lab protocols and much more, you will need to click the exclusion option for the "unknown" category in the "Publication Type" facet in the opened sidebar to add them back.
As we give more attention to the "publication type" feature, we will develop more customized presets of the data that you can explore separately or jointly. We also include here, Conference information, Field of Study and improved Author and Affiliation data.
As mentioned, The Lens - with our many publicly spirited collaborators - is currently subjecting the affiliation data to an increasingly rigorous quality control process and we can not assert the accuracy of such data yet. With a few months of work with the community and with Microsoft Academic's great Machine Learning team, we're confident this will become Gold Standard. One thing to remember about using machine learning: each week they scan the entire web, and each week they've improved the algorithms with our (and your) feedback. So every week, it gets better and sometimes massively so.
We thank all the institutions who have been sharing their public scholarly data with the Lens and we encourage others to do so so we can harmonize and ingest any contextual relevant information and make it publicly available.
Scholarly improvements
- Facet filters now support autosuggest
- Facets 'Choose More' explorer improved
- Affiliation Logo Grid visualisation added
- Ability to customise and share displayed visualisations
- Added helpful information to explain data sources for different facets
Patent Analysis improvements
- New logos added for over 1000 institutions and companies added
- Column layout now optional
Classification Explorer improvements
- Search on all classification symbols within a main group or higher level code
- Don't show folder icon for symbols that don't have any subgroups (user's request)
- Select and search / filter on multiple symbols from Classification Explorer (user's request)
Notes UI overhaul
- Inline patent notes design improvements
- Add/edit/delete notes UI improvements
- Work Area notes page added
- Patent Collections can now have notes associated with them
Bug Fixes
- Fix occasional analysis chart rendering issues on some screen sizes
- Fix filter links in Article > Patent Citations > Analysis View charts
- Fix saved query failure for long queries
- Fix Scholar facets "Clear All" button clearing the query
We enjoy hearing from you and are always looking to understand how we can further support your decision-making. If you need guidance to navigate The Lens and utilise its many different features, please visit our Support Center to check out the latest tutorials or contact us at support@about.lens.org.
