UPScope is a planning grant from the Mellon Foundation to the AAUP to develop a natural language search platform for books published by AAUP member presses. The platform will create networks between texts through semantic search that can discover connections across disciplines. It aims to improve discovery, visibility, and usage of humanities monographs. Advisory councils and working groups were formed to provide guidance on technical development, business models, and outreach. Next steps include submitting a proof of concept proposal to test ingesting press content, building the discovery tool, and validating usage by scholars.
2. UPScope expands the network
UPScope is the working title of an AAUP planning grant to
develop a platform enabled by natural language full-text search
and discovery of high quality, peer-reviewed books published by
AAUP member presses.
• Creates a network between the texts published by AAUP
member presses
• Natural language search facilitates:
• crosses disciplinary silos
• results free of bias based on commercial factors (i.e. advertising)
• Links texts into the networks of scholars
4. Raison d’etre:
Monograph Discovery
Monograph discovery is not optimal for humanities research
Phrases and key words carry multiple meanings
Ideas across disciplines have different names
5. Why did we do this:
Improve discovery, visibility, and
usage of AAUP member press
books
6. Inference Engines: semantic
search
• An inference engine reads the text of the books and identifies
the concepts and key phrases within the tests
• Maps the books against one another based on those common
key phrases/key words, creating idea pathways
7. Testing of Inference Engines
• 200 files used to test inference engines
• Proved the files & metadata have today will work with these
engines
8. Governance Structure
for the planning grant
Two Advisory Councils:
• Executive Advisory Council
• Technical Advisory Council
The two councils split into nine working groups:
• Semantic Tech Evaluation
• Scholar as customer
• Up content analysis
• Access models and fulfillment
• Identifiers & metadata standards
• Reimagine Curation
• Aggregator Interfaces
• Library as customer
• AAUP member communication
9. Outreach to the community
• Surveyed AAUP member presses
• Libraries
• Scholarly Societies
• Individual Scholars
• Technology Companies
• Inference Engine Developers
10. Business Modeling
Members of the advisory council created potential business
model projections for:
book sales
discovery service sales
open access hosting fees
research services to scholars
11. Member Content Survey
166,000 digital book files exist from 104 reporting AAUP member
presses. Most are PDF
Presses publish more than 10,000 new titles a year
12. Next Steps
• Drafting a Proof of Concept project proposal to submit to the
Mellon Foundation
• Core goals of proof of concept
1. Ingest files metadata from presses
2. Develop the inference engine (discovery tool)
3. Validating scholars will use it