Re-invigorating a middle-aged publisher with machine learning, AI and open data - PowerPoint Presentation

343 views
Uploaded On 2019-11-03

Re-invigorating a middle-aged publisher with machine learning, AI and open data - PPT Presentation

Reinvigorating a middleaged publisher with machine learning AI and open data Jonathan Griffin Managing Director IFIS Publishing amp Jignesh Bhate CEO Molecular Connections Who are IFIS educational charity ID: 762680

connections content data molecular content connections molecular data 2018 copyright pvt amp expertise product linked legacy description domain standards

Link:

Copy

Embed:

<iframe width="560" height="315" src="https://www.docslides.com/embed/762680" frameborder="0" allowfullscreen></iframe>

Download Presentation from below link

Download Presentation The PPT/PDF document "Re-invigorating a middle-aged publisher ..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.

Presentation Transcript

Re-invigorating a middle-aged publisher with machine learning, AI and open data Jonathan Griffin, Managing Director, IFIS Publishing & Jignesh Bhate , CEO, Molecular Connections

Who are IFIS? educational charity founded 50 years ago Three trade associations concerned about difficulties of locating relevant research

What does IFIS do? Publish an abstracting & indexing database for food science community Team of food scientists curate content Used by universities worldwide

Huge increases in scientific output How are things going? The difficulties of locating relevant research are greater than they were High levels of innovation

How are things going? There is now an extensive range of search tools Time to retire?

How are things going? Google Scholar is seriously flawed Text string searches (not AI) Large numbers of irrelevant results Relevant results missed Articles from predatory journals not filtered out is the most widely used tool in food science Faculty members are concerned … So, it’s not time to retire!

Training & education to promote best practice Diversifying portfolio to balance riskDeploy latest technologies to further improve quality & relevance of information search How are we reinvigorating our activities? PROBLEM: We are an editorial driven organization & lack technical expertise SOLUTION: Partnership

India's leading informatics company + Charity based in a barn in SE England = Productive partnership

FSTA (abstracting & indexing database) Increase in number of records Accuracy of search results (enhanced thesaurus) Cost savingsEscalexNew online service in an adjacent market Solves problems arising from using web to locate legislation Nominated for industry award last year New product pipeline Analytics Subsets of data

Copyright © 2018 Molecular Connections Pvt. Ltd. 10 SCOPE & CHALLENGES SCOPE Processing regulatory & compliance information pertaining to Food Sciences/Industry of different regions and languages – Real time CHALLENGES Quick search and indexing on huge datasets (Text) Handle unstructured text across different types of datasets (Documents, WEB APIs) Managing updates (Up to date information)

Copyright © 2018 Molecular Connections Pvt. Ltd. 12 Legacy Curated Content LEGACY CURATED CONTENT WEB CONTENT Different file format Million + Abstracts 10 Million + Metadata 50 + years of Legacy content

Copyright © 2018 Molecular Connections Pvt. Ltd. 13 MC’s proprietary platforms Description: End to end flexible content parser ML/AI ML based content segmentation And structure recognition modules Word, Excel, LaTeX , PDFs, XMLs, Social media texts Standards: TEI/XML/APIs Description: A high throughput semantic fingerprinting system ML/AI Topic Modelling ML based Entity Extraction Feature driven Classification Ontology based tagging/indexing Standards: APIs Heuristics/Domain Expertise: Yes Description: A Named entity recognition platform ML/AI Conditional Random fields(ML) based models with plug and lay ensemble capabilities Feedback ingestion and logs (Active training in AI terms) Standards: APIs Heuristics/Domain Expertise: Yes Description: A complete ontology management solution ML/AI ML modules that identify missing ‘Concepts’ and in parallel suggest candidate concepts Parse and mine large amounts of resources for candidate or lead generation in real time Standards: SKOS/RDF-XML/OWL/APIs Heuristics/Domain Expertise: Yes Description: A visual summary and analytics studio ML/AI Plug and play NERs and ontologies Standards: Embed, exchange formats

New Product development engine Copyright © 2018 Molecular Connections Pvt. Ltd. 15 Linked Data New Product Multiple New Products Content Slicing Granular Analytics Superior visualization Better Discoverability Benefits

Enhancing existing datasets Copyright © 2018 Molecular Connections Pvt. Ltd. 16 Linked Data Automated data processing significantly increased capacity AI-enhanced tools used to move from print-centric to enhanced digital thesaurus enabling more accurate search results