SHARIAsource–universal access to Islam through data science

SHARIAsource seeks to meet an urgent need for preservation of a significant part of our shared human cultural heritage. Islam arises in the news daily, and directly affects a fifth of the world’s population. Shari’a is of increasing relevance: from religion and law to policy and public debate. But its historical sources are inaccessible or rapidly disappearing through decay and war (Syria, Yemen, Iraq). Our mission is to collect the world’s information on Islamic law and history, and to facilitate universal access and promote new research using data science and AI tools. We will digitize the extensive Harvard collections and add sources from endangered or inaccessible caches globally. Our partners innovate in data science and AI to turn the sources into readable data, for expanding access and research. Preservation and access to these sources is essential to changing notions of law, women’s rights, and access to education and knowledge worldwide.

Access to sources for the global cultural heritage of Islam–a 1400-year history–are rapidly disappearing through war (Syria, Yemen, Iraq) and decay; sources are generally unavailable in the small number of libraries that house them. Regional preservation is unlikely: political or religious-sectarian sensitivities makes it difficult to get neutral partners to collaborate in the Middle East, and there is too little AI innovation for meaningful access. A sense of ephemeralness, impossibility, and delay pervades an age that demands a sense of urgency. Meanwhile, shari’a is of increasing relevance globally: from religion and law to policy and public debate. But conversations and education proceed with anemic data; media outlets prefer sensation to information. The problem requires innovation driven by local communities in the Muslim world together with global researchers with the wherewithal to access and digitize the sources and to inform significant conversations around law history and education more broadly arising from them. Addressing this global problem thus requires a remarkable lineup of institutional infrastructure and vision, access to global sources and pathways to local partnerships, and subject matter expertise in Islam and in data science. We combine these elements. We are committed to leading solutions to the problem long-term.

Islam crops up in news headlines almost every day, and directly affects 1.5 billion+ Muslims around the world and the US and other countries interacting with its past and present. But its disappearing and otherwise inaccessible heritage makes it almost impossible to engage in informed conversation, education, and law and policymaking. Barriers to access are too high. SHARIAsource will lower them. We will fill the rather urgent need for preserving the 1400-year old cultural heritage of Islam through collecting and making the sources accessible. We will combine the latest data science and AI tools with the expertise of our global network of Islamic law experts. This joinder of Islam and AI is essential in these times. We will make it possible for the first time for local populations in the Middle East and larger Muslim world to access their own histories and make informed determinations about the laws that govern them, researchers around the glove to do the same, and data scientists to pioneer innovative new methods that will radically advance the possibilities for access and new conclusions never imagined or possible without the use of data science and AI.

SHARIAsource is making possible a world in which everyone can access the world's quickly disappearing cultural heritage of Islam developed over 1400 years, and developing AI tools to facilitate access its use. We will radically change the face of access to knowledge, inform our notions of global history and law, and innovate new uses for AI in education and cultural preservation as we know it.

Harvard, where SHARIAsource is housed, boasts one of the world’s largest library collections in Islamic sources and the leading experts in the field. We are well-positioned with infrastructure, knowledge, and ethos to lead preservation, excellence in scholarship, and the use of data science and AI. We propose to build a Portal, People, and Publications designed to foster knowledge preservation and production.

• Portal - cultural preservation through data science & AI: Already in beta form (beta.shariasource.com), we plan to expand the SHARIAsource Portal to scale up—with identification, aggregation, and digitization of all sources of Islamic law and history. We will dynamically organize and present the data to provide universal access, and create digital and other AI tools for dynamic use. AI tools slated for development that will facilitate access and education include: a "StackLife" tool to identify and filter local and global holdings in sources of Islam together with an ability to virtually browse them, an Arabic OCR tool to convert digital material to machine readable text-as-data, and a network mapping tool for historical figures and sources through algorithms and machine learning developed in collaboration with subject-matter experts on how to read and make sense of the texts.
• People - data driven research and innovation: We will expand the collaborations with leading scholars of Islamic law and history from around the world who provide subject-matter expertise and analysis, and we will foster partnerships with global libraries and tech innovators to facilitate access and new research.
• Publication - access and education: We will publish and otherwise disseminate the sources, related open-access AI tools, and the cutting-edge research emerging out of increased access to both through our newly launched Blog, Journal, and Book Series in Islamic Law (in agreement with Harvard University Press), to inform the world’s knowledge about Islamic law and history globally. 

Our platform will make it possible to see, for the first time, how Islamic law developed and changed over the course of time and place and from one school of thought to another. The Muslim world is plagued by social, economic, and political challenges. The ability to preserve, access, and analyze Islamic sources can play an important role in changing notions of law, finance and development, crime and punishment, women’s rights, and education for the better. This approach is unique: it combines experts and expertise in Islam and data science. It is possible: we have access to one of the most extensive collections of materials—curated over the past 300 years at Harvard—and to partners who are leading innovators in data science and AI methods. 

We envision an on-line web portal where 1400 years of digitized Islamic historical and contemporary records will be archived, annotated, and publicly searchable. Importantly, the portal will host a collection of tools that allow for advanced semantic and syntactic search and analysis of documents, the relationships between documents, and the relationships between authors of documents over time.

A preliminary version of the portal was developed in 2017 by the Program in Islamic Law, available online at beta.shariasource.com. To attain our overall vision, the portal will be enhanced in three phases:

• Phase 1 [Start-up]: We will develop a set of data science tools that autonomously search for, identify the names of, and pinpoint the physical locations of, historical Islamic source materials. To identify this information, the search tools will ingest data from public and private library web Application Programming Interfaces (APIs) from Harvard University, Princeton University, and the Library of Congress.
• Phase 2 [Ramp-up]: We will select an exemplary subset of the identified source materials from Phase 1 (from the Harvard Library), digitize the contents of the identified subset, and develop tools for:
  • Automated transcription of documents via Optical Character Recognition (OCR).
  • Human transcription of document text using crowd-powered web-based tools.
• Phase 3 [Steady-state]: We will utilize the transcribed data from Phase 2, along with document meta-data (e.g., date of publication, author, geography, etc.) to visualize and map the evolution of Islamic ideas presented within the sources as a function of time, space, and geopolitical shifts. 

Our efforts will make it possible, for the first time, for local populations in the Middle East and the larger Muslim world to access their own histories and make informed determinations about the laws that govern them. There are ample historical examples of how access to knowledge has empowered positive change in the Muslim world. In Pakistan for instance, women who were victims of rape were themselves punished under the country’s first Islamic criminal law code (the “Hudood Ordinances”) when they pushed for prosecution. It was only after scholars and activists grounded their opposition to these practice in Islamic sources, which prohibited punishment of victims of sexual violence, that lawmakers and advocates implemented reform.

We are currently serving 5000 people. The SHARIAsource Portal has 120 accounts of scholars from around the world who have advanced degrees in Islamic law or data science, and are recognized by their peers as leaders or rising stars in their fields. Our mailing list comprises 2063 recipients. And our blog currently gets 2200 views per month, with 1100 unique visitors per month.

In one year, we plan to double that number of users and readers to 10,000 people. In five years, we plan to multiply that number many-fold with 1,000,000 readers, researchers, and data-scientists using or benefiting from the tool around the world.

We are seeking approximately $10.6 million in funding to support the 3 phases of development over the next 3 years.

The startup phase of the project (Phase 1, July 2019 – September 2019) will require approximately $250,000 of support to enable the rapid development of engineering foundations necessary for the larger fundraising and development activities of Phase 2.

The ramp-up phase of the project (Phase 2, October 2019 – October 2020) will require the convergence of engineering efforts, research efforts, and data curation efforts. As the most critical phase of the project, it will cost approximately $5 million. A vast majority of these funds ($3 million) will be spent for data acquisition, digitization, and transcription of both printed and fragile historical documents in the Harvard University archives. Importantly, the digitized and transcribed data will serve as the cornerstone of the efforts of engineering staff, historians and legal researchers, and AI experts for the development of the platform and its associated tools.

Upon completion of the ramp-up phase, the project will enter steady-state (Phase 3, October 2020 – October 2021). At this stage, the core technology and archiving work will be 50% complete, and our project will be well-poised to attract additional funding through grants and licensing arrangements. 

Aside from finances, the main barriers have to do with resources, geopolitics, and culture/law. In 1258, the House of Wisdom in Baghdad, Iraq was the intellectual center of global knowledge and discourse. It housed classic and contemporary works of Arab, Greek, Persian, Indian, Chinese, and Egyptian thinkers. In that same year, the Mongols laid siege to Baghdad, destroying the city and tossing its collection of books into the Tigris river. Survivors described the river as black from the ink of knowledge forever lost. Today, access to sources for the global cultural heritage of Islam – a 1400-year history – are once again rapidly disappearing through war’s destruction (Syria, Yemen, Iraq) and time’s decay. What sources remain are limited to a small number of libraries that are difficult for the public to access.

Regional preservation is unlikely; political and/or religious-sectarian sensitivities makes it difficult for neutral partners to collaborate in the Middle East. A sense of ephemeralness, impossibility, and delay pervades in an age that demands a sense of urgency. Meanwhile, sharīʿa is more important than ever before all over the world. Islam appears in news headlines daily, it directly affects over 1.5 billion Muslims around the world, and knowledge of it drives law and policy in the US, in Muslim counties, and globally. We propose to solve the resources problem by combining expertise and resources in Islam and data science, by opening up the unique collection of sources at Harvard and making possible universal access through AI tools.

We plan on overcoming barriers to knowledge acquisition of Islamic sources by providing access and tools for use. The preservation of the global cultural heritage of Islam will require innovative collaborations between law and history scholars and data scientists to (1) digitize and navigate Islamic content, (2) reliably place that content into historical, social, and geo-political context, and (3) disseminate information in a manner in which key institutions and the general public may engage. These elements must combine to inform the significant conversations around law and society related to Islam that are ever more vibrant today. 

This preservation effort will require a remarkable lineup of institutional infrastructure and vision, access to global sources and pathways to local partnerships, and subject-matter expertise in Islam and data science—all with a commitment to getting it done over the long term. We combine these elements.

The Program in Islamic Law at Harvard Law School houses SHARIAsource and will lead this overall effort. The Program is dedicated to promoting research and providing resources for the academic study of Islamic law. It hosts a suite of projects toward that end: a Portal for organizing the world’s information on Islamic law (SHARIAsource), a set of Publications for cutting-edge scholarship in Islamic legal studies that emerges from the sources (a book series, occasional papers, and a peer-reviewed journal), and myriad Program events and support for students, scholars, and partners working to build and disseminate knowledge in fields of Islamic law and history through traditional methods and AI. The Program works with the Harvard Library, the Institute for Quantitative Social Science, and a global team of scholars and institutional partners to advance Islamic law scholarship, including the use of data science and AI tools to facilitate new research and analysis in this field. The initial SHARIAsource platform was built with support from the MacArthur Foundation and the Luce Foundation, as a promising first step to address the problem of access to knowledge in the world.

We have 14 people on our team: 1 full-time Program Director at the Program in Islamic Law at Harvard Law School (Professor Intisar Rabb), 2 data science partners at Ghamut (Mohammad Ghassemi and Tuka Alhanai), 2 full-time staff members at PIL (Sharon Tai and Nataly Castro), and 5 part-time or contractor workers for discrete software development and web editing (Jasmine Jemeison, Kāmil Ciemniewski, Sebastian Diaz, Salma Kazmi), and 5 staff partners (Kevin Garewal, Gayle Fischer, Jessica Fjeld, Jocelyn Kennedy, Elizabeth Hess).

This preservation effort will require a remarkable lineup of institutional infrastructure and vision, access to global sources and pathways to local partnerships, and subject-matter expertise in Islam and data science—all with a commitment to getting it done over the long term. We combine these elements.

Professor Intisar Rabb leads the Program in Islamic Law and SHARIAsource as faculty director. She is a Professor of Law and History at Harvard, and holds an appointment at its Institute for Quantitative Social Science. She has published on Islamic law in historical and modern contexts, including the monograph, Doubt in Islamic Law (2015), two edited volumes, and numerous articles on Islamic legal history, Islamic constitutionalism, and the early history of the Qur'an text. She received a BA from Georgetown, a JD from Yale, and an MA and PhD from Princeton. She has conducted research in Egypt, Iran, Syria, and elsewhere.

Dr. Ghassemi (https://ghassemi.xyz) and Dr. Alhanai (talhanai.com), combined, have over 12 years of experience working in technical and strategic consulting with numerous Fortune 500 companies (e.g. Allstate, Samsung, etc.), , and have built machine learning technologies. Together, they have published over 30 peer-reviewed papers in several highly-respected scientific journals, including: Nature Scientific Data and Science Translational Medicine. Dr. Ghassemi was involved in the de-identification, storage, and distribution of MIMIC-III, a research database that contains the electronic health records of approximately 60,000 critical care patients. Dr. Alhanai has also worked extensively with Big Data, including data from the illustrious Framingham Heart Study.

SHARIAsource currently has 3 partner Organizations:

• The Harvard Library – provides the core sources of materials and organization / metadata for Islamic sources, with one of the largest collections in the world, curated over the past 300 years.
• The Institute for Quantitative Social Sciences – provides the data backbone for the project and space (see Dataverse) where an AI team, once hired, can plug in, collaborate, and grow the project.
• The Ghamut Corporation—provides leadership for the technical development in Phase 1 of the project, and guidance for the engineering and AI research efforts in Phases 2 and 3.

Ghamut was founded by Prof. Mohammad Ghassemi and Dr. Tuka Alhanai in 2016 while they were pursuing their Computer Science doctoral degrees at the Massachusetts Institute of Technology (MIT).  Ghamut’s mission is to use AI technology to solve the world’s most pressing problems. The firm has extensive first-hand experience building systems that store, process, and derive value from large volumes of data. The company has received entrepreneurial awards from MassChallenge and the MIT Sandbox Innovation Fund.

Upon completion of the ramp-up phase, the project will enter steady-state (Phase 3, October 2020 – October 2021). At this stage, the core technology and archiving work will be 50% complete, and our project will be well-poised to attract additional funding through grants and licensing arrangements. The cost of operations at steady state is approximately $2 million per annum (not including the additional $3 million required for continued digitization), and is estimated to be the long-term yearly operational cost to maintain and continuously grow this project. 

To meet our annual funding requirement of $2m for operations and $3m for further digitization, we will submit applications to several granting bodies, and also sell licenses to financial, legal, and research institutions that may benefit from a subscription to the platform’s content. For an annual license, we propose charges to for-profit ventures of $100,000 per year, and to not-for-profit ventures of $50,000 per-year. In the spirit of empowering the public, all content will be freely available to non-commercial, non-academic users. The proposed project will be performed within a (yet-to-be formed) 501(c)(3) tax-exempt, non-profit organization whose long-term mission will be the preservation of the world’s heritage through technology. Individuals, or representatives of institutions/governments, who contribute in excess of 10% of the funding target of this proposal will be invited to join a board of advisers who will convene annually in Cambridge, Massachusetts to receive updates on our progress, and provide guidance on our ongoing efforts.   

We have a great concept, and all of the puzzle pieces align to bring a much needed picture of cultural heritage preservation – of Islam’s historical documents – back into view. We seek bridge funding for the immediate next steps in gaining software developers to build AI tools, and connections to partnerships that will help accelerate the work in business planning, information architecture, fundraising, and scaling of our solution.

N/A

- Business development solutions for a research / academic context

- Global libraries and repositories for information

We envision an on-line web portal where 1400 years of digitized Islamic historical and contemporary records will be archived, annotated, and publicly searchable. Importantly, the portal will host a collection of tools that allow for advanced semantic and syntactic search and analysis of documents, the relationships between documents, and the relationships between authors of documents over time.

In tandem with the technology development process, we will expand collaborations with leading scholars of Islamic law and history from around the world who will provide subject-matter expertise and analysis, and we will increase partnerships with global libraries and tech innovators to facilitate access and new research. To foster interest, recruitment, and sponsorship, we will present our progress toward the development of the platform, and its associated tools to universities and research institutes, the broader Islamic community, and other venues. We will publish and otherwise disseminate the sources, related open-access AI tools, and cutting-edge research emerging out of increased access to both, through our newly launched Blog, Journal, and Book Series in Islamic Law (in agreement with Harvard University Press), to inform the world’s knowledge about Islamic law and history globally – with primary source materials, reliable expert commentary, and AI tools to process both.      

Our leader is a woman who well understands focus on women in Muslim societies and beyond. In tandem with the technology development process, we will expand collaborations with leading scholars of Islamic law and history from around the world who will provide subject-matter expertise and analysis, and we will increase partnerships with global libraries and tech innovators to facilitate access and new research -- with particular emphasis one women. To foster interest, recruitment, and sponsorship, we will present our progress toward the development of the platform, and its associated tools to universities and research institutes, the broader Islamic community, and other venues. We will publish and otherwise disseminate the sources, related open-access AI tools, and cutting-edge research emerging out of increased access to both, through our newly launched Blog, Journal, and Book Series in Islamic Law (in agreement with Harvard University Press), to inform the world’s knowledge about Islamic law and history globally – with primary source materials, reliable expert commentary, and AI tools to process both.      

N/A

N/A

N/A

N/A