Panel 3 International Mathematical Knowledge Trust IMKT: Update on - - PowerPoint PPT Presentation

Panel 3 International Mathematical Knowledge Trust IMKT: Update on the Global Digital Mathematics Library

Rio de Janeiro 7 August 2018

Rio de Janeiro Update on the Global Digital Mathematics Library

Panel Introduction

Patrick D. F . Ion (moderator) Stephen M. Watt Gadadhar Misra Alf A. Onshuus Liu Zheng Thierry Bouche

Rio de Janeiro Update on the Global Digital Mathematics Library

Stephen M. Watt

Dean of Mathematics, University of Waterloo, Ontario, Canada Western University, Ontario; IBM T.J. Watson Research Center; INRIA and University of Nice Computer algebra: co-author of Maple and Axiom, Aldor; W3C MathML and InkML

Rio de Janeiro Update on the Global Digital Mathematics Library

In International Mathematical Knowl wledge Trust - IM IMKT:

an Update on the Global Digital Mathematics Library

Stephen M. Watt

David R. Cheriton School of Computer Science University of Waterloo

2018 International Congress of Mathematicians Rio de Janeiro, Brazil 2018-08-01 – 2018-08-09

Mathematical Knowledge

Can be

• Of permanent relevance,
• Represented exactly,
• Treated mechanically.

2 × 3 × 7 = 42

Global Digital Math Library – a longstanding vision

• Many advocates over time, likely including some in this room.
• Envisaged as linked digitization of (much or) all of past math literature in

an enhanced, open article repositories.

• Scope: 106 → 107 items.

IMU GDML Working Group

Initiated by Ingrid Daubechies at Seoul ICM. Thierry Bouche (U Grenoble, France) Bruno Buchberger (RISC Linz, Austria) Patrick Ion, chair (Math Reviews, USA) Michael Kohlhase (FAU Erlangen-Nürnberg, Germany) Jim Pitman (UC Berkeley, USA) Olaf Teschke (ZB Math, Germany) Stephen Watt (U Waterloo, Canada) Eric Weisstein (Wolfram Research, USA)

WG Remit

The GDML WG is charged with the tasks of

• designing a road map for the practical next steps towards the GDML
• determining its organizational structure
• prioritizing the different requirements for its implementation
• estimating an incremental budget, both start-up and sustaining funds
• fostering the writing of proposals to funding organizations

WG Activities

• Workshops, sessions and panels
• Specific presentations and conference presentations
• Seed funding grant
• Founding not-for-profit
• Seed projects

What the Future Was

• A linked repository of articles is nice, but don’t we have this already?

The Present Day

• Article Repositories

Publisher archives www.springer .com/online+journals+archive JSTOR www.jstor.org arXiv arxiv.org EuDML eudml.org NUMDAM numdam.org GöttingerDigitalisierungszentrum gdz.sub.uni-goettingen.de/gdz RusDML

www.rusdml.de/rusdml

Private Lists www.math.uni-bielefeld.de/~rehmann/DML …

The Present Day

• Indexes, Reviews, Author Databases

Inspec inspecdirect.theiet.org zbMATH zbmath.org MathReviews/MathSciNet www.ams.org/mathscinet Math Genealogy Project www.genealogy.math.ndsu.nodak.edu . . . . . .

The Present Day

• General Summary Works

Wikipedia Mathematics Project www.wikipeda.org MathWorld mathworld.wolfram.com Wolfram Functions Site functions.wolfram.com NIST Digital Library of Mathematical Functions dlmf.nist.gov Digital Repository of Math Formulae

gw32.iu.xsede.org/index.php/Main_Page

…

The Present Day

• Specialized Tools and Databases

Online Encyclopedia of Integer Sequences oeis.org Dynamic Dictionary of Mathematical Functions ddmf.msr-inria.inria.fr Online Integral Calculator www.wolframalpha.com/calculators/integral-calculator Inverse Symbolic Calculator isc.carma.newcastle.edu.au Atlas of Finite Simple Groups brauer.maths.qmul.ac.uk/Atlas/v3 L-functions and modular forms database www.lmfdb.org Combinatorial Statistic Finder www.findstat.org A Catalogue of Lattices www.math.rwth-aachen.de/~Gabriele.Nebe/LATTICES Encyclopedia of Triangle Centers faculty.evansville.edu/ck6/encyclopedia/ETC.html …

The Present Day

• Proof Libraries

Big Proofs (4 Colour, Feit Thompson, Kepler, …) Mizar Mathematical Library www.mizar.org/library Archive of Formal Proofs www.isa-afp.org MetaMath Proof Explorer us.metamath.org/mpegif/mmset.html …

The Present Day

• Other Projects and Activities

Encyclopedia of Continued Fractions s

blog.wolframalpha.com/2013/05/16/computational-knowledge-of-continued-fractions

Maple Function Advisor r

www.maplesoft.com/support/help/MapleSim/view.aspx?path=FunctionAdvisor

Maple impl. of Exact Solutions of the Einstein Field Equations s jointmathematicsmeetings.org/amsmtgs/2181_abstracts/1116-00-2335.pdf GAP Package liealgdb www.gap-system.org/Packages/liealgdb.html Matrix Market math.nist.gov/MatrixMarket EqWorld eqworld.ipmnet.ru Formulasheet formulasheet.com SymbolicData wiki.symbolicdata.org …

Current Direction

• From page images

to parsed pages to semantic mark up to corpus analysis

• Support tools for dealing with the meaning of content.
• Mathematical Digital Library
• Mathematical Knowledge Base
• Mathematical Knowledge Tools

Pando, the Trembling Giant

Quaking aspen, estimated

• 47,000 trunks

433,000 m2

• 6,000,000 Kg

80,000 y.o.

Credit: Google Maps Credit: J. Zapell/Public domain

IMKT Projects

• F Abstracts
• Formal Harmony
• Special Function Concordance
• Mathematical Document Classification

(N-Grams and Machine Learning)

William Casselman Tom Hales

Several Groups’ Work

Slightly Less Obvious Statements about Getting Something Done

• Can afford to hand-annotate/capture a small subset
• Rely on improvements of technology/AI

By Hand Supervised automation Automation + checking Automation J?

• Approximation goes a long way
• Future-proof the data

Open Participation and Return on Effort

Gadadhar Misra

Indian Institute of Science, Bangalore Vice-President of the Indian National Science Academy for publications and informatics Operator theory

Rio de Janeiro Update on the Global Digital Mathematics Library

Gadadhar Misra, Indian Institute of Science

Preprint Archives in India

Preprint archives provide a platform for permanently storing sofu copies of scientific manuscripts, ofuen before peer-review, with

• pen access to any interested person.

PREPRINTS

“Scientists issue preprints to speed dissemination, establish priority, obtain feedback, and 


• fgset publication bias”.

NIH Notes:

The ArXiV

❖ With the advent of internet, one of the first and popular

• nline preprint archives was ‘arXiv’ (http://arXiv.org)

which in over 25 years of its existence, strongly influenced many publishers and impacted how science is disseminated.

The ArXiV

❖ With the advent of internet, one of the first and popular

• nline preprint archives was ‘arXiv’ (http://arXiv.org)

which in over 25 years of its existence, strongly influenced many publishers and impacted how science is disseminated.

❖ The ArXiV originated at Los Alamos (for physics) and at

Duke University (for mathematics).

The ArXiV

❖ With the advent of internet, one of the first and popular

• nline preprint archives was ‘arXiv’ (http://arXiv.org)

which in over 25 years of its existence, strongly influenced many publishers and impacted how science is disseminated.

❖ The ArXiV originated at Los Alamos (for physics) and at

Duke University (for mathematics).

❖ These sites had then been combined at the Los Alamos

(LANL) centre and have since moved to Cornell University.

The ArXiV

ArXiV @IMSc

❖ The IMSc mirror at https://in.arxiv.org is a mirror site for

ArXiV that has been functional since the late nineties.

ArXiV @IMSc

❖ The IMSc mirror at https://in.arxiv.org is a mirror site for

ArXiV that has been functional since the late nineties.

❖ Currently, the mirror is connected to the internet with an

sufgiciently high-speed link so that it is up-to-date within about 20-30 minutes from the time that the primary site is updated.

ArXiV @IMSc

❖ The IMSc mirror at https://in.arxiv.org is a mirror site for

ArXiV that has been functional since the late nineties.

❖ Currently, the mirror is connected to the internet with an

sufgiciently high-speed link so that it is up-to-date within about 20-30 minutes from the time that the primary site is updated.

❖ There is enough processing power and disk-space to last

us for quite a while.

ArXiV @IMSc

Open Access India

Open Access India is in the process of forming a steering committee to manage a Preprints Repository for India, indiaRxiv.

Indian Academy of Sciences Online Repository

“This is an Open Access Repository of publications of Fellows of the Indian Academy of Sciences. It attempts to collect, preserve and disseminate the intellectual output of the Academy available as publications by its Fellows in peer-reviewed journals.”

Indian Institute of Science ePrint Archives

“ePrints@IISc repository collects, preserves and disseminates in digital format the research output created by the IISc research community. It enables the Institute community to deposit their preprints, postprints and other scholarly publications using a web interface, and organizes these publications for easy retrieval. ”

episciences.org

Episciences.org is an innovative combination

• f the two routes of free access:

the gold route by hosting journals in open access (overlay journals) the green route where articles are submitted to these journals by depositing them in an open archive +

The Hardy-Ramanujan Journal

• n Episciences

Declaration By Open Access Communities & Attendees of OpenCon 2018, New Delhi

We will strive to publish our interim research outputs as preprints or postprints (e.g. Institutional Repositories) and encourage our peers and supervisors to do the same to 
 make our research open and actionable in a timely manner.

3 FEBRUARY 2018

http://openaccessindia.org/

Recommendation By The Indian National Science Academy (INSA)

Various agencies/organizations in India that fund research should take cognizance of articles that have been deposited in established free open access Pre-Print Archives as a proof

• f prior-data.

2 JUNE 2018

Alf A. Onshuus

Universidad de los Andes, Colombia IMU CEIC and CDC Logic and Foundations of Mathematics, Number Theory and Algebra

Rio de Janeiro Update on the Global Digital Mathematics Library

Some ideas from developing countries for GDML

Overview

• Most universities in developing countries, even research universities,

don’t subscribe to all the main mathematical journals.

• Article access is achieved through:
• 1. Personal contacts,
• 2. Library access of researchers Alma Mater.
• 3. Library agreements.
• Arxiv has become a very mainstream tool.
• Internet access and speed, has become better in most places. Speed
• f downloading does not seem a significant obstacle, contrary to

what was happening some years ago.

Some initiatives in Colombia to access mathematical literature.

• Library agreements for sharing journal articles.
• Joint access to MathScinet from many small universities.

Challenge we are talking about.

The “Gold Open Standard” initiative has been discussed. Changing the model from a “pay‐to‐read” to a “pay to publish” model can have complicated consequences for scientists in Colombia and the developing world who do not, in general, belong to institutions with large number of subscriptions to journals.

Some quotes.

“It would be great to have access to information at more basic levels than what is found in Arxiv. For example, to support a graduate student in Cambodia or Honduras. I have not explored what is available but what I have seen is at the level of research that a student (and most teachers) in many developing countries could not read. Something that would be of special interest at the level of developing countries would be to have documents edited by experts on how one enters some areas of research.”

A couple more replies.

• “This is a very interesting initiative. It should be open access. One

thing to explore is to map mathematical knowledge, understanding for example how many written articles are in the interface of two subfields.”

• “It would be important to discuss first the balance between novelty

and soundness, between the number of publications and their quality, between open source and funding for this purpose.”

Liu Zheng

National Science Library, Chinese Academy of Sciences Collection & Knowledge Organization Center

Rio de Janeiro Update on the Global Digital Mathematics Library

National Science Library, Chinese Academic Sciences.

Global digital mathematics library views from China

Liu Zheng Aug, 7, 2018

2 page National Science Library，Chinese Academic Sciences.

Mathematics in China

• Mathematics education

Graduate （people） University and research institutions* Bachelor 66100-71500 649 Master 6000 246 Doctor 1000 75 *statistical data in 2017

50 100 150 200 250 mathematics basic mathematics computationalmathematics probability theory and mathematical statistics applied mathematics

• perations research and cybernetics

Universities and research institutions

Doctor Master 100 200 300 400 500 600 mathematics and applied mathematics information and computing science mathematical basic science

Universities and colleges

bachelor

3 page

Mathematics in China

• Mathematics research
• Number of papers published

increased year by year

• The funding for Mathematics

research is 500 million yuan by National natural science foundation of China

• The quality of mathematics

research is still low

6457 7541 8759 10207 12125 13594 14057 71.0% 82.0% 85.9% 88.2% 90.3% 90.4% 91.6% 3000 6000 9000 12000 15000 18000 2009 2010 2011 2012 2013 2014 2015 0% 20% 40% 60% 80% 100% paper funding 受资助的论文数 论文的受资助比例

4 page National Science Library，Chinese Academic Sciences.

Digital Mathematics Library in China

• No digital mathematics library
• By electronic database subscription
• Construct mathematics library homepage, subject portal, mathematics

electronic journal navigation

• Very few mathematical papers submitted to reprint platform、institutional

repository

5 page National Science Library，Chinese Academic Sciences.

We hope and suggest

GDML

Decreasing in time for retrieval & discovery Understanding the progress

• f similar research topic

Introducing Chinese research results to international peers Promoting the mathematics research

6 page National Science Library，Chinese Academic Sciences.

We hope and suggest

• From document retrieval to domain knowledge structure discovery

7 page National Science Library，Chinese Academic Sciences.

We hope and suggest

Integrated and sharing warehousing Relationship network of mathematical researcher Intelligent tools/assistants

Collection: Journal Proceeding of conference Monographs Textbooks Tutorial papers …… Unit: University college institution Person: student teacher researcher user Domain subject: Formularies mathematical modeling Philosophy of mathematics

Classi ficati

• n

article dictio nary wiki

Basic data

湿地 植物 动物 insitu tion Perso n Doc 地理

entities

Domain entities general entities

？ Doc event Institut ion person ？ model formul a applie d

concept relation

Thierry Bouche

Universit Grenoble Alpes, Institut Fourier EuDML, Cellule Mathdoc; IMU CEIC & GDML WG Global analysis; Digitisation, e-publishing, DML

Rio de Janeiro Update on the Global Digital Mathematics Library

The Digital Mathematics Library A European perspective

Thierry Bouche

Cellule Mathdoc, Université Grenoble Alpes

ICM  Rio de Janeiro August th, 

DML OA Examples

The Mathematics Library

The libraries’s main functions are: Selection of collections by subject, document type Not necessarily formally published: theses, reports, preprints.. . Some grey literature needs to be archived or made accessible to patrons (e.g. Nash, Perelman.. . ) Acquisition of actual documents (books, journals.. . ) Cataloguing for collection maintenance and easy retrieval Archiving and Preservation of the curated collections Access provision maintained for patrons

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 1 / 23

DML OA Examples

The Digital Mathematics Library

EuDML tries to recreate this in the digital era A curated digital archive Physically hosted at public (at least non-profit, science-driven) institutions Free to use (possibly after some moving wall) (Conceived as a regional integrator ready for worldwide integration, but missing international partners.. . ) GDML/IMKT could add Worldwide content Enhanced with intelligent math-aware agents More accessible to humans with disabilities Free to mine (more accessible to machines)

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 2 / 23

DML OA Examples

The Digital Mathematics Library

EuDML tries to recreate this in the digital era A curated digital archive Physically hosted at public (at least non-profit, science-driven) institutions Free to use (possibly after some moving wall) (Conceived as a regional integrator ready for worldwide integration, but missing international partners.. . ) GDML/IMKT could add Worldwide content Enhanced with intelligent math-aware agents More accessible to humans with disabilities Free to mine (more accessible to machines)

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 2 / 23

DML OA Examples

The Digital Mathematics Library

Vision A global (distributed) facility dedicated to archive historical (digitised) material as well as newly published (born digital) An up-to-date registry of all available resources Mechanisms for interlinking the holdings with existing and future infrastructures Seamless navigation across the whole corpus Instant and perpetual access

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 3 / 23

DML OA Examples

The mathematical literature

The reference mathematical corpus is About ,, items (135·106 pages?) Heavily multilingual Growing by about , new items a year.. . (mostly English)

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 4 / 23

DML OA Examples

The Digital Mathematics Library

Warning

Beware confusion DML vs. repositories mixing various flavours of eprints

= ⇒ We need reliable references

Trusted archive (original sources) vs. Encyclopedia

= ⇒ we need access to the original sources!

Mathematical corpus vs. math. knowledge as big data

= ⇒ True vs. plausible theorems?

However our main current problem is that we lack support to significatively enlarge the available content

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 5 / 23

DML OA Examples

The Digital Mathematics Library

Warning

Beware confusion DML vs. repositories mixing various flavours of eprints

= ⇒ We need reliable references

Trusted archive (original sources) vs. Encyclopedia

= ⇒ we need access to the original sources!

Mathematical corpus vs. math. knowledge as big data

= ⇒ True vs. plausible theorems?

However our main current problem is that we lack support to significatively enlarge the available content

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 5 / 23

DML OA Examples

Open access

Mathematicians have an overall favorable attitude towards Open Access, but not at any cost Good options Eventual perpetual OA to all (through GDML/IMKT) Diamond OA, with some regulation (e.g. funded and monitored by a scientific organization, selective support by library goups.. . ) Reasonably priced subscription journals are better than crap OA journals! Concerns An obvious dangerous cost: APCs Putting societies’ publications at risk Research integrity might be endangered by open access mandates that could encourage bad practices

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 6 / 23

DML OA Examples

Open access

Mathematicians have an overall favorable attitude towards Open Access, but not at any cost Good options Eventual perpetual OA to all (through GDML/IMKT) Diamond OA, with some regulation (e.g. funded and monitored by a scientific organization, selective support by library goups.. . ) Reasonably priced subscription journals are better than crap OA journals! Concerns An obvious dangerous cost: APCs Putting societies’ publications at risk Research integrity might be endangered by open access mandates that could encourage bad practices

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 6 / 23

DML OA Examples

Open access

Mathematicians have an overall favorable attitude towards Open Access, but not at any cost Good options Eventual perpetual OA to all (through GDML/IMKT) Diamond OA, with some regulation (e.g. funded and monitored by a scientific organization, selective support by library goups.. . ) Reasonably priced subscription journals are better than crap OA journals! Concerns An obvious dangerous cost: APCs Putting societies’ publications at risk Research integrity might be endangered by open access mandates that could encourage bad practices

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 6 / 23

DML OA Examples

Open access

Mathematicians have an overall favorable attitude towards Open Access, but not at any cost Good options Eventual perpetual OA to all (through GDML/IMKT) Diamond OA, with some regulation (e.g. funded and monitored by a scientific organization, selective support by library goups.. . ) Reasonably priced subscription journals are better than crap OA journals! Concerns An obvious dangerous cost: APCs Putting societies’ publications at risk Research integrity might be endangered by open access mandates that could encourage bad practices

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 6 / 23

DML OA Examples

ICSU recommendation

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 7 / 23

DML OA Examples

ICSU recommendation

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 8 / 23

DML OA Examples

The mathematical literature

Towards a not-for-profit system?

Journal articles in MR: commercial vs. non-commercial publishers

• D. E. McClure, AMS 
• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 9 / 23

DML OA Examples

The mathematical literature

Everything on the arXiv?

Percentage of arXiv coverage for main MSC subject classes.

Müller, F., Teschke, O.: Will all mathematics be on the arXiv (soon)?

• Eur. Math. Soc. Newslett.  ()
• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 10 / 23

DML OA Examples

Jussieu call

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 11 / 23

DML OA Examples

Jussieu call

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 12 / 23

DML OA Examples

Math OA

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 13 / 23

DML OA Examples

Free journal network

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 14 / 23

DML OA Examples

ELibM

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 15 / 23

DML OA Examples

Centre Mersenne

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 16 / 23

DML OA Examples

DEAL

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 17 / 23

DML OA Examples

SciPost

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 18 / 23

DML OA Examples

SciPost

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 19 / 23

DML OA Examples

Peer Community in X

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 20 / 23

DML OA Examples

Open Library of Humanities

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 21 / 23

DML OA Examples

Knowledge unlatched

• Th. Bouche (Grenoble)

EuDML ICM Rio, 07/08/2018 22 / 23

Thank you

Thierry BOUCHE Institut Fourier & Cellule Mathdoc, Grenoble http://www-fourier.ujf-grenoble.fr/~bouche/ Mathdoc director EuDML initiative chair EMS Publications and Electronic Dissemination Committee chair IMU Committee on Electronic Information and Communication member

Further Discussion of GDML

Room 103 (Pavilion 5) 10:45 – 11:50, Wednesday 8 August 2018

Rio de Janeiro Update on the Global Digital Mathematics Library

Thank You All