View source for Open Science Information Infrastructure

This is an initial draft of the website.
Its wording, scope, structure, and layout are subject to revision.

__NOTOC__


Research on open, reproducible, and sovereign scientific information systems

We build open and reproducible ways to organise, connect, preserve, and make use of openly accessible scientific information.

Our research spans mathematical knowledge, scholarly knowledge graphs, research software, and sovereign digital infrastructure.

We work specifically with scientific information that is openly available and can be inspected, copied, processed, and redistributed without access restrictions.

This limitation is fundamental to our technical approach. It allows us to investigate transparent, distributed, and content-addressed infrastructure, including technologies such as IPFS, without introducing access-control mechanisms that would undermine openness, reproducibility, or independent verification.

Our work does not generally concern confidential or access-restricted information, sensitive information, or information containing personally identifiable information, or PII.

By concentrating on open scientific information, we aim to create workflows in which the source material, software, processing steps, and results can be independently examined and reproduced.

* [[#Research|Research]]
* [[#People|People]]
* [[#Projects and ecosystems|Projects and ecosystems]]
* [[#Software and demonstrators|Software and demonstrators]]
* [[#Work with us|Work with us]]
* [https://schubotz.org/ Moritz Schubotz]
* [https://www.fiz-karlsruhe.de/de/bereiche/mathematische-informationsinfrastruktur Mathematical Information Infrastructure at FIZ Karlsruhe]

== Our mission ==

Scientific knowledge increasingly depends on digital infrastructure.

Search systems, repositories, knowledge graphs, software archives, and computational services influence:

* what information can be discovered;
* how research results can be verified;
* whether data and software can be reused;
* how knowledge is connected and preserved;
* who controls access to the scholarly record;
* and who can participate in maintaining it.

Open science therefore requires more than openly accessible publications.

It requires information infrastructure that research communities can inspect, operate, adapt, govern, and preserve.

We aim to develop and strengthen infrastructure that is:

* based on open standards;
* implemented through inspectable and reusable software;
* built around openly available scientific information;
* designed for reproducible information processing;
* interoperable across institutions and systems;
* and governed in the public interest.

Our aim is not to create another isolated platform.

We work on components, standards, information models, and connections that contribute to a diverse and interoperable open-science ecosystem.

<!--
LAYOUT NOTE:
The following three sections could appear next to each other as three introductory blocks:

* Scope of the information we work with
* Reproducible information infrastructure
* Sovereign information infrastructure
-->
<!--
=== Scope of the information we work with ===

Our research focuses on scientific information that is openly accessible and can be used within open and reproducible workflows.

This includes:

* openly accessible scholarly publications;
* openly licensed metadata;
* mathematical formulas and concepts;
* public knowledge graphs;
* openly available research software;
* public software repositories;
* openly shared datasets;
* bibliographic and citation information;
* public institutional and project information.

We do not primarily develop infrastructure for information that is:

* confidential;
* commercially restricted;
* subject to access controls;
* covered by professional secrecy;
* sensitive;
* or dependent on processing PII.

This is not only an ethical or organisational limitation. It is also an architectural choice.

Infrastructure for access-restricted or sensitive information requires mechanisms for identity management, authorisation, revocation, auditing, and controlled disclosure.

These mechanisms can be necessary and appropriate in other contexts, but they lead to a different class of technical systems.

By focusing on open scientific information, we can investigate infrastructure in which information may be:

* copied without prior authorisation;
* independently mirrored;
* verified through content identifiers;
* distributed across different institutions;
* processed through reproducible workflows;
* and preserved without dependence on a single service provider.

Even when working with public information, infrastructure should follow responsible practices such as data minimisation, transparent provenance, and careful consideration of the effects of aggregation and reuse.

=== Reproducible information infrastructure ===

Reproducibility requires more than making a dataset or publication available.

A reproducible information workflow should make it possible to determine:

* where the input information came from;
* under which licence it can be used;
* which version of the information was processed;
* which software and software versions were involved;
* which transformations were applied;
* which parameters and configurations were used;
* when the processing took place;
* and how the resulting information or claims can be verified.

We therefore treat provenance, versioning, documentation, persistent identifiers, and executable workflows as core parts of information infrastructure.

=== Sovereign information infrastructure ===

A sovereign information infrastructure allows scholarly communities and public-interest institutions to retain meaningful control over their information, software, standards, and governance.

Sovereignty does not mean isolation.

Scientific infrastructure should remain interoperable across institutions, disciplines, and national borders.

Important characteristics include:

; Transparency
: Software, information models, interfaces, and decision-making processes should be understandable and inspectable.

; Portability
: Scientific information and scholarly records should be exportable through documented, open formats.

; Replaceability
: Individual components and providers should be replaceable without losing access to the underlying information.

; Interoperability
: Systems should connect through open standards, persistent identifiers, and documented interfaces.

; Community governance
: Research communities and public institutions should have a meaningful role in setting priorities and policies.

; Provenance
: Users should be able to determine where information originated and how it was transformed.

; Reproducibility
: Information processing and computational results should be accompanied by sufficient information to inspect and repeat the relevant workflow.

; Independent operation
: Communities and institutions should be able to operate their own components or instances where this is technically and organisationally appropriate.

; Sustainability
: Infrastructure should be designed for maintenance, migration, and long-term preservation.

== Research ==

<!--
LAYOUT NOTE:
The following six research areas could be displayed as a grid of blocks.
A two-column or three-column layout would work well.

Row 1:
* Machine-actionable mathematics
* Scholarly knowledge graphs

Row 2:
* Open research software
* Reproducibility and provenance

Row 3:
* Responsible artificial intelligence
* Federated and decentralised infrastructure
-->

=== Machine-actionable mathematics ===

We investigate how mathematical expressions, concepts, proofs, and documents can be represented for both human and machine use.

Research topics include:

* mathematical information retrieval;
* formula search;
* semantic enrichment of mathematical documents;
* recognition and disambiguation of mathematical entities;
* connections between formulas and their meanings;
* machine-readable representations of mathematical knowledge;
* links between mathematical literature, software, and research data.

Mathematical notation is a particularly challenging form of scientific information.

The meaning of a symbol often depends on its surrounding document, discipline, and local context.

Making mathematics machine-actionable therefore requires more than extracting visual formula structures.

=== Scholarly knowledge graphs ===

We study how publications, researchers, institutions, software, datasets, and mathematical objects can be connected through open knowledge graphs.

Our goal is not simply to produce more metadata.

We aim to create reliable, inspectable, and verifiable paths through the scholarly record.

Relevant topics include:

* persistent identifiers;
* entity linking;
* provenance;
* metadata interoperability;
* Wikidata and other open knowledge bases;
* links between literature, software, and research data;
* federated queries across independent information sources;
* reconciliation between different identifier systems.

=== Open research software ===

Software is a first-class research output.

We work on methods and infrastructure for:

* software discovery;
* software metadata;
* software citation;
* source-code preservation;
* links between software and publications;
* connections to long-term archives such as Software Heritage;
* recognition of software contributions in scholarly communication;
* reproducible software environments;
* open software registries and knowledge graphs.

Research software is both a scholarly output and a dependency of other research outputs.

Preserving a publication without preserving the software needed to reproduce it may leave an incomplete scholarly record.

=== Reproducibility and provenance ===

Computational research depends on more than a publication.

It may also depend on source code, input information, software environments, external services, and undocumented processing steps.

We explore ways to preserve and communicate this context through:

* executable workflows;
* computational notebooks;
* dependency information;
* structured provenance;
* links between claims and supporting artefacts;
* reproducible research environments;
* versioned information pipelines;
* content-based identification of research objects.

=== Responsible artificial intelligence for scientific information ===

Retrieval systems, recommendation systems, and language models increasingly mediate access to scientific knowledge.

We investigate how such systems can support research without becoming opaque gatekeepers.

Important questions include:

* Which sources support an answer or recommendation?
* How was the information selected and transformed?
* Can users inspect the limitations of the system?
* Can the system be independently evaluated?
* Can communities operate or replace its components?
* Are generated statements connected to verifiable sources?
* Can the underlying information and processing steps be reproduced?
* Does the system depend on inaccessible or proprietary information?

We are particularly interested in systems whose outputs can be traced back to open sources and independently verified.

=== Federated and decentralised infrastructure ===

We explore architectures that allow institutions and communities to cooperate without surrendering control to a single central platform.

Relevant concepts include:

* federation;
* decentralised identifiers;
* content-addressed storage;
* distributed knowledge graphs;
* community-operated services;
* synchronisation between independent repositories;
* portable identity and attribution mechanisms;
* peer-to-peer distribution;
* independently verifiable information objects.

Technologies such as IPFS are relevant because they identify content by its cryptographic representation rather than only by the location of a particular server.

This can support:

* verifiable references to specific versions of information;
* replication across independent hosts;
* resilience against the disappearance of a single service;
* reproducible access to exact information objects;
* separation between the identity of information and the provider currently serving it.

Content-addressed systems do not by themselves solve questions of governance, discoverability, preservation, or scholarly quality.

We therefore study them as components within a wider information infrastructure rather than as complete solutions.

== People ==

Our work takes place in and around the [https://www.fiz-karlsruhe.de/de/bereiche/mathematische-informationsinfrastruktur Mathematical Information Infrastructure] department at FIZ Karlsruhe – Leibniz Institute for Information Infrastructure.

The group also collaborates with universities, libraries, archives, open-source projects, Wikimedia communities, and national and international research-infrastructure initiatives.

<!--
LAYOUT NOTE:
The Moritz Schubotz profile could appear as one wide introductory block.

The group-member profiles below could then appear as individual cards in a two-column or three-column layout.

A separate template could later be created, for example:

{{Person
 |name=
 |role=
 |focus=
 |url=
}}

The profiles are deliberately not presented as a table because wide tables render poorly on smaller screens and with longer descriptions.
-->

=== Moritz Schubotz ===

Dr. Moritz Schubotz is a computer scientist and theoretical physicist working on information infrastructure for mathematics and open science.

He heads Research and Projects Mathematics at FIZ Karlsruhe and teaches at Humboldt University of Berlin.

His research connects:

* mathematical information retrieval;
* semantic technologies;
* scholarly knowledge graphs;
* research software;
* Wikimedia;
* decentralised systems;
* open-science infrastructure.

A recurring question throughout this work is how scientific knowledge can remain understandable, verifiable, and reusable when it moves between documents, databases, software, and artificial-intelligence-supported interfaces.

External profiles:

* [https://schubotz.org/ Personal website]
* [https://www.fiz-karlsruhe.de/de/bereiche/lebenslauf-und-publikationen-dr-moritz-schubotz FIZ Karlsruhe profile and publications]
* [https://orcid.org/0000-0001-7141-4997 ORCID]
* [https://scholar.google.com/citations?user=rqUlEN4AAAAJ Google Scholar]

See also:

* [[Publications]]
* [[Teaching and supervision]]
* [[Projects]]
* [[Software]]

=== Group members ===

The following overview is based on the staff information published by FIZ Karlsruhe.

For current roles and affiliations, consult the [https://www.fiz-karlsruhe.de/de/bereiche/mathematische-informationsinfrastruktur#mitarbeitende official staff list].

==== Moritz Schubotz ====

Role:

Head of Research and Projects Mathematics

Research interests:

Machine-readable mathematics, open science, knowledge graphs, research software, and decentralised information infrastructure.

Profile:

[https://schubotz.org/ Personal website]

==== Daniel Mietchen ====

Role:

Researcher

Research interests:

Open research workflows, Wikimedia, Wikidata, citizen science, reproducibility, and human-machine collaboration.

Profile:

[https://www.fiz-karlsruhe.de/de/bereiche/lebenslauf-und-publikationen-dr-daniel-mietchen FIZ Karlsruhe profile]

==== Madhurima Deb ====

Role:

Researcher

Research interests:

Research data, data curation, the MaRDI Portal, and interdisciplinary scientific applications.

Profile:

[https://www.fiz-karlsruhe.de/de/bereiche/lebenslauf-und-publikationen-dr-madhurima-deb FIZ Karlsruhe profile]

==== Ankit Satpute ====

Role:

Doctoral researcher

Research interests:

Mathematical information retrieval, recommendation systems, plagiarism detection, and neural methods.

Profile:

[https://www.fiz-karlsruhe.de/de/bereiche/lebenslauf-und-publikationen-von-ankit-satpute FIZ Karlsruhe profile]

==== Maxence Azzouz-Thuderoz ====

Role:

Doctoral researcher

Research interests:

Open-source research software, software discovery, software citation, source-code search, and Software Heritage.

Profile:

[https://www.fiz-karlsruhe.de/de/bereiche/lebenslauf-und-publikationen-maxence-azzouz-thuderoz FIZ Karlsruhe profile]


==== Noah Giessing ====

Role:

Researcher

Research interests:

Mathematical information retrieval, natural-language processing, machine learning, and mathematical physics.

Profile:

[https://www.fiz-karlsruhe.de/de/bereiche/lebenslauf-und-publikationen-noah-giessing FIZ Karlsruhe profile]


== Principles ==

<!--
LAYOUT NOTE:
The following principles could be shown as small blocks in a grid.

Suggested grouping:

Row 1:
* Open by default
* Community control
* Interoperability rather than isolation

Row 2:
* Replaceable components
* Verifiable information
* Long-term stewardship

Row 3:
* Decentralisation where useful
-->

=== Open by default ===

We prefer:

* open-source software;
* openly documented interfaces;
* open standards;
* openly licensed metadata;
* reproducible methods;
* publicly accessible research outputs.

Our focus is not on opening information that should remain restricted.

Instead, we design infrastructure around scientific information that is already suitable for open use, redistribution, and independent processing.

=== Community control ===

Infrastructure should serve scholarly communities rather than capture them.

Researchers and public-interest institutions should be able to influence:

* governance;
* technical priorities;
* information policies;
* preservation strategies;
* interoperability requirements;
* and the conditions under which infrastructure components are operated.

=== Interoperability rather than isolation ===

No single repository, knowledge graph, or service can represent all of science.

We therefore work towards systems that can exchange information while retaining their institutional and technical independence.

Interoperability should not require all participants to use the same software or depend on the same provider.

=== Replaceable components ===

A service should not become indispensable merely because its information cannot be exported or its interfaces cannot be reproduced.

Where possible, infrastructure should support:

* documented information exports;
* standard protocols;
* modular architectures;
* independent implementations;
* migration between providers;
* local or community-operated instances;
* reproducible reconstruction of derived information.

=== Verifiable information ===

Scientific information systems should expose their sources and transformations.

Users should be able to distinguish between:

* original research outputs;
* curated metadata;
* automatically extracted information;
* algorithmic recommendations;
* model-generated summaries;
* human corrections and annotations;
* derived information produced by reproducible workflows.

=== Long-term stewardship ===

Maintenance is part of research infrastructure.

Documentation, testing, migration, curation, preservation, and governance are not secondary activities.

They determine whether scientific information remains usable after a project or funding period ends.

=== Decentralisation where useful ===

Decentralisation is not an end in itself.

We consider decentralised or federated technologies where they improve:

* independent verification;
* resilience;
* portability;
* reproducibility;
* institutional autonomy;
* long-term accessibility.

A decentralised system can still reproduce centralised power through governance, discovery services, or control of important software components.

Technical distribution must therefore be considered together with institutional and community governance.

== Projects and ecosystems ==

<!--
LAYOUT NOTE:
The following five project areas could be displayed as separate blocks.

Suggested layout:

Row 1:
* Mathematical knowledge infrastructure
* Mathematical Research Data Initiative

Row 2:
* Research software infrastructure
* Open scholarly communication

Row 3:
* National and European infrastructure
-->

=== Mathematical knowledge infrastructure ===

We work on services and methods related to:

* zbMATH Open;
* mathematical search;
* machine-actionable formulas;
* mathematical classifications;
* semantic enrichment;
* links between mathematical literature, software, and research data.

=== Mathematical Research Data Initiative ===

The Mathematical Research Data Initiative, or MaRDI, develops infrastructure and standards for mathematical research data.

Our interests include:

* the MaRDI Portal;
* mathematical knowledge graphs;
* metadata standards;
* links between publications, software, and datasets;
* reusable research workflows;
* support for FAIR mathematical research data.

=== Research software infrastructure ===

We contribute to efforts involving:

* swMATH;
* software metadata;
* software citation;
* source-code discovery;
* Software Heritage;
* links between research software and the scholarly literature;
* open registries of research software;
* persistent and verifiable references to source code.

=== Open scholarly communication ===

We support approaches to scholarly communication based on:

* open access;
* community governance;
* transparent evaluation;
* reusable metadata;
* open licensing;
* sustainable public infrastructure;
* verifiable links between claims and sources.

=== National and European infrastructure ===

Our work connects to broader initiatives such as:

* the German National Research Data Infrastructure;
* the European Open Science Cloud;
* disciplinary and interdisciplinary research-data initiatives;
* international persistent-identifier and metadata communities.

A useful practical test for infrastructure sovereignty is:

Can a research community understand the system, copy and export its information, replace a component, operate an independent instance, and continue its work if a provider or project disappears?

== Software and demonstrators ==

<!--
LAYOUT NOTE:
Software projects and demonstrators could be shown as cards.

Each card could contain:

* project name;
* short description;
* project URL;
* source-code URL;
* documentation URL;
* status.
-->

=== RelaX ===

[https://relax.labs.schubotz.org/ RelaX] is an interactive relational-algebra calculator for teaching, learning, and experimentation.

=== Further resources ===

* [[Software|Research software developed or maintained by the group]]
* [[Datasets|Open datasets and knowledge-graph resources]]
* [[Demonstrators|Prototypes and experimental services]]
* [[Student projects|Student and thesis projects]]
* [[Publications|Publications and research outputs]]

== How we work ==

# Start from public value. Infrastructure should serve research communities and society.
# Work with open scientific information. Our technical approach assumes that the information may be copied, inspected, and redistributed.
# Prefer open standards. Interfaces and information models should be documented and reusable.
# Use replaceable components. No individual service should become an unavoidable dependency.
# Treat maintenance as infrastructure work. Documentation, testing, and migration are part of the scholarly contribution.
# Design for provenance. Users should be able to determine where information came from.
# Support pluralism. Different communities may need different interfaces, governance models, and technical arrangements.
# Connect rather than capture. We contribute to shared ecosystems and avoid creating new information silos.
# Use decentralisation selectively. Distributed technologies should address concrete problems rather than being adopted as ends in themselves.
# Consider environmental and social costs. Sustainable infrastructure must account for computation, energy, labour, and long-term stewardship.

== Work with us ==

We welcome collaboration with:

* researchers;
* students;
* libraries;
* archives;
* infrastructure providers;
* open-source communities;
* Wikimedia communities;
* scholarly publishers;
* standards organisations.

Possible forms of collaboration include:

* joint research projects;
* infrastructure development;
* Bachelor's and Master's theses;
* doctoral research;
* open-source software development;
* software maintenance;
* dataset and benchmark creation;
* interoperability experiments;
* seminars and workshops;
* community-governance and policy work.

Projects should generally align with our focus on open scientific information and reproducible information workflows.

For enquiries, contact [mailto:contact@schubotz.org contact@schubotz.org].

== News and updates ==

<!--
LAYOUT NOTE:
This section could later transclude a separate Updates page:

{{:Updates}}

Individual updates could also be displayed as cards with a date, title, summary, and link.
-->

* [[News]]
* [[Events]]
* [[Publications]]
* [[Open positions]]
* [[Student projects]]

== Further information ==

* [https://schubotz.org/ schubotz.org]
* [https://www.fiz-karlsruhe.de/de/bereiche/mathematische-informationsinfrastruktur Mathematical Information Infrastructure at FIZ Karlsruhe]
* [[About]]
* [[Contact]]
* [[Imprint]]
* [[Privacy policy]]
[[Category:Open science]]
[[Category:Research infrastructure]]
[[Category:Mathematical knowledge management]]
[[Category:Research software]]