Theoretical and technological building blocks for an innovation accelerator

Frank van Harmelen, G. Kampis, Katy Börner, Peter van den Besselaar, Erik Schultes, Carole Goble, Paul Groth, Barend Mons, Stuart Anderson, Stefan Decker, Conor Hayes, Thierry Buecheler, Dirk Helbing

Abstract

The current scientific enterprise operates on a model largely unchanged since the 17th century, relying on monolithic papers and slow, conservative peer-review processes. This system is increasingly inadequate for the Information and Communication Technology (ICT) age, where the volume of data and the speed of discovery outpace human cognitive limits. This paper proposes the "Innovation Accelerator" (IA), a multi-dimensional platform designed to modernize science. By deconstructing papers into machine-readable "nanopublications," redesigning incentive schemes beyond the h-index, and leveraging Semantic Web technologies, the IA aims to transform computers from mere word processors into active research partners.

The Crisis in Modern Science

While science remains the primary driver of technological innovation, the process itself has become inefficient and conservative. The authors identify several critical bottlenecks that hinder progress:

  • Conservatism in Peer Review: The traditional system often penalizes high-risk, high-reward ideas, favoring "safe" research that aligns with established paradigms.
  • The PDF Trap: Scientific results are often trapped in static PDF files. This makes the underlying data, code, and workflows difficult to extract, reuse, or validate, leading to a lack of transparency.
  • Data Hoarding: A significant portion of raw data and "negative" results are never published. This leads to redundant efforts and a pervasive lack of reproducibility across disciplines.
  • Information Overload: The sheer volume of new literature—such as the publication of one Medline paper every minute—exceeds the capacity of any individual researcher to stay current. Current tools do not allow machines to effectively assist in reasoning or large-scale discovery.

The Innovation Accelerator Framework

To address these systemic failures, the authors propose a framework for an Innovation Accelerator built upon four foundational pillars:

1. Replacing the Monolithic Publication

The IA seeks to deconstruct the traditional paper into modular "research objects." These objects include the raw data, the code used for analysis, the specific workflows, and the individual claims. By treating these as independent but linked entities, science becomes more granular and verifiable.

2. Redesigning Incentives

Current academic success is heavily weighted toward citation counts and the h-index. The IA proposes new metrics that reward a broader range of scientific contributions, including data sharing, software development, and high-quality peer-reviewing. This shift is essential to encourage the behaviors necessary for a collaborative, open-science ecosystem.

3. Machine-Readable Semantics

By utilizing Semantic Web technologies (such as RDF and OWL), scientific claims can be made "computable." This allows for automated inconsistency detection, trend analysis, and the identification of implicit links across massive datasets. In this model, the computer moves from being a tool for documentation to a partner in hypothesis generation.

4. Online Collaboration Platforms

The IA envisions a "web of trust" and reputation among scientists, funders, and industry stakeholders. This infrastructure facilitates interdisciplinary work and moves innovation from a linear model (Research → Development → Market) to a cyclic system with continuous feedback loops between science and society.

Nanopublications: The Atomic Unit of Knowledge

A core concept of the IA is the "nanopublication." This is defined as the smallest unit of publishable information: a single machine-readable assertion consisting of a Subject-Predicate-Object triple (e.g., Protein X – inhibits – Enzyme Y) accompanied by its provenance. This "liquid" form of science allows individual claims to be cited, verified, and integrated into larger knowledge networks independently of a full-length narrative paper.

To manage the validity of these claims, the authors suggest moving away from anonymous, one-time peer review toward crowdsourced "futures markets" or "Bet Books." In these systems, experts record their estimated odds on a specific nanopublication being true, creating a dynamic, real-time measure of scientific consensus.

Evidence and Feasibility

The paper highlights several existing building blocks that demonstrate the feasibility of the IA:

  • CIShell/OSGi: A modular framework that facilitates code reuse in network science and epidemic research.
  • Taverna: A workflow management system that enables the in silico reproduction of complex experiments.
  • Scholarly Databases: Systems like VIVO and the Scholarly Database (SDB) provide the semantic infrastructure needed to map researcher networks and funding data.

The urgency of this transition is underscored by a Bayer validation study, which found that 65% of published "new drug target" claims could not be validated by in-house data. This staggering failure rate suggests that the current "monolithic" publication model is failing to provide a reliable foundation for industrial innovation.

Limitations and the "Essential Tension"

Transitioning to an Innovation Accelerator is not without significant hurdles. The authors acknowledge an "essential tension" between the tradition required for quality control and the innovation required for new formats. Changing a system as entrenched as global science requires more than just technology; it requires a massive socio-political shift in how universities and funding agencies evaluate merit.

Furthermore, there is a risk that breaking science into nanopublications could lead to an even more unmanageable "ocean of scientific bits." To prevent this, the IA must be paired with highly sophisticated recommendation engines and semantic filters. The authors emphasize that these concepts must be rigorously tested in small-scale "exploratories" before being scaled to the global scientific ecosystem.

Summary updated

Abstract

Modern science is a main driver of technological innovation. The efficiency of the scientific system is of key importance to ensure the competitiveness of a nation or region. However, the scientific system that we use today was devised centuries ago and is inadequate for our current ICT-based society: the peer review system encourages conservatism, journal publications are monolithic and slow, data is often not available to other scientists, and the independent validation of results is limited. The resulting scientific process is hence slow and sloppy. Building on the Innovation Accelerator paper by Helbing and Balietti [1], this paper takes the initial global vision and reviews the theoretical and technological building blocks that can be used for implementing an innovation (in first place: science) accelerator platform driven by re-imagining the science system. The envisioned platform would rest on four pillars: (i) Redesign the incentive scheme to reduce behavior such as conservatism, herding and hyping; (ii) Advance scientific publications by breaking up the monolithic paper unit and introducing other building blocks such as data, tools, experiment workflows, resources; (iii) Use machine readable semantics for publications, debate structures, provenance etc. in order to include the computer as a partner in the scientific process, and (iv) Build an online platform for collaboration, including a network of trust and reputation among the different types of stakeholders in the scientific system: scientists, educators, funding agencies, policy makers, students and industrial innovators among others. Any such improvements to the scientific system must support the entire scientific process (unlike current tools that chop up the scientific process into disconnected pieces), must facilitate and encourage collaboration and interdisciplinarity (again unlike current tools), must facilitate the inclusion of intelligent computing in the scientific process, must facilitate not only the core scientific process, but also accommodate other stakeholders such science policy makers, industrial innovators, and the general public. We first describe the current state of the scientific system together with up to a dozen new key initiatives, including an analysis of the role of science as an innovation accelerator. Our brief survey will show that there exist many separate ideas and concepts and diverse stand-alone demonstrator systems for different components of the ecosystem with many parts are still unexplored, and overall integration lacking. By analyzing a matrix of stakeholders vs. functionalities, we identify the required innovations. We (non-exhaustively) discuss a few of them: Publications that are meaningful to machines, innovative reviewing processes, data publication, workflow archiving and reuse, alternative impact metrics, tools for the detection of trends, community formation and emergence, as well as modular publications, citation objects and debate graphs. To summarize, the core idea behind the Innovation Accelerator is to develop new incentive models, rules, and interaction mechanisms to stimulate true innovation, revolutionizing the way in which we create knowledge and disseminate information.
BibTeX
@article{harmelen2012theoretical,
  author = {Frank van Harmelen and G. Kampis and Katy Börner and Peter van den Besselaar and Erik Schultes and Carole Goble and Paul Groth and Barend Mons and Stuart Anderson and Stefan Decker and Conor Hayes and Thierry Buecheler and Dirk Helbing},
  title = {{Theoretical and technological building blocks for an innovation accelerator}},
  journal = {The European Physical Journal Special Topics},
  year = {2012},
  doi = {10.1140/epjst/e2012-01692-1},
}
Figure