OTS Open Grant Call

Online Trust And Safety Research Programme
Open Grant Call For Deep Tech Research Proposals And Solutions

1 Nov 2023 – 10 Jan 2024

Background

Harmful online content causes severe threats and damage to people's well-being, costs lives, and undermines societal stability. Being a digitalised and connected country, Singapore is highly susceptible to harms from the online space. Local mainstream media and public agencies are intensively dealing with false claims targeted at Singapore-specific named entities or their identities. While leading international social media platforms are taking steps to address harmful content or misinformation spread, and the private sector offers commercial tools, the effectiveness and timeliness of existing solutions vary significantly when applied in the Singapore context. While we shall tap into global tech advances in the Online Trust and Safety (OTS) space to grow in tandem, Singapore needs context-aware tools to ensure the technological means to defend ourselves from such internet harms.

The OTS Research Programme is a funding initiative from the Ministry of Communications and Information (MCI). It is part of the Smart Nation and Digital Economy (SNDE) domain of the Research, Innovation and Enterprise 2025 programme (RIE2025) administered by the National Research Foundation (NRF). The overarching aim of the OTS Research Programme is to develop technological capabilities to combat online falsehood and harms and align translation efforts with high-impact use cases.

With support from the MCI, NRF and partner organisations, the Agency for Science, Technology and Research (A*STAR) established the Centre for Advanced Technologies in Online Safety (CATOS) in April 2023 to host the OTS Research Programme as its flagship programme. Based in A*STAR’s Institute of High Performance Computing (IHPC), CATOS comprises functions responsible for key strategic pillars for the OTS Research Programme, including Deep Tech Research Pillar (focusing on selecting and funding low-TRL research with high competitive advantages and potential to move to mid-TRL), Systems Engineering Pillar (focusing on mid-high TRL technology evaluation, integration and translation of technology outputs into needle-moving use cases), and Programme Coordination activities (focusing on engaging research and practitioner communities through workshops, forums, and collaborative networks).

CATOS works in close consultation with MCI to prioritise technologies that address use cases affecting Singapore at the national level. Based in Singapore, we differentiate by building context-aware OTS tech capabilities that can take into account the local user behaviour, local context, and local languages. We envisage that upon successful execution, technological capabilities developed from the OTS Research Programme will significantly complement existing legislation and education efforts to detect and deal with the rapidly evolving threat of online harms to Singapore and our society.

Objectives

This OTS Research Programme - Open Grant Call for Deep Tech Research Proposals and Solutions (hereafter referred to as "OTS Open Grant Call") is launched to solicit high-quality projects that can deliver a strong suite of robust technological solutions that address significant unmet problems in the OTS space.

We seek projects that can potentially yield technology R&D outcomes to:

Help Singapore develop technology leadership in the OTS space through research excellence such as notable publications and presentations at top scientific and industry conferences.
Form strong pipeline capabilities to equip Singapore-based companies and public agencies with effective tools to combat online harms.

The projects supported by the OTS Open Grant Call should deliver technologies at mid-TRL in the form of code, software development kits, Docker images or similar, with user documentation that describes the training method and validation results. Deliverables should be ready for industry-strength evaluation to test the delivered technology's robustness in out-of-training/sample test cases that come from real-world or simulated scenarios.

Research Themes and Technology Topics of Interest

The Online Trust and Safety (OTS) space is exceptionally complex and rapidly evolving. To effectively define a technology research agenda and prepare for high-impact translational use cases, CATOS has consulted academia, industry and public agencies and developed a framework to define the key research-worthy technology topics of interest based on MCI's “systems perspective of the online harms space” and the associated challenges identified in each of the systems (Figure 1).

Figure 1. The “Systems of Stakeholders” in OTS Space, Challenges, and Key Research-worthy Technology Topics of Interests over Three Themes

Specifically, the OTS Open Grant Call looks for proposals of technologies tackling online harms under the following three research themes. Ten Technology Topics have been identified under the three Research Themes.

Research Theme 1: Prevention
Research Theme 2: Early Detection
Research Theme 3: Mitigation & Empowerment

Research Theme 1: Prevention

Prevention engines and tools that enhance trust and/or reduce mistrust towards mainstream media and authorities and identify the common sources of mis/disinformation for actions to stem the spread from the root.

Tech 1

Content Provenance & Authenticity for Trustability by Design

A content signature generator system (based on asymmetric encryption and distributed ledger technology) that publishers and content generators can use to allow info consumers to verify the provenance and uniqueness of content. Such a solution can alert the info consumer or info propagator if the content has been flagged or tampered with.

Tech 2

Misinformation and DisInformation Source Attribution

Technology that can help relevant users to identify and attribute sources and key spreaders of identified or suspected mis/disinformation. This technology shall present features such as those that can help users decode digital narratives and achieve a better understanding of unfolding events in the world.

For the Prevention thematic tools, we will prioritise methods that can leverage an advanced understanding of the notion of trust (e.g., [1-3]), and can demonstrate a substantial advancement in existing approaches (e.g., [4-10]) in real-world settings.

Research Theme 2: Early Detection

Early Detection engines that help assess the existence and intensity of harmful online content, including deepfakes, non-factual claims, propaganda, extremism and hate speech.

Tech 3

Multimodal Deepfakes Detector

Technology that detects inauthentic media with audio-visual content manipulation, including but especially the manipulation at the messaging and narrative level through cross-modal manipulation. The models shall facilitate users with research-grounded explainable features to discern manipulation surrounding facial/non-facial regions, speaker/voice manipulation, lip-synching, lighting, language use, and even the malicious vs. non-malicious intent behind the author who posts the content.

Tech 5

Propaganda Detector

Technology that can assist a human user and/or is deployable in systems to automatically detect the presence of propaganda at both the overall level (i.e., propaganda vs. non-propaganda) and the propaganda technique level (e.g., loaded language, name-calling and labelling, flag-waving, doubt). This content can be in news segments, messages or social media content. This engine would help identify propagandistic content harbouring ill intentions against Singapore and our society.

Tech 7

Malicious Account & Bot Detector

Technology that can detect inauthentic user accounts and identify bots, particularly those that undertake malicious activities to generate online harm, such as illegal and ill-intended uses of logos of creditable organisations. This can include technology to detect large numbers of bots that are coordinated and controlled by a group of common users to amplify mis/disinformation.

Tech 4

Non-Factual Claims Detector

Technology that assists a human user and/or is deployable in systems throughout the fact-checking process to verify claims as early as possible and be sensitive to Singapore and the region's multilingual context.

Tech 6

Extremism & Hate Detector

Technology that can automatically detect online extremism and hate speech from social media text/audio/video. This includes anything that encourages hate or violence towards a person/group based on race, religion, ethnic origin, age, gender, sexual orientation, and/or disability. Detectors sensitive to Singapore and the region's specific socio, linguistic and cultural values, such as racial and religious harmony, are particularly needed.

For the Early Detection thematic tools, we will prioritise and select methods and systems that can present a significant advance over existing solutions, such as a) displaying high detection accuracy for real-world data, with applicability and robustness for the Singapore context, b) presenting features that are explainable to facilitate user adoption, and c) advancing the understanding of the underlying phenomenon.

Research Theme 3: Mitigation & Empowerment

Mitigation & Empowerment is a promising area where technology researchers, social, behavioural and communication scientists and practitioners can join forces to create theoretically and empirically sound tools for alerting, debunking, or even pre-bunking online misinformation and disinformation.

Tech 8

Tools to Improve Debunking Effectiveness

Technology that can algorithmically assist content generators and communication professionals in evaluating and recommending more effective mis/disinformation debunking messaging, such as the clarity, creditability, vividness, emotionality, empathy, and contextual dimensions of a crafted message intending to mitigate the negative impact of a piece of mis/disinformation which has already emerged or spread.

Tech 10

Media Literacy Enhancement Tools

Development and testing of “gamified education” tools such as mobile apps that seek to increase media literacy and critical thinking through gamification, quizzes, knowledge modules, etc. The tools and studies shall advance the understanding of our population's vulnerabilities from existing studies and examine with efficacy studies the extent to which such media literacy enhancement tools can increase “psychological defence" by educating citizens effectively and inoculating citizens against the effects of online falsehood and harms.

Tech 9

Online Trust & Safety Policy Testing System

Technology that can assist policymakers, analysts, content moderation leads and info operation teams to evaluate the impact of policy decision-making based on personas that represent the communities. This technology needs to factor in Singapore's local context, such as local demographics and psychographics and Singapore's values, to enable policy testing that is not feasible to generate and evaluate using conventional approaches.

For the Mitigation & Empowerment thematic tools, we will prioritise projects that present highly novel and effective methods based on sound psychological knowledge and smart design for different demographic groups [e.g., 25-32].

Please download the OTS Open Grant Call Document Package and refer to Annex A in OTS Open Grant Call_Info Sheet.pdf for more details on the Research Themes and the respective Tech Topics under the Research Themes.

Proposals can either tackle one or more of the tech topics of interest. Please clearly indicate the Tech Topic(s) that the proposal will address on the Proposal Cover Page.

With solid justification and rationale, proposals are allowed to propose tech topics not identified in Annex A. Such proposals must clearly explain the value of the proposed research approach in meeting the broader aims of the OTS Open Grant Call, provide comparative evaluations against state-of-the-art approaches (including, if necessary, the ten identified tech topics), and eventually be demonstratable and robust when evaluated using real-world local context data.

Funding Support

The OTS Open Grant Call offers two categories of funding support:

Category 1:
Funding up to S$500K (inclusive of 30% indirect cost) for a proposal focusing on addressing one Tech Topic
Category 2:
Funding up to S$3M (inclusive of 30% indirect cost) for a proposal focusing on addressing two or more Tech Topics

This Open Grant Call offers funding support for up to 2.5 years.

Applicants are strongly encouraged to collaborate with industry partners (e.g., mainstream media companies, digital platform companies, independent fact-checkers, etc.) and government public sector agencies to develop innovative solutions to address the grant call objectives and demonstrate strong potential (or, more preferably, actual technology deployment and use cases) for real-world applications within and beyond Singapore.

Eligibility Criteria

At the point of application, the Principal Investigators (PIs) and Co-Principal Investigators (Co-PIs) must hold a full-time appointment and be salaried in a local publicly funded institution. The full-time appointment is defined as at least 9 months of service a year based in Singapore or 75% appointment.

The PI and Co-PIs must be based in Singapore. Collaboration with foreign organisations and experts in the capacity of Collaborator is allowed. Research work should be done in Singapore and should not be carried out overseas unless expressly approved by the grantor.

Where applicable, we encourage the integration of relevant real-world use cases or social and behavioural research to complement the technology-centric R&D work under this grant call to ensure the practicality, user-centricity and acceptability of the solutions proposed.

Proposals that other government agencies already fund will not be considered under this grant call. PIs shall declare their other funding sources and participation in other funding initiatives during application. Proposals with similar scope, currently funded or under evaluation by other funding initiatives, will not be considered.

In addition to the above criteria, applicants should not have any outstanding reports from previous A*STAR grants, NMRC grants and other national competitive grants.

Evaluation Criteria

Selection of successful proposals will be based on, but not limited to, the following evaluation criteria:

The novelty, intellectual and innovative merit of the proposal
The competitive advantage of the proposed solution(s) in meeting the aims of the OTS Open Grant Call
Potential technological impact based on the proposed deliverables, in particular, the likelihood of technology being adopted and used by the industry through licensing
The feasibility of implementing and deploying the proposed methodology
The track record of the PI and the team

The proposals will be reviewed by international scientific reviewers and the OTS Open Grant Call Proposal Review Panel based on the evaluation criteria. PIs of shortlisted proposals may be invited to give presentations and answer questions from the Proposal Review Panel.

All decisions are final, and no appeals will be entertained.

Submission Instructions

PIs should use the proposal and budget templates provided and follow the instructions stated in the templates.

Only the Principal Investigator from each research team can submit the application. The applicant will be the primary contact for the research team.

The application must be endorsed by the relevant institutional authority/director of research (or equivalent) of PIs and Co-PIs whenever applicable. Incomplete applications will not be accepted.

The application must be submitted through the iGrants system (igrants-app.a-star.edu.sg) before Wednesday, 10th Jan 2024, 1700 hrs (Singapore Time). Late applications will not be accepted.

Inquires can be addressed to CATOS-OTS@hq.a-star.edu.sg.

More Information

For the full details of the OTS Open Grant Call, please download document package (in zip format).

Download Grant Call Document Package