Fifth Call For AI Research Proposal

In the AI Research Grant Call 2022, AI Singapore invites proposals that advance the state of the art in fundamental AI research for enabling and realising the following visions of the future:

Vision 1: Discriminating Information

• This vision centres on advancing the state of the art in fundamental AI research to improve our ability to discriminate information from mis/disinformation, to understand the negative impacts of false information and how to prevent / mitigate such impacts.
• Enabling fundamental AI research topics include:
  – Information authentication, bias, benchmarks and evaluation
  – Interpretation and generation
  – Assessment of impact and user profiling

Vision 2: Towards Human Omnitasking

• This vision centres on advancing the state of the art in fundamental AI research to enable humans to perform a range of tasks more efficiently, and to facilitate effective multitasking.
• Enabling fundamental AI research topics include:
  – Multimodal understanding and summarisation / Conversational NLP
  – Cognitive state tracking / Common sense reasoning
  – Behaviour understanding / Personalisation
  – Trustworthy and explainable AI / Human-AI collaboration

Vision 3: Removing Unwanted Digital Footprints

• This vision centres on advancing the state of the art in fundamental AI research to improve our ability to control the footprints that we leave behind us on AI systems that use our data.
• Enabling fundamental AI research topics include:
  – Definition of unwanted footprints
  – Footprint removal algorithms
  – Footprint removal auditing
  – Evaluation of organisational trustworthiness

Vision 4: AI for Science

• This vision centres on advancing the state of the art in fundamental AI research to aid and accelerate the scientific discovery process.
• Enabling fundamental AI research topics include:
  – AI for natural law extraction and understanding
  – Foundational AI models for fast, accurate, large-scale scientific discovery 
  – AI for inverse design in science
  – AI that maximises a priori knowledge
  – AI for data-sparse science
  – Learning representations for scientific creativity
  – AI for quantum science