Proposals are being accepted for the Spring 2025 round of Amazon Research Awards, closing 30 April 2025.
Funding of up to USD$80,000 – USD$100,000 and/or AWS Promotional Credits is available, depending on the Theme.
AI for Information Security – advancing possible solutions for some of the most challenging problems in information security:
- Agentic AI for security
- Generative AI and foundation models for information security
- Securing generative AI and foundation models
- Threat, intrusion and anomaly detection for cloud security
- Vulnerability detection and remediation using AI
- Incident response using AI
- Reinforcement learning for information security
- Learning with limited/noisy labels and weakly supervised learning
- Graph modeling and anomaly detection on graphs
- ML for malware analysis and detection, with a focus on cloud environments or devices
Amazon Ads – advancing customer protections in the era of artificial intelligence in digital advertising:
- Fraud, Abuse, and financial scams
- Behavioral foundational models to distinguish human and bot behavior
- Multi-modal classification of online content and websites
- Device and website identification
- Detection of plagiarized content
- Cyber threat activity, malicious and adversarial misuse of AI and LLMs
- Decentralized identifiers, verifiable credentials, data lineage and transparency
- Large language models for labeling, annotation and auditing of model performance
- Detection and mitigation of hallucination in LLMs for content analysis
- Protection of vulnerable individuals, specially child and teen safety
Build on Tranium – building the future of AI with AWS Tranium:
- Novel kernels and compiler extensions for Trainium
- Novel algorithms for large language models
- Systems improvements for distributed training and hosting
- Streamline development
Think Big – advancing the frontiers of science through transformative ideas; research that might be considered speculative or contains out-of-the-box thinking. Ideas that
- Challenge fundamental assumptions in your field
- Propose novel frameworks for solving long-standing problems
- Could enable order-of-magnitude improvements over current approaches
- Have the potential to create new technical capabilities
- Could open unexplored frontiers in science and technology
