Subject areas
AEX CONTINUUM is organised into subject areas that cut across all five collections. A record can belong to more than one.
Language Models & Agents
Architectures, training and prompting of language models; tool use, planning and multi-step agents.
Efficient & On-Device ML
Compression, quantisation, distillation, parameter-efficient adaptation and fast inference under resource limits.
Evaluation & Benchmarks
Benchmark design, metrics, measurement validity, statistical practice and human evaluation.
Retrieval & Knowledge
Retrieval-augmented generation, indexing, knowledge grounding and uncertainty about retrieved evidence.
Safety, Security & Governance
Robustness, adversarial and prompt-injection attacks, oversight, accountability and policy for AI systems.
Research Practice & Reproducibility
Open science, research software, reporting standards and the sociology of machine-learning research.
Data & Research Infrastructure
Datasets, documentation, repositories, persistent identifiers and the tooling that keeps research outputs usable.