AI paper roundups
Dated selections of recent papers with reading notes on their methods and evidence.
Reading order
- Five research directions, from multimodal memory to small-model agents
- Five papers on agent world models, scaling laws, multi-image reasoning, 3D reconstruction and watermarking
- Five papers on pixel-level multimodality, 3D video constraints, clinical evaluation, RoPE and classification theory
- Five AI papers for a closer reading on May 9, 2026
- Five papers on test-time scaling, driving data, 3D representations, image generation and KGQA
- Five AI papers to read on May 12, 2026
- Five papers on embodied-agent verifiers, benchmark integrity, VLM failures, GUI grounding and implicit preferences