A procedural self-critique loop can make LLM planners markedly more reliable—but only when reflection is converted into explicit rule checking, state tracking, and conservative approval.
ERIQ and GenieReasoner reveal why understanding the right action and physically executing it are separate engineering problems that robotics teams must diagnose separately.
A practical reading of why LLM memorization becomes hard to remove once training entangles recall with general capability.
A controlled study of LLM-generated neural networks shows why moderate prompt context can improve architecture synthesis—and why more examples eventually break the pipeline.
A systems-level reading of how AI scale can amplify ecological, social, linguistic, and institutional risks—and what organizations can do about it.
A four-year experiment in hands-on NLP ethics shows why responsibility is learned through difficult choices, public explanation, and repeated practice—not compliance slides.
LeanCat reveals why verified AI reasoning still fails when agents must navigate large libraries, preserve abstraction, and construct missing conceptual bridges.
MI-ZO shows how a lightweight inference-time controller can improve 2D-trained vision-language models on 3D scenes by learning which views contain useful, non-redundant evidence.
HiGR shows that generative recommendation becomes practical only when item representation, slate planning, and preference alignment are designed as one coordinated system.
Encyclo-K replaces fixed benchmark questions with dynamically composed knowledge statements, creating a reusable evaluation engine that exposes the gap between knowing facts and reliably combining them.