2 papers
cs.CV2026
Compositional Semantics for Open Vocabulary Spatio-semantic Representations
Robin Karlsson, Francisco Lepe-Salazar, Kazuya Takeda
Vision-language models (VLMs) transform environment percepts into vision-language semantics interpretable by LLMs. However, completing complex tasks often requires reasoning about…
cs.RO2026
CSR: Infinite-Horizon Real-Time Policies with Massive Cached State Representations
Robin Karlsson, Go Suzui
Deploying massive large language models (LLMs) as continuous cognitive engines for robotics is bottlenecked by the time-to-first-token (TTFT) latency required to process extensive…