3 papers
cs.LG2026
Compiler-First State Space Duality and Portable Autoregressive Caching for Inference
Cosmo Santoni, Anmol Thapar
High-throughput Mamba-2 inference is usually tied to fused CUDA and Triton kernels, limiting portability across accelerator backends. We show that the state space duality (SSD) rec…
cs.LG2026
Tokenised Flow Matching for Hierarchical Simulation Based Inference
Giovanni Charles, Cosmo Santoni, Seth Flaxman +1
The cost of simulator evaluations is a key practical bottleneck for Simulation Based Inference (SBI). In hierarchical settings with shared global parameters and exchangeable site-l…
cs.SE2026
Contextual Memory Virtualisation: DAG-Based State Management and Structurally Lossless Trimming for LLM Agents
Cosmo Santoni
As large language models engage in extended reasoning tasks, they accumulate significant state -- architectural mappings, trade-off decisions, codebase conventions -- within the co…