Showing cs.LGShow all
3 papers · 1 filter
cs.LG2026
LAWS: Learning from Actual Workloads Symbolically -- A Self-Certifying Parametrized Cache Architecture for Neural Inference, Robotics, and Edge Deployment
Gregory Magarshak
We introduce LAWS (Learning from Actual Workloads Symbolically), a self-certifying inference caching architecture that builds a growing library of certified expert functions from d…
cs.LG2026
Sequential KV Cache Compression via Probabilistic Language Tries: Beyond the Per-Vector Shannon Limit
Gregory Magarshak
Recent work on KV cache quantization, culminating in TurboQuant, has approached the Shannon entropy limit for per-vector compression of transformer key-value caches. We observe tha…
cs.LG2026
Probabilistic Language Tries: A Unified Framework for Compression, Decision Policies, and Execution Reuse
Gregory Magarshak
We introduce probabilistic language tries (PLTs), a unified representation that makes explicit the prefix structure implicitly defined by any generative model over sequences. By as…