Skip to content

Author

Sukanta Ganguly

1 paper indexed here

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

Fractional Decay KV-Cache: Ownership-Aware Memory Management for Improved Inference Relevancy in Dialog Systems

Fractional Decay KV-Cache is proposed, a novel algorithm that maintains a dual-channel scoring mechanism for each cached KV pair: a cumulative attention channel that tracks aggregate importance (akin to H2O), and a recency-weighted relevance channel governed by temporal decay and reinforcement-inspired updates.

Sukanta Ganguly · 0 citations