MambaCSP: Hybrid-Attention State Space Models for Hardware-Efficient Channel State Prediction

arXiv 2604.21957•3d4150dd2a6a6b87237ad00b7c0a1a15bfd740dac8d96cf146324504c85288dd
CSI predictionGamma-distributed constellationISACMIMOMambaCSPPCRBattentionbeamformingchannel-state-informationconstellation designdiffusiongenerative modelsgrouped patternintegrated sensing and communicationslow-latencynull-space flow matchingrange estimationsecrecy outage probability','wireless secrecy','reconfigurable-@semantic HARQsemantic communicationsemantic error correctionshort block codesstate-space modelstransformerswireless

Paper metadata

arXiv ID
2604.21957
Version
Not specified by this published record
Category
Mathematics — Information Theory (math.IT)

The PDF link points to arxiv.org. Baitaphish does not expose a private stored PDF.

Evidence and limitations

Source ID
arxiv_math_it
Record identifier
3d4150dd2a6a6b87237ad00b7c0a1a15bfd740dac8d96cf146324504c85288dd
Enrichment time
2026-04-27T08:51:49Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.