Chimera combines kernelized attention approximations with symbolic fusion mechanisms to enable high-fidelity neuro-symbolic inference inside commodity programmable switches.
Stable, fast and accurate: Kernelized attention with relative positional encoding.Advances in Neural Information Processing Systems, 34:22795–22807, 2021
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.NI 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Chimera: Neuro-Symbolic Attention Primitives for Trustworthy Dataplane Intelligence
Chimera combines kernelized attention approximations with symbolic fusion mechanisms to enable high-fidelity neuro-symbolic inference inside commodity programmable switches.