3 papers
cs.AI2026
MANAR: Memory-augmented Attention with Navigational Abstract Conceptual Representation
Zuher Jahshan, Ben Ben Ishay, Leonid Yavits
MANAR (Memory-augmented Attention with Navigational Abstract Conceptual Representation), contextualization layer generalizes standard multi-head attention (MHA) by instantiating th…
q-bio.GN2026
Processing-in-memory for genomics workloads
William Andrew Simon, Leonid Yavits, Konstantina Koliogeorgi +12
Low-cost, high-throughput DNA and RNA sequencing (HTS) data is the backbone of the life sciences. Genome sequencing is now becoming a part of Predictive, Preventive, Personalized,…
cs.AR2026
PiC-BNN: A 128-kbit 65 nm Processing-in-CAM-Based End-to-End Binary Neural Network Accelerator
Yuval Harary, Almog Sharoni, Esteban Garzón +3
Binary Neural Networks (BNNs), where weights and activations are constrained to binary values (+1, -1), are a highly efficient alternative to traditional neural networks. Unfortuna…