6 papers
BLINK: Batch Normalization-based Integrity Checkpoints for In-Situ Detection and Mitigation of Diverse Weight Corruptions in DNN Accelerators
Marzia Khan, Akul Malhotra, Sumeet Kumar Gupta
In safety-critical deployments, AI hardware must remain reliable against a broad spectrum of threats such as aging, soft errors, hard faults, and adversarial attacks (e.g. progress…
Weight Transformations in Bit-Sliced Crossbar Arrays for Fault Tolerant Computing-in-Memory: Design Techniques and Evaluation Framework
Akul Malhotra, Sumeet Kumar Gupta
The deployment of deep neural networks (DNNs) on compute-in-memory (CiM) accelerators offers significant energy savings and speed-up by reducing data movement during inference. How…
Oxygen Vacancy-Induced Monoclinic Dead Layers in Ferroelectric With Metal Electrodes
Tanmoy Kumar Paul, Atanu Kumar Saha, Sumeet Kumar Gupta
In this work, we analyze dead layer comprising non-polar monoclinic (m) phase in (HZO)-based ferroelectric (FE) material using first principles analysis. We show…
BinSparX: Sparsified Binary Neural Networks for Reduced Hardware Non-Idealities in Xbar Arrays
Akul Malhotra, Sumeet Kumar Gupta
Compute-in-memory (CiM)-based binary neural network (CiM-BNN) accelerators marry the benefits of CiM and ultra-low precision quantization, making them highly suitable for edge comp…
SiTe CiM: Signed Ternary Computing-in-Memory for Ultra-Low Precision Deep Neural Networks
Niharika Thakuria, Akul Malhotra, Sandeep K. Thirumala +3
Ternary Deep Neural Networks (DNN) have shown a large potential for highly energy-constrained systems by virtue of their low power operation (due to ultra-low precision) with only…
Memory Faults in Activation-sparse Quantized Deep Neural Networks: Analysis and Mitigation using Sharpness-aware Training
Akul Malhotra, Sumeet Kumar Gupta
Improving the hardware efficiency of deep neural network (DNN) accelerators with techniques such as quantization and sparsity enhancement have shown an immense promise. However, th…