Soumendu Kumar Ghosh

17 papers A 3B 1Misc 1Journal 9Unranked 3
YearRankTypeTitle / Venue / Authors
2026 J jnl
Softprox: A Post-Finetuning Methodology to Mitigate Softmax Bottlenecks in Transformer Workloads.
IEEE Trans. Circuits Syst. Artif. Intell.
Aradhana Mohan Parvathy, Sourjya Roy, Soumendu Kumar Ghosh, Arnab Raha, Deepak Mathaikutty, Anand Raghunathan
2025 A conf
Demo Abstract: ECO: Low Power Context-Aware Multimodal AI on NPUs.
ISLPED
Arghadip Das, Yatharth Agarwal, Soumendu Kumar Ghosh, Arnab Raha, Vijay Raghunathan
2025 B conf
GraNNite: Enabling High-Performance Execution of Graph Neural Networks on Resource-Constrained Neural Processing Units.
IJCNN
Arghadip Das, Shamik Kundu, Arnab Raha, Soumendu Kumar Ghosh, Deepak Mathaikutty, Vijay Raghunathan
2025 J jnl
GraNNite: Enabling High-Performance Execution of Graph Neural Networks on Resource-Constrained Neural Processing Units.
CoRR
Arghadip Das, Shamik Kundu, Arnab Raha, Soumendu Kumar Ghosh, Deepak Mathaikutty, Vijay Raghunathan
2025 — conf
LLM-NPU: Towards Efficient Foundation Model Inference on Low-Power Neural Processing Units.
COINS
Arnab Raha, Souvik Kundu, Sharath Nittur Sridhar, Shamik Kundu, Soumendu Kumar Ghosh, Alessandro Palla, Arghadip Das, Darren Crews, Deepak A. Mathaikutty
2025 J jnl
Towards Energy-Accuracy Scalable Multimodal Cognitive Systems.
IEEE Embed. Syst. Lett.
Soumendu Kumar Ghosh, Arghadip Das, Arnab Raha, Vijay Raghunathan
2025 J jnl
XAMBA: Enabling Efficient State Space Models on Resource-Constrained Neural Processing Units.
CoRR
Arghadip Das, Arnab Raha, Shamik Kundu, Soumendu Kumar Ghosh, Deepak Mathaikutty, Vijay Raghunathan
2024 J jnl
FlexNN: A Dataflow-aware Flexible Deep Learning Accelerator for Energy-Efficient Edge Devices.
CoRR
Arnab Raha, Deepak A. Mathaikutty, Soumendu Kumar Ghosh, Shamik Kundu
2024 Misc conf
HARVEST: Towards Efficient Sparse DNN Accelerators using Programmable Thresholds.
VLSID
Soumendu Kumar Ghosh, Shamik Kundu, Arnab Raha, Deepak A. Mathaikutty, Vijay Raghunathan
2024 J jnl
PArtNNer: Platform-Agnostic Adaptive Edge-Cloud DNN Partitioning for Minimizing End-to-End Latency.
ACM Trans. Embed. Comput. Syst.
Soumendu Kumar Ghosh, Arnab Raha, Vijay Raghunathan, Anand Raghunathan
2024 A conf
SwiSS: Switchable Single-Sided Sparsity-based DNN Accelerators.
ISLPED
Shamik Kundu, Soumendu Kumar Ghosh, Arnab Raha, Deepak A. Mathaikutty
2024 J jnl
Toward Energy-Efficient Collaborative Inference Using Multisystem Approximations.
IEEE Internet Things J.
Arghadip Das, Soumendu Kumar Ghosh, Arnab Raha, Vijay Raghunathan
2023 J jnl
Energy-Efficient Approximate Edge Inference Systems.
ACM Trans. Embed. Comput. Syst.
Soumendu Kumar Ghosh, Arnab Raha, Vijay Raghunathan
2020 A conf
Approximate inference systems (AxIS): end-to-end approximations for energy-efficient inference at the edge.
ISLPED
Soumendu Kumar Ghosh, Arnab Raha, Vijay Raghunathan
2017 — conf
Railway bridge health monitoring system using smart wireless sensor network: demo.
WISEC
Sathvik Dev Velagandula, Nirjhar Dhang, Raja Datta, Soumendu Kumar Ghosh, Maroju Suman
2016 J jnl
CEEDS: A Cost Effective Event Detection System for Energy Efficient Railway Bridge Monitoring with Wireless Sensor Network.
EAI Endorsed Trans. Ubiquitous Environ.
Soumendu Kumar Ghosh, Maroju Suman, Raja Datta, Prabir Kumar Biswas
2014 — conf
Power efficient event detection scheme in wireless sensor networks for railway bridge health monitoring system.
IEEE ANTS
Soumendu Kumar Ghosh, Maroju Suman, Raja Datta, Prabir Kumar Biswas