Tata Consultancy Services Limited

Gitesh Kulkarni

Arpan Pal

Chirabrata Bhaumik

Tapas Chakravarty

Arijit Mukherjee

Arijit Ukil

Arijit Sinharay

Swarnava Dey

Amit Swain

Ishan Sahu

Abhijeet Gorey

Jayeeta Mondal

Aakash Tyagi

Small and compact Deep Learning models are required for embedded AI in several domains. In many industrial use-cases, there are requirements to transform already trained models to ensemble embedded systems or re-train those for a given deployment scenario, with limited data for transfer learning. Moreover, the hardware platforms used in embedded application include FPGAs, AI hardware accelerators, System-on-Chips and on-premises computing elements (Fog/Network Edge). These are interconnected through heterogenous bus/network with different capacities. Method of the present disclosure finds how to automatically partition a given DNN into ensemble devices, considering the effect of accuracy—latency power—tradeoff, due to intermediate compression and effect of quantization due to conversion to AI accelerator SDKs. Method of the present disclosure is an iterative approach to obtain a set of partitions by repeatedly refining the partitions and generating a cascaded model for inference and training on ensemble hardware.

Combining compression, partitioning and quantization of DL models for fitment in hardware processors

Fluids are normally transported from one place to another through pipelines. It is essential to monitor the pipeline to avoid leakage or theft. It is expensive and not feasible to install cameras and sensors along the whole length of the pipeline. A system and method for inspecting and detecting fluid leakage in a pipeline has been provided. The system is using vibration sensors along with pressure sensors to detect the leakage or theft along with the exact location of the leakage or theft. The pressure sensors are mounted on the pipeline so that the fluid touches the diaphragm of the pressure sensors to sense the wave generated due to leakage. The vibration sensors are mounted on top of the pipeline surface and on the nearby ground to eliminate general noise conditions. Moreover, two pressure sensors are also installed at opposite sides to pinpoint the leakage location.

Method and system for inspecting and detecting fluid in a pipeline

Growing community of inventors

Bengaluru, India

Gitesh Kulkarni