Most natural signals are sparse in some basis; hence compressive sensing is a natural approach for information acquisition.

Compressive projections:

  • Can be realized using inner products with binary coefficients, which is attractive for hardware implementation.

  • Are expected to be resilient to a broad class of impairments.

Hence they are a promising approach for low-power front ends for downstream learning and inference. 

Compressive projections are a promising approach for general purpose front ends: 

  • They can be realized using inner products with binary coefficients, which is attractive for hardware implementation.

  • They are expected to be resilient to a broad class of impairments.

Compressive projections are a promising approach for general purpose front ends: 

  • They can be realized using inner products with binary coefficients, which is attractive for hardware implementation.

  • They are expected to be resilient to a broad class of impairments.

Compressive projections are a promising approach for general purpose front ends: 

  • They can be realized using inner products with binary coefficients, which is attractive for hardware implementation.

  • They are expected to be resilient to a broad class of impairments.

Compressive projections are a promising approach for general purpose front ends: 

  • They can be realized using inner products with binary coefficients, which is attractive for hardware implementation.

  • They are expected to be resilient to a broad class of impairments.

We investigate the impact of real-world nonlinearities on compressive front ends, and study design changes needed to target deep learning applications. 

Students

Soorya Gopalakrishnan

Faculty

Upamanyu Madhow, Naveen Verma (Princeton University)

Publications

S. Gopalakrishnan, T. Moy, U. Madhow, N. Verma, "Compressive Information Acquisition with Hardware Impairments and Constraints: A Case Study", in 2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), New Orleans, LA, March 2017, pp. 6075-6079.