Anthony Jayakumar

Title: Anthony Jayakumar and Neural-Network Innovations

Introduction

Anthony Jayakumar is an inventor associated with Somerset, New Jersey. His work includes inventions involving self-learning neural networks and signal processing. He is credited with 2 patents and worked at Bell Communications Research, Inc.

Latest Patents

Jayakumar’s patent, “Adaptive Equalizer Using Self-Learning Neural Network,” describes a channel equalizer that learns a channel’s response during training and then compensates for signal distortions. The design uses a neural network that may include input, hidden, and output layers connected by bidirectional, symmetric weighted synapses. Direct communication between the input and output layers supports faster channel acquisition, and the approach can be extended to modulation schemes such as QAM and PSK.

His patent, “Neuron for Use in Self-Learning Neural Network,” describes a variable-gain, bidirectional current-transimpedance neuron with a controllable noise input. The neuron sums synaptic currents, normalizes the result, converts it to voltage, and amplifies the output. Gain control and noise inputs support settling through Mean Field Approximation and simulated annealing.

Career Highlights

At Bell Communications Research, Inc., Jayakumar’s work addressed neural-network components and their application to communications systems. His inventions connect circuit-level neuron designs with adaptive signal equalization.

Collaborations

Jayakumar’s listed coworkers include Joshua Alspector and Timothy X. Brown.

Conclusion

Anthony Jayakumar’s patents reflect contributions to self-learning neural networks, adaptive equalization, and neuron circuitry for communications applications.