Diksha Garg

Title: **Innovations by Diksha Garg: Transforming Recommendation Systems**

Introduction

Diksha Garg is a prominent inventor based in Noida, India, recognized for his innovative contributions to the field of recommendation systems. With an impressive portfolio of two patents, he is making significant strides in addressing challenges in user interaction and recommendation biases, particularly through advanced machine learning techniques.

Latest Patents

Diksha's latest inventions include two notable patents that focus on improving session-based recommendation systems. The first patent, titled "System and method for training recommendation policies," introduces a novel approach for recommending the next item based on historical user interactions. Unlike existing methods that depend on costly real-time user interactions or biased behavioral models, his system employs Distributional Reinforcement Learning (DRL) to derive recommendation policies from offline data. This innovative approach eliminates the need for real-user interactions, enabling efficient model training and implementation.

The second patent, "System, method, and non-transitory machine-readable information storage medium for handling popularity bias in item recommendations," addresses the problem of popularity bias in recommendations. It outlines a method that initializes an item embedding look-up matrix from a sequence of item clicks and applies L2 norm learning to normalize item embeddings. Utilizing a neural network, the patent develops session embeddings and computes relevance scores to generate a prioritized list of recommendations. This patent represents a significant step forward in creating unbiased and responsive recommendation systems.

Career Highlights

Diksha Garg currently works at Tata Consultancy Services Limited, where he leverages his expertise to drive innovation in technology solutions. Throughout his career, he has played a crucial role in developing cutting-edge systems that enhance user experience and optimize recommendation processes.

Collaborations

In his endeavors, Diksha collaborates with skilled professionals, including his coworkers Pankaj Malhotra and Priyanka Gupta. Together, they contribute to fostering an innovative environment that encourages the development of forward-thinking technological solutions.

Conclusion

Diksha Garg stands out as a forward-thinking inventor in Noida, India, with a passion for transforming recommendation systems. His patents not only enhance the efficiency and accuracy of recommendations but also mitigate biases inherent in current models. With continued innovation and collaboration, Diksha is poised to make a lasting impact in the realm of technological advancements and user experience enhancements.