Abhishek Kumar

Title: Abhishek Kumar: Innovator in Classifier Assistance Technology

Introduction

Abhishek Kumar is a notable inventor based in Redmond, WA, who has made significant contributions to the field of machine learning and data classification. He is currently associated with Microsoft Technology Licensing, LLC, where he focuses on innovative solutions that enhance data processing and classification techniques.

Latest Patents

Abhishek Kumar holds a patent titled "Classifier assistance using domain-trained embedding." This invention addresses the challenge of training classifiers with limited manually annotated datasets. The approach involves using a small subset of labeled data as seeds, from which data augmentations can generate additional seed verbatim sets for known topics. The method assigns temporary pseudo labels to the remaining data based on their proximity in vector space to the labeled seeds. The training process includes classification epochs that update embeddings, assuming pseudo labels are ground-truth labels, and labeling epochs that refine the vectors corresponding to the verbatims. As the training progresses, the embeddings converge, leading to improved classification accuracy.

Career Highlights

Throughout his career, Abhishek has demonstrated a strong commitment to advancing technology in the field of data science. His work at Microsoft Technology Licensing, LLC, showcases his ability to innovate and develop practical solutions that address real-world challenges in data classification.

Collaborations

Abhishek Kumar has collaborated with talented colleagues such as Soyoung Peraud and Alexandre Rochette, contributing to a dynamic work environment that fosters innovation and creativity.

Conclusion

Abhishek Kumar's contributions to classifier assistance technology exemplify the impact of innovative thinking in the realm of machine learning. His patent reflects a forward-thinking approach to data classification, paving the way for future advancements in the field.