Sifei Li

Title: Sifei Li and Access-Controlled Machine Learning Fine-Tuning.

Introduction

Sifei Li is an inventor based in Bothell, Washington, and is associated with Amazon Technologies, Inc. His patent work addresses access controls in the fine-tuning of machine learning models.

Latest Patents

Li is named on 1 patent: “Enforcing Access Restrictions for Fine-Tuning Machine Learning Models.” The invention describes receiving a request to fine-tune a model and accounting for both the model provider’s access restrictions and the tuning-data consumer’s access restrictions. It then generates a tuned set of weights that can be combined with the trained model’s weights for inference.

Career Highlights

The patent reflects work at the intersection of machine learning and data governance. Its approach aims to support model customization while respecting restrictions imposed by both providers and consumers.

Collaborations

Li’s listed coworkers include Arun Babu Nagarajan and Paras Mehra. The patent’s focus on coordinating access restrictions highlights the importance of collaboration among contributors working on machine learning systems.

Conclusion

Sifei Li’s patent presents an approach to fine-tuning machine learning models that accounts for access restrictions on both models and tuning data.