This inventor holds 1 USPTO granted patent. Top assignee: Adobe, Inc.. Active years: 2026.
Company Filing History:
Years Active: 2026
Title: The Innovations of Abel Salinas
Introduction
Abel Salinas is an inventive mind based in West Covina, CA (US). He is known for his contributions to the field of language models, particularly in the area of training methods. His innovative approach focuses on enhancing the effectiveness of language models through advanced data generation techniques.
Latest Patent Applications
Abel Salinas has filed a patent application titled "IMPROVED TRAINING METHODS FOR LANGUAGE MODELS USING DATA GENERATION AND REINFORCEMENT LEARNING." This method aims to generate helpful training data for language models, specifically for tasks such as punctuation restoration in real-world Automatic Speech Recognition (ASR) texts. The approach utilizes a reinforcement learning method in conjunction with a generative AI model to create additional training data. This allows the generative AI model to learn from actual ASR text, thereby producing more effective training examples based on gradient feedback from the language model.
Conclusion
Abel Salinas is an emerging inventor whose work in language model training methods showcases the potential of generative AI in improving real-world applications. His innovative patent application reflects a commitment to advancing technology in this field.