This inventor holds 3 USPTO granted patents and 3 published patent applications. Top assignee: Oracle International Corporation. Active years: 2026.
Company Filing History:
Years Active: 2026
Title: Innovations of Tomas Feith
Introduction
Tomas Feith is a notable inventor based in Zurich, Switzerland. He has made significant contributions to the field of natural language processing and machine learning. With a total of 3 patents, his work focuses on enhancing the efficiency and accuracy of language models.
Latest Patents
Tomas Feith's latest patents include a "Graph path prediction and masked language modelling joint training algorithm for language models." This invention provides a method for generating a histogram that accurately represents a graph of lexical text. It also involves a token sequence encoder that can be both trainable and untrained. During the pretraining phase, the encoder infers an encoded sequence that may not accurately represent the lexical text, while also being space-efficient. The accuracy of this sequence encoder is improved by adjusting it based on a numeric difference between the encoded sequence and the correct histogram representation.
Another significant patent is the "AST branch prediction and next token prediction joint pre-training task for code generative models." In this invention, three machine learning models are generated during pretraining: a token sequence encoder, a token predictor, and a path predictor. The process involves generating a sequence of lexical tokens and a corresponding graph that represents the lexical text. The path predictor and token predictor utilize the encoded sequence for inferencing, and their training losses are combined to enhance the accuracy of the models through backpropagation.
Career Highlights
Tomas Feith is currently employed at Oracle International Corporation, where he continues to innovate in the field of technology. His work has been instrumental in advancing the capabilities of machine learning and natural language processing.
Collaborations
Some of his notable coworkers include Matteo Casserini and Felix Schmidt, who contribute to the collaborative environment that fosters innovation at Oracle.
Conclusion
Tomas Feith's contributions to the field of natural language processing and machine learning are noteworthy. His innovative patents reflect his commitment to advancing technology and improving the efficiency of language models.