Fangyu Liu

Title: Fangyu Liu: Innovator in Language Processing Technologies

Introduction

Fangyu Liu is a prominent inventor based in Cambridge, GB. He has made significant contributions to the field of language processing technologies, holding 2 patents that showcase his innovative approaches to complex problems.

Latest Patents

One of his latest patents is titled "Zero-shot domain transfer with a text-to-text model." This invention provides example solutions for zero-shot domain transfer by training a text-to-text model for a target domain using unlabeled in-domain text training data. Concurrently, the model is trained using labeled general-domain task training data. The in-domain training involves masked language modeling (MLM) training, while the task training includes both natural language generation (NLG) and natural language understanding (NLU) training. The NLG training encompasses natural language inference (NLI) training, and the NLU training includes summarization training. This trained model acquires domain-specific task competency, enabling it to perform language tasks within various target domains, including radiology, biomedical, and other medical, legal, and scientific fields. This approach effectively utilizes large volumes of general-domain task training data and abundant unlabeled in-domain text, addressing the challenges posed by the unavailability or high cost of labeled in-domain training data in specialized domains.

Another notable patent is "One-shot visual language reasoning over graphical depictions of data." This invention offers a one-shot solution to visual language reasoning by decomposing the challenge into two steps: translating a graphical depiction of data, such as a plot or chart, into text, followed by reasoning over the translated text. The systems described can include a machine-learned visual-to-language conversion model that translates graphical depictions of datasets into descriptive text. The output from this model can then be used to prompt a language model, such as a pretrained large language model (LLM), leveraging the few-shot reasoning capabilities of the language model.

Career Highlights

Fangyu Liu has worked with leading technology companies, including Google Inc. and Microsoft Technology Licensing, LLC. His experience in these organizations has allowed him to refine his skills and contribute to groundbreaking projects in language processing.

Collaborations

Throughout his career, Fangyu has collaborated with notable professionals, including Kenton Chiu Tsun Lee and Yasemin Altun. These collaborations have further enriched his work and expanded the impact of his