Dollars for Profs
Dig Into University Researchers' Outside Income and Conflicts of Interest
Published Dec. 6, 2019
This database was last updated in December 2019 and should only be used as a historical snapshot. There may be new or amended records not reflected here.
Conflict of Interest
Institutions must file significant disclosures to the National Institutes of Health if they determine financial relationships could affect the design, conduct or reporting of the NIH-funded research. The NIH provided us with their entire financial conflict of interest database, with filings from 2012 through 2019.
Should you be removed from our database? Contact us at [email protected]. Read more below.
Hua Xu
University of Texas Hlth Sci Ctr Houston, Department: None
Should you be removed from our database? Contact us at [email protected]. Read more below.
Melax Technologies, Inc.
Equity Interest - Non-publicly traded entity ( e.g., stock, stock option, or other ownership interest)
Dr. Xu is the founder, shareholder, and consultant to Melax Technologies, Inc., a private, for-profit company which will provide natural language processing software (NLP) software and service for healthcare. Melax Technologies, Inc., has an exclusive license agreement with The University of Texas Health Science Center Houston. The technology licensed to the company involves Clinical Language Annotation Modeling and Processing (CLAMP) toolkit and the license is exclusive to software and includes services. The research will develop new interactive machine learning methods for various clinical natural language processing tasks, including named entity recognition, word sense disambiguation, and clinical phenotyping. The research contributes to the development of CLAMP.
Interactive machine learning methods for clinical natural language processing
Project Narrative In this project, we propose to develop interactive machine learning methods to process clinical text stored in electronic health records (EHRs) systems. Such methods can efficiently integrate domain and expert knowledge with machine learning processes to quickly build high-quality probabilistic models with minimum annotation cost, thus improving performance of text processors. This technology will allow more accurate data extraction from clinical documents, thus to facilitate clinical research that rely on EHRs data.
Filed on June 08, 2018.
Tell us what you know about Hua Xu's disclosure
We're still reporting about conflicts of interest. Is there something you'd like to tell us about this disclosure?
Hua Xu filed other conflict of interest disclosures with the NIH:
Name | Institution | Type | Company | Disclosed Value |
---|---|---|---|---|
Hua Xu | University of Texas Hlth Sci Ctr Houston | Conflict of Interest | Melax Technologies, inc. | $40,000 - $59,999 |
Hua Xu | University of Texas Hlth Sci Ctr Houston | Conflict of Interest | Melax Technologies, inc. | $5,000 - $9,999 |
Hua Xu | University of Texas Hlth Sci Ctr Houston | Conflict of Interest | Melax Technologies, Inc. | Value cannot be readily determined |
Hua Xu | University of Texas Hlth Sci Ctr Houston | Conflict of Interest | Melax Technologies, inc. | Value cannot be readily determined |
Hua Xu | University of California, San Diego | Conflict of Interest | Melax Technologies, Inc. | Value cannot be readily determined |
Notes: When a more specific filing date is not available for an individual financial disclosure or conflict of interest form, we use the year the form was filed. If the year was not disclosed, we report the range of years covered by our public records requests. In a few cases, a start date was provided instead of a filing date. In those cases, we use the start date instead.
Fewer than 10% of records from the University of Florida and fewer than 1% of records from the University of Texas system were removed because they did not contain enough information.
ProPublica obtained additional financial disclosures and conflict of interest forms that we have not yet digitized and added to the database. You can download those disclosures in the ProPublica Data Store.