The IQ application will be unavailable on Sunday, November 24th, starting at 9:00am ET while we make system improvements. Access will be restored as quickly as possible.

Month of January 2019 - Page Number 22

Showing 211 - 220 of 513 from January 2019
Browse Prior Art Database
  1. 211.
    A method and system is disclosed for scaling up a factorization machine by integrating a parameter server with a map/reduce framework. The method and system exploits data parallelism achieved via the map/reduce framework along with model parallelism of the parameter server in the factorization machine for learning...
    IPCOM000257106D | 2019-Jan-15
  2. 212.
    IPCOM000257105D | 2019-Jan-15
  3. 213.
    A method and system is disclosed for facilitating users to customize search page results. The method and system provides a Search Engine Product Filter (SEPF), which allows a user to interact with search results (products) and customize the search results on search results page, thereby increasing the users'...
    IPCOM000257104D | 2019-Jan-15
  4. 214.
    A method and system is disclosed for performing Student's T-Test in a simple moving operation by using an implementation of K-minimum values sketch with associated summary objects. Thus, the method and system obviates the need to process the entire data set in a single pass for each analysis, and enables analysis of...
    IPCOM000257103D | 2019-Jan-15
  5. 215.
    A method and system is disclosed for optimizing delivery and return on investment (ROI) with target margin for Real-time Bidding (RTB). The method and system includes an automated RTB auction for displaying advertisements (ads) based on performance of the ads to maximize revenue while supporting fixed positive margin.
    IPCOM000257102D | 2019-Jan-15
  6. 216.
    A method and system is disclosed for providing an ensemble model for efficiently predicting conversion rates. The method and system utilizes inverse variance weighted average of predictions from various weak models to significantly improve conversion rate predictions.
    IPCOM000257101D | 2019-Jan-15
  7. 217.
    A method and system is disclosed for utilizing field-weighted factorization machines (FwFM) for Click-through Rate (CTR) prediction and recommendation. FwFM analyze different correlations between different field pairs, and labels the field pairs by considering field information using several different latent vectors...
    IPCOM000257100D | 2019-Jan-15
  8. 218.
    A method and system is disclosed for computing a user reputation score based on various social signals in an iterative and distributed fashion, utilizing EigenTrust, a probabilistic trust measurement algorithm used in peer-to-peer network applications. The algorithm uses various social signals between interacting and...
    IPCOM000257099D | 2019-Jan-15
  9. 219.
    A method and system is disclosed for improving predictions of multi-label learning (MLL) approaches using post-processing techniques. The method and system presents a suite of MLL algorithm with independent post-processing techniques, which utilizes conditional and directional label-dependences in order to make the...
    IPCOM000257098D | 2019-Jan-15
  10. 220.
    IPCOM000257097D | 2019-Jan-14