Yahoo! Inc.

Oath Inc.

Yahoo Assets LLC

Excalibur Ip, LLC

Verizon Media Inc.

Oracle International Corporation

Yahoo Ad Tech LLC

Yi Chang

Anlei Dong

Dawei Yin

Zhaohui Zheng

Hongbo Deng

Amit Goyal

Lizi Zhang

Jiliang Tang

Ciya Liao

Achint Oommen Thomas

Hua Ouyang

Jing Bai

Georges-Eric Albert Marie Robert Dupret

Amanda Joy Stent

Bo Long

Jean-Marc Langlois

Jia Li

Keiko Horiguchi

Ricardo Alberto Baeza-Yates

Narayanan Sadagopan

Xiangnan Kong

Nageswara R Pobbathi

Yunlong He

Ricardo Baeza-Yates

Pranam Kolari

Manjana Chandrasekharan

Jeffrey Kuwano

Ruiqiang Zhang

Changsung Kang

Yuening Hu

Yoshiyuki Inagaki

Weize Kong

Shiyu Chang

Makoto Yamada

Scott J Gaffney

Shenhong Zhu

Srinivas Vadrevu

Gilbert Leung

Fernando David Diaz

Jiuhe Gan

Xuanhui Wang

Karolina Buchner

Sudarshan Lamkhede

Chao Qin

Gilad Avraham Mishne

Tim Daly

Pengyuan Wang

Avishek Saha

Jie Luo

Vignesh Murugesan

Chi-Chung Chan

Chao Tan

Robert A M Seaman, Ii

Jonathan Fan

Shihao Ji

Ling Liu

Taesup Moon

Yuanhua Lv

Disclosed are systems and methods for improving interactions with and between computers in content searching, generating, hosting and/or providing systems supported by or configured with personal computing devices, servers and/or platforms. The systems interact to identify and retrieve data within or across platforms, which can be used to improve the quality of data used in processing interactions between or among processors in such systems. The disclosed systems and methods provide a unified digital content discovery framework that implements a combination of a logistic loss function and a pair-wise loss function for information retrieval. The logistic loss function reduces non-relevant images from appearing in the retrieved results, while the pair-wise loss function ensures that the highest-quality content is included in such results. The combination of such functions provides a search information retrieval system with the novel functionality of quantifying a search results' relevance and quality in accordance with the searcher's intent.

Computerized system and method for high-quality and high-ranking digital content discovery

The present teaching relates to recommending content by analyzing the streamed data. A request is received from a user requesting one or more recommendations from a set of items. A first distribution indicative of an interest distribution of the user in a plurality of topics is obtained. For each item, a second distribution indicative of a classification distribution of the item with respect to the plurality of topics is obtained. A score is estimated based on the first distribution and the second distribution, wherein the score indicates likelihood that the user is interested in the item. The scores associated with the set of items are ranked. The one or more recommendations are presented based on the ranked scores.

Method and system for recommending content

Methods, systems and programming for providing query suggestions based on user feedback. In one example, a prefix of a query is first received. An input including a prefix of a query is received from a user in a search session. A plurality of query suggestions are fetched based on the prefix of the query. Rankings of the plurality of query suggestions are determined based, at least in part, on the user's previous interactions in the search session with respect to at least one of the plurality of query suggestions. The at least one of the plurality of query suggestions has been previously provided to the user in the search session. The plurality of query suggestions are provided in the search session based on their rankings as a response to the input.

Method and system for providing query suggestions based on user feedback

The present teaching relates to ranking search content. In one example, a plurality of documents is received to be ranked with respect to a query. Features are extracted from the query and the plurality of documents. The plurality of documents is ranked based on a ranking model and the extracted features. The ranking model is derived to remove one or more documents from the plurality of documents that are less relevant to the query and order remaining documents based on their relevance to the query. The ordered remaining documents are provided as a search result with respect to the query.

Method and system for ranking search content

At least one label prediction model is trained, or learned, using training data that may comprise training instances that may be missing one or more labels. The at least one label prediction model may be used in identifying a content item's ground-truth label set comprising an indicator for each label in the label set indicating whether or not the label is applicable to the content item.

System and method for large-scale multi-label learning using incomplete label assignments

The present teaching relates to rewriting a query and providing search results. In one example, a plurality of queries is obtained. For each of the plurality of queries, one or more search results are identified. The one or more search results have been obtained in response to the query and have been previously selected by a user submitting the query. A plurality of titles is obtained. Each of the titles corresponds to one of the one or more search results with respect to one of the plurality of queries. A model is generated based on the plurality of queries and the plurality of titles. The model is to be used for rewriting a query.

Method and system for rewriting a query

In one embodiment, a set of training data consisting of inliers may be obtained. A supervised classification model may be trained using the set of training data to identify outliers. The supervised classification model may be applied to generate an anomaly score for a data point. It may be determined whether the data point is an outlier based, at least in part, upon the anomaly score.

Large-scale anomaly detection with relative density-ratio estimation

Disclosed are systems and methods for improving interactions with and between computers in content searching, generating, hosting and/or providing systems supported by or configured with personal computing devices, servers and/or platforms. The systems interact to identify and retrieve data within or across platforms, which can be used to improve the quality of data used in processing interactions between or among processors in such systems. The disclosed systems and methods provide systems and methods for determining and suggesting query auto-completions (QACs). In some embodiments, when a user is inputting a search query, the disclosed systems and methods can provide a QAC suggestion based on the inputted text in addition to application programs installed and/or executing on the user's device.

Computerized system and method for search query auto-completion

One or more techniques and/or systems are provided for audio verification. An audio signal, comprising a code for user verification, may be identified. A second audio signal is created comprising speech. The audio signal and the second audio signal may be altered to comprise a same or similar volume, pitch, amplitude, and/or speech rate. The audio signal and the second audio signal may be combined to generate a verification audio signal. The verification audio signal may be presented to a user for the user verification. Verification may be performed to determine whether the user has access to content or a service based upon user input, obtained in response to the user verification audio signal, matching the code within the user verification audio signal. In an example, the user verification may comprise verifying that the user is human.

Audio verification

A location prediction framework is described for applying location labels or tags to target documents and/or identifying location-sensitive queries. Terms in content and queries are represented by corresponding term locations vectors (TLVs) in which the term is represented as a weighted distribution across locations. Each element of a TLV represents a probability that the term corresponding to the TLV relates to a particular location. Predicted locations may be introduced as features to a ranking framework to improve the identification and ranking of search results for a given query.

Location-sensitive ranking for search and related techniques

Predicting locations for web pages and related techniques

A system and method is described for large-scale, automated classification of products. The system and method receives information about products, wherein such information includes one or more text metadata fields associated with each product, receives a set of categories, and automatically selects one or more categories from the set of categories to which each product belongs based upon at least one of the one or more text metadata fields associated with each product. A machine learning classifier may be used to automatically select the one or more categories to which each product belongs by operating upon a feature vector for each product derived from text metadata fields of the product description. The machine learning classifier may be trained using a set of pre-categorized product descriptions. The product-category associations generated by the system and method can be used to improve search engine results or product recommendations to consumers.

Automated categorization of products in a merchant catalog

Growing community of inventors

Sunnyvale, CA, United States of America

Yi Chang