Yahoo Assets LLC

Yahoo! Inc.

Oath Inc.

Verizon Media Inc.

Kostas Tsioutsiouliklis

Donghyun Kim

Rao Shen

Prakhar Biyani

Yufeng Ma

Liuqing Li

Cem Akkaya

Siddhartha Banerjee

Yu Wang

Gilbert Leung

Farzin Maghoul

Jean-Marc Langlois

Su-Lin Wu

Arkady Borkovsky

Lei Duan

Dmitri Pavlovski

Su Chan

Alexandrin Popescul

Hongwei Shang

Changsung Kang

Clifford A Brunk

Alexander Smola

Novi Quadrianto

Nadav A Zimak

Sainath Vellal

Sanika Shirwadkar

Nicolas Torzec

In an example, in response to a request for content, content items may be identified. Content item scores associated with the content items may be determined using a content scoring model. A first ranked list may be generated based upon the content item scores using an exploration model. The first ranked list is indicative of rankings of the content items. A content item selection interface may be generated based upon the first ranked list. The content item selection interface may include selectable objects, corresponding to the content items, arranged according to rankings indicated by the first ranked list.

System and method for ranking content items

The disclosed systems and methods provide a novel framework that provides mechanisms for a Deep & Cross Network (DCN) framework that performs distilled deep prediction for personalized stream ranking on portal websites. The disclosed framework is scalable to satisfy the much more stringent latency and computational requirements required by current network operating environments. The disclosed framework is able to dynamically evaluate and leverage live traffic on network sites in order to provide, update and maintain current recommendations for users as they traverse to a portal and when they navigate within the portal. The disclosed framework implements a DCN model(s) that is capable of being compressed into a model size for a unified optimization within a live traffic environment by combining knowledge distillation and model compression techniques. The disclosed framework is built as a light-weight deep learning model that can be served in production and perform on par with large models.

Computerized system and method for distilled deep prediction for personalized stream ranking

Techniques for automatic intelligent information extraction from an electronic document are disclosed. In one embodiment, a computerized method is disclosed comprising training a label prediction model to generate a set of label predictions, obtaining an electronic document, analyzing the electronic document and determining a set of features for each of a set of information items identified in the electronic document, obtaining model output from the label prediction model for each information item, the model output comprising, for a respective information item, a set of probabilities corresponding to a set of information classes, and generating an information extraction comprising a set of labels corresponding to the set of information items.

Electronic information extraction using a machine-learned model architecture method and apparatus

One or more computing devices, systems, and/or methods for debiasing training data based upon information seeking behaviors are provided. Users associated with a set of training data are segmented into information seeking behavior groups corresponding to varying degrees of information seeking behaviors of the users. Biases for the information seeking behavior groups may be estimated based upon information seeking behaviors of users within the information seeking behavior groups. The training set of data is debiased using the biases to generate a debiased training set of data. A model may be trained to perform a task based upon the debiased training set of data.

Debiasing training data based upon information seeking behaviors

One or more computing devices, systems, and/or methods for oversampling for imbalanced test data are provided. A classifier is configured to classify data points as either belonging to a first class or a second class. A determination may be made that the first class and the second class are imbalanced where a first number of data points estimated to be part of the first class is a threshold amount less than a second number of data points estimated to be part of the second class. An oversampling ratio is determined for the first class. The oversampling ratio is used to select a sample set of data points for editorial labeling, where the sampling set of data points comprises a total number of data points below a threshold amount.

Oversampling for imbalanced test data

In an example, first entities are extracted from user profiles. Second entities are extracted from content information associated with content item. User-associated metrics associated with the first entities are determined based upon the user profiles and/or content events. First vector representations of the first entities and second vector representations of the second entities are processed to generate an attention distribution array. Each value of the attention distribution array represents, for a user interested in an entity of the first entities, a proportion of (i) entity-specific activity, of the user, related to an entity of the second entities relative to (ii) an entirety of activity of the user. An inferred activity distribution array is generated by applying the user-associated metrics to the attention distribution array. A filtered subset of activity distribution values is generated by pruning values from the inferred activity distribution array. Transmission of content is controlled using the filtered subset of activity distribution values.

System and method for feature determination and content selection

Methods, programming, and system for modifying a slot value are described herein. In a non-limiting embodiment, an intent may be determined based on a first utterance. A first slot-value pair may be obtained for the first utterance based on the intent, the first slot-value pair including a first slot and a first value associated with the first slot. A second value associated with the first slot may be identified, the second value being identified from a second utterance that was previously received. Based on the intent and the first slot, a type of update to be performed with respect to the second value may be determined. The second value may then be updated based on the first value and the type of update.

System and method for rule based modifications to variable slots based on context

The present teaching relates to method and system for generating a stream of content items. A plurality of entities associated with a time-window are obtained, wherein each entity of the plurality of entities is associated with at least one content item. For each entity, a first parameter with respect to the time-window, a second parameter with respect to previous time-windows, and a trendiness score based on a function of the first parameter and the second parameter are respectively calculated. A graph based on one or more entity-pairs is generated, wherein each entity-pair of the one or more entity-pairs satisfies a first criterion. A stream of content items is generated based on the graph, wherein each content item in the stream of content items corresponds to at least one of the one or more entity pairs.

Method and system for data mining

The present teaching relates to personalized content recommendation. A webpage is contrasted for a user having a plurality of slots each of which is to be allocated with a content item. For each of the plurality of slots, a plurality of content items in a plurality of types of content are accessed. For each of the plurality of types of content, a personalized score is predicted for each content item in the type of content, wherein the personalized score is obtained based on a trained model trained. A recommended content item of the type of content is selected based on personalized scores. An overall recommended content item is selected and allocated to a slot based on criteria associated with the personalized scores of the recommended content items and a business rule. The webpage with the plurality of slots allocated with content items is provided to the user.

Method and system of personalized blending for content recommendation

Methods, programming, and system for identifying one or more variable slots within an utterance are described herein. In a non-limiting embodiment, a first slot-value pair for an utterance may be obtained. The first slot-value pair may include a first slot and a first value associated with the slot. The first slot may be of a first entity type, where an intent and a data object are estimated based on the first utterance. A data structure representing the data object may be identified. Based on the intent, a first variable slot in the data structure associated with the first entity type may be determined, where the first variable slot may be associated with at least one of: multiple values and an adjustable value. Based on the intent, the first value may be assigned to the first variable slot in the data structure.

System and method for identifying and replacing slots with variable slots

The present teaching relates to method, system, and medium for generating training data for generating a relationship identification model. Sentences are received as input. Each of the sentences is aligned with a fact previously stored to create an alignment. Confidence scores for the alignments are computed and then used, together with the alignments to train a relationship identification model.

System and method for relationship identification

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 disclosure provides a computerized framework for automatically generating chatbot responses to produce domain-specific responses that mimic native styles unique to particular domains. The disclosed systems and methods construct domain-specific word-graphs based on account activity from specific domains and generate word-patterns. New words obtained from the patterns in the graph are introduced to transform the regular response. The graph is then pruned using data-driven thresholds in order to avoid spurious transformations, and paragraph vectors are also utilized to assign relevance scores to generated patterns such that only the patterns that are contextually similar to the original response (generic/regular response) are used. As result, the regular chatbot response is rewritten using an optimized set of patterns.

Computerized system and method for automatically transforming and providing domain specific chatbot responses

A system and method is described herein that cache web documents in two or more caches in an optimized manner. The web documents result from a plurality of queries presented to a web search engine. Each web document has a tier score indicating a position of the web document in the caches. The tier scores are modified in a manner that rearranges the web documents in caches. This rearrangement of the web documents in the caches can advantageously improve the retrieval time for the web documents.

Growing community of inventors

San Jose, CA, United States of America

Kostas Tsioutsiouliklis