Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training machine learning systems. One of the methods includes receiving a plurality of training examples; and training a machine learning system on each of the plurality of training examples to determine trained values for weights of a machine learning model, wherein training the machine learning system comprises: assigning an initial value for a regularization penalty for a particular weight for a particular feature; and adjusting the initial value for the regularization penalty for the particular weight for the particular feature during the training of the machine learning system.

Abstract: Described are systems and methods for establishing and generating collections of sets that contain object identifiers based on user provided annotations for the object identifiers. A set may include one or more object identifiers and each object identifier may include one or more user provided annotations. Annotations from all object identifiers within a set are processed to form a set profile signature representative of the set. The set profile signatures are then compared between different sets to identify similar sets. Similar sets are included in a collection. Utilizing set profile signatures for a set formed based on user provided annotations provides useful relationships between sets that might otherwise not exist.

Abstract: This disclosure describes systems and methods for establishing a unit group dictionary based on user provided annotations. The unit group dictionary may be used to identify relationships between multiple items in a corpus. Those relationships may facilitate the display of object identifiers and/or other aspects used and/or provided by the object management service.

Abstract: The present disclosure provides methods and systems for using variable length representations of machine learning statistics. A method may include storing an n-bit representation of a first statistic at a first n-bit storage cell. A first update to the first statistic may be received, and it may be determined that the first update causes a first loss of precision of the first statistic as stored in the first n-bit storage cell. Accordingly, an m-bit representation of the first statistic may be stored at a first m-bit storage cell based on the determination. The first m-bit storage cell may be associated with the first n-bit storage cell. As a result, upon receiving an instruction to use the first statistic in a calculation, a combination of the n-bit representation and the m-bit representation may be used to perform the calculation.

Abstract: Methods and systems for replacing feature values of features in training data with integer values selected based on a ranking of the feature values. The methods and systems are suitable for preprocessing large-scale machine learning training data.

Abstract: The present disclosure provides systems and techniques for efficient locking of datasets in a database when updates to a dataset may be delayed. A method may include accumulating a plurality of updates to a first set of one or more values associated with one or more features. The first set of one or more values may be stored within a first database column. Next, it may be determined that a first database column update aggregation rule is satisfied. A lock assigned to at least a portion of at least a first database column may be acquired. Accordingly, one or more values in the first set within the first database column may be updated based on the plurality of updates. In an implementation, the first set of one or more values may be associated with the first lock.

Abstract: This disclosure describes systems and methods for establishing a unit group dictionary based on user provided annotations. The unit group dictionary may be used to identify relationships between multiple items in a corpus. Those relationships may facilitate the display of object identifiers and/or other aspects used and/or provided by the object management service.

Abstract: Implementations of the disclosed subject matter provide methods and systems for using a multistage learner for efficiently boosting large datasets in a machine learning system. A method may include obtaining a first plurality of examples for a machine learning system and selecting a first point in time. Next, a second point in time occurring subsequent to the first point in time may be selected. The machine learning system may be trained using m of the first plurality of examples. Each of the m examples may include a feature initially occurring after the second point in time. In addition, the machine learning system may be trained using n of the first plurality of examples, and each of the n examples may include a feature initially occurring after the first point in time.

Abstract: Systems and techniques are disclosed for training a machine learning model based on one or more regularization penalties associated with one or more features. A template having a lower regularization penalty may be given preference over a template having a higher regularization penalty. A regularization penalty may be determined based on domain knowledge. A restrictive regularization penalty may be assigned to a template based on determining that a template occurrence is below a stability threshold and may be modified if the template occurrence meets or exceeds the stability threshold.

Abstract: Systems and techniques are disclosed for generating weighted machine learned models using multi-shard combiners. A learner in a machine learning system may receive labeled positive and negative examples and workers within the learner may be configured to receive either positive or negative examples. A positive and negative statistic may be calculated for a given feature and may either be applied separately in a model or may be combined to generate an overall statistic.

Abstract: Implementations of the disclosed subject matter provide methods and systems for using a multistage learner for efficiently boosting large datasets in a machine learning system. A method may include obtaining a first plurality of examples for a machine learning system and selecting a first point in time. Next, a second point in time occurring subsequent to the first point in time may be selected. The machine learning system may be trained using m of the first plurality of examples. Each of the m examples may include a feature initially occurring after the second point in time. In addition, the machine learning system may be trained using n of the first plurality of examples, and each of the n examples may include a feature initially occurring after the first point in time.

Abstract: Systems and techniques are disclosed for training a machine learning model based on one or more regularization penalties associated with one or more features. A template having a lower regularization penalty may be given preference over a template having a higher regularization penalty. A regularization penalty may be determined based on domain knowledge. A restrictive regularization penalty may be assigned to a template based on determining that a template occurrence is below a stability threshold and may be modified if the template occurrence meets or exceeds the stability threshold.

Abstract: In general, the subject matter described in this specification can be embodied in methods, systems, and program products for providing a query to a search engine for searching a corpus of documents. A plurality of result documents are received from the search engine, each result document associated with a ranking. For a first document in the plurality, the following is performed. First, a second document in a second corpus is identified as containing content that identifies the same physical object as the first document. The second document was included in a plurality of result documents responsive to a second query of the second corpus, similar to the first query. Second, a new ranking is determined for the first document based on its ranking and relevance data associated with the second document. The relevance data is indicative of the second document's popularity as a result for the second query.

Abstract: In general, the subject matter described in this specification can be embodied in methods, systems, and program products for providing a query to a search engine for searching a corpus of documents. A plurality of result documents are received from the search engine, each result document associated with a ranking. For a first document in the plurality, the following is performed. First, a second document in a second corpus is identified as containing content that identifies the same physical object as the first document. The second document was included in a plurality of result documents responsive to a second query of the second corpus, similar to the first query. Second, a new ranking is determined for the first document based on its ranking and relevance data associated with the second document. The relevance data is indicative of the second document's popularity as a result for the second query.

Abstract: The present invention relates to systems and methods for identifying captions associated with images in media material. A captioner includes a selector module and a caption identifier module. The selector module identifies text-blocks potentially associated with images in the media material. The caption identifier module identifies which text-blocks are captions associated with images in the media material, based on the textual and proximity features of the text-block and the images. The captioner may also include a caption feedback module to modify the determining of the caption identifier module.