Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a machine learning model. One of the methods includes receiving a complete set of training data; receiving instructions to train a predictive model having a plurality of parameters on an initial subset of the complete set of training data; training the predictive model on the initial subset; storing data representing a first state of the predictive model after training the predictive model on the initial subset; receiving updated parameter values and instructions to train the predictive model on a new subset of the complete set of training data; and training the predictive model on the new subset.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training a machine learning model. One of the methods includes performing experiments to select a training strategy for use in training the model on a particular data set. The selected training strategy includes a binning strategy for binning the raw feature vectors before the raw feature vectors are provided to the predictive model.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for training machine learning models. The models can include models for predicting a next transaction price or a next transaction price direction for one or more financial products, for classifying particular debit or credit card transactions as likely being anomalous or fraudulent or not, or for classifying particular financial claims processing transactions, e.g., insurance, health care, or employee expense claims transactions, as likely being anomalous or fraudulent or not.

Abstract: A security component determines whether a request for a resource poses a security risk to a computing device and verifies the integrity of the requested resource before the request is allowed. For a request having arguments and a resource path with a filename that identifies the resource, the security component determines that the request does not pose a security risk if the resource path does not exceed a maximum number of characters, individual arguments do not exceed a maximum number of characters, the arguments combined do not exceed a maximum number of characters, and the filename has a valid extension. The security component verifies the integrity of a requested resource by formulating a descriptor corresponding to the resource and comparing the descriptor with a cached descriptor corresponding to the resource.

Abstract: A system and method that evaluates privacy policies from web sites to determine whether each site is permitted to perform operations (e.g., store, retrieve or delete) directed to cookies on a user's computer. Various properties of each cookie and the context in which it is being used are evaluated against a user's privacy preference settings to make the determination. An evaluation engine accomplishes the evaluation and determination via a number of criteria and considerations, including the cookie properties, its current context, the site, the zone that contains the site, and any P3P data (compact policy) provided with the site's response. The user privacy preferences are evaluated against these criteria to determine whether a requested cookie operation is allowed, denied or modified. A formalized distinction between first-party cookies versus third-party cookies may be used in the determination, along with whether the cookie is a persistent cookie or a session cookie.