Abstract: Systems and methods for training machine learning models using emulated datasets are provided. In some instances, it may be undesirable for certain portions of an initial training dataset to be used to train a machine learning model. In other instances, a model may be trained using an initial training dataset and it may be desired to cause the machine learning model to “forget” certain aspects of the training. To address both of these scenarios, a generative machine learning model may be used to generate an emulated training dataset that is based on the initial training dataset. The emulated training dataset is then used to train the machine learning model that is ultimately used to perform downstream tasks. This allows the machine learning model to still be trained to perform the task, without the risk of the machine learning model being exposed to the portions of the training data that are undesired to be used for training.

Abstract: The breath synchronization lighting devices and apparatus helps the user to regulate their breathing rate by moving a light source with respect to a stationary shade or cover. The shade or cover, optionally perforated or with an integrated obscuring pattern, partially blocks the light created by the light source. The result is the generation of a shadow pattern on adjacent surfaces. As the light moves with respect to the cover, the shadow pattern correspondingly moves on the surface is surrounding the device. The result is a cycling shadow pattern on adjacent surfaces, the pattern timed to match the ideal breathing rate of the user, helping the user to slow their breathing and reach a calm and meditative state.

Abstract: An algorithm releases answers to very large numbers of statistical queries, e.g., k-way marginals, subject to differential privacy. The algorithm answers queries on a private dataset using simple perturbation, and then attempts to find a synthetic dataset that most closely matches the noisy answers. The algorithm uses a continuous relaxation of the synthetic dataset domain which makes the projection loss differentiable, and allows the use of efficient machine learning optimization techniques and tooling. Rather than answering all queries up front, the algorithm makes judicious use of a privacy budget by iteratively and adaptively finding queries for which relaxed synthetic data has high error, and then repeating the projection. The algorithm is effective across a range of parameters and datasets, especially when a privacy budget is small or a query class is large.

Abstract: An algorithm releases answers to very large numbers of statistical queries, e.g., k-way marginals, subject to differential privacy. The algorithm answers queries on a private dataset using simple perturbation, and then attempts to find a synthetic dataset that most closely matches the noisy answers. The algorithm uses a continuous relaxation of the synthetic dataset domain which makes the projection loss differentiable, and allows the use of efficient machine learning optimization techniques and tooling. Rather than answering all queries up front, the algorithm makes judicious use of a privacy budget by iteratively and adaptively finding queries for which relaxed synthetic data has high error, and then repeating the projection. The algorithm is effective across a range of parameters and datasets, especially when a privacy budget is small or a query class is large.

Abstract: In an idle state transition from a macro mobile network to a femtocell, a mobile station looks for the femtocell to operate in a frequency band typically assigned to a macro network. For normal macro network operations, mobile stations look for base stations operating on various channels, e.g. based on hashing of respective mobile station identifiers. Femtocells generate a pilot beacon on a carrier frequency to enable mobile station acquisition, e.g. to redirect a mobile station to the operating frequency of the femtocell. To avoid the need to continually cycle or hop through all possible macro assigned carrier frequencies, pilot beacon transmissions of a particular femtocell are prioritized to correspond to the mobile station(s) expected to utilize that femtocell, e.g. by using existing network information, provisioning device information or self learning/calculating information regarding the carrier most likely to be used by any particular mobile station expected to use the femtocell.

Abstract: A novel business method for selling automobiles includes a unique multiply form that includes three different data fields. The form is used in a paperwork tracking technique that comprises multiple sheets having a common data field in a pre-selected area. Each sheet includes other data fields that are unique to that particular form. Each sheet is stripped from the multi-ply form and forwarded to an appropriate department in an automobile dealership.

Abstract: A part sequencing system is shown that can receive requests for a plurality of inventory items to be supplied in a specified order. The specified order is related to the order in which the inventory will be consumed in a selected manufacturing process. The system may utilize object oriented programming to reduce input errors.