Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for post-hoc deferral for classification tasks. In particular, a system can perform either post-hoc threshold correction or post-hoc rejector training to account for the cost of deferring model inputs to an expert system for classification.

Abstract: The disclosure is directed towards automatically tuning quantization-based approximate nearest neighbors (ANN) search methods and systems (e.g., search engines) to perform at the speed-recall pareto frontier. With a desired search cost or recall as input, the embodiments employ Lagrangian-based methods to perform constrained optimization on theoretically-grounded search cost and recall models. The resulting tunings, when paired with the efficient quantization-based ANN implementation of the embodiments, exhibit excellent performance on standard benchmarks while requiring minimal tuning or configuration complexity.

Abstract: The present disclosure provides systems and methods for communication efficient distributed mean estimation. In particular, aspects of the present disclosure can be implemented by a system in which a number of vectors reside on a number of different clients, and a centralized server device seeks to estimate the mean of such vectors. According to one aspect of the present disclosure, a client computing device can rotate a vector by a random rotation matrix and then subsequently perform probabilistic quantization on the rotated vector. According to another aspect of the present disclosure, subsequent to quantization but prior to transmission, the client computing can encode the quantized vector according to a variable length coding scheme (e.g., by computing variable length codes).

Abstract: A wireline operation method is disclosed. The method includes disposing an outer diameter (OD) meter ring on a wireline, the OD meter ring having spring-loaded sensors, releasing, from a spooling drum into the wellbore, the wireline via a wireline passage having a tight spot, a released portion of the wireline progressively passing through an opening of the OD meter ring, generating, using the spring-loaded sensors, a real time OD measurement of a current location of the wireline passing through the opening of the OD meter ring at the time of measurement, generating, in response to the real time OD measurement of a particular location of the wireline exceeding a pre-determined threshold, a fail-safe signal, and stopping, in response to the fail-safe signal, the spooling drum from continuing to release the wireline before the particular location of the wireline reaches the tight spot of the wireline passage.

Abstract: Generally, the present disclosure is directed to systems and methods of quantizing a database with respect to a novel loss or quantization error function which applies a weight to an error measurement of quantized elements respectively corresponding to the datapoints in the database. The weight is determined based on the magnitude of an inner product between the respective datapoints and a query compared therewith. In contrast to previous work, embodiments of the proposed loss function are responsive to the expected magnitude of an inner product between the respective datapoints and a query compared therewith and can prioritize error reduction for higher-ranked pairings of the query and the datapoints. Thus, the systems and methods of the present disclosure provide solutions to some of the problems with traditional quantization approaches, which regard all error as equally impactful.

Abstract: A cylindrical main body is configured to be inserted into a well tree. A laser emitter is attached to the main body. The laser emitter is arranged to emit a laser beam configured to cut through a metal valve within a well tree. The laser emitter is arranged to emit the laser beam towards the metal valve. A magnet is on a distal end of the body. The magnet is arranged to magnetically capture a sliced portion of the metal valve cut by the laser beam.

Abstract: The present disclosure provides systems and methods that include or otherwise leverage use of a multiscale quantization model that is configured to provide a quantized dataset. In particular, the multiscale quantization model can receive and perform vector quantization of a first dataset. The multiscale quantization model can generate a residual dataset based at least in part on a result of the vector quantization. The multiscale quantization model can apply a rotation matrix to the residual dataset to generate a rotated residual dataset that includes a plurality of rotated residuals. The multiscale quantization model can perform reparameterization of each rotated residual in the rotated residual dataset into a direction component and a scale component. The multiscale quantization model can perform product quantization of the direction components of the plurality of rotated residuals, and perform scalar quantization of the scale components of the plurality of rotated residuals.

Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for performing a kNN computation using a hardware accelerator. One of the methods includes obtaining a set of one or more query vectors; obtaining a set of database vectors; and performing, on a hardware accelerator and for each query vector in the set, a search for the k most similar database vectors to the query vector, comprising: computing, by circuitry of the hardware accelerator and for each query vector, a respective similarity value between the query vector and each database vector; and for each query vector, identifying, by the hardware accelerator and for each bin, (i) an index of the most similar database vector within the bin and (ii) the respective similarity value for the most similar database vector within the bin.

Abstract: A cylindrical main body is configured to be inserted into a well tree. A laser emitter is attached to the main body. The laser emitter is arranged to emit a laser beam configured to cut through a metal valve within a well tree. The laser emitter is arranged to emit the laser beam towards the metal valve. A magnet is on a distal end of the body. The magnet is arranged to magnetically capture a sliced portion of the metal valve cut by the laser beam.

Abstract: Generally, the present disclosure is directed to systems and methods that perform adaptive optimization with improved convergence properties. The adaptive optimization techniques described herein are useful in various optimization scenarios, including, for example, training a machine-learned model such as, for example, a neural network. In particular, according to one aspect of the present disclosure, a system implementing the adaptive optimization technique can, over a plurality of iterations, employ an adaptive effective learning rate while also ensuring that the effective learning rate is non-increasing.

Abstract: A system includes a flexi-string having a first lateral end, a second lateral end, and an outer diameter. The outer diameter defines an outer circumferential surface, and the outer diameter is smaller than an inner diameter of a casing shoe. The system further includes an anchor disposed on the outer circumferential surface of the first lateral end, where the anchor interacts with the casing shoe, or a last joint of casing, to hold the first lateral end of the flexi-string within the casing shoe. The system also has centralizers located between the first lateral end and the second lateral end of the flexi-string. The centralizers are configured to center the flexi-string within the casing shoe and the washout section. Finally, the system includes a roller guide for lowering the flexi-string inside the casing shoe and to a depth below the washout section of the well.

Abstract: The present disclosure provides systems and methods for communication efficient distributed mean estimation. In particular, aspects of the present disclosure can be implemented by a system in which a number of vectors reside on a number of different clients, and a centralized server device seeks to estimate the mean of such vectors. According to one aspect of the present disclosure, a client computing device can rotate a vector by a random rotation matrix and then subsequently perform probabilistic quantization on the rotated vector. According to another aspect of the present disclosure, subsequent to quantization but prior to transmission, the client computing can encode the quantized vector according to a variable length coding scheme (e.g., by computing variable length codes).

Abstract: Generally, the present disclosure is directed to systems and methods of quantizing a database with respect to a novel loss or quantization error function which applies a weight to an error measurement of quantized elements respectively corresponding to the datapoints in the database. The weight is determined based on the magnitude of an inner product between the respective datapoints and a query compared therewith. In contrast to previous work, embodiments of the proposed loss function are responsive to the expected magnitude of an inner product between the respective datapoints and a query compared therewith and can prioritize error reduction for higher-ranked pairings of the query and the datapoints. Thus, the systems and methods of the present disclosure provide solutions to some of the problems with traditional quantization approaches, which regard all error as equally impactful.

Abstract: Generally, the present disclosure is directed to systems and methods that perform adaptive optimization with improved convergence properties. The adaptive optimization techniques described herein are useful in various optimization scenarios, including, for example, training a machine-learned model such as, for example, a neural network. In particular, according to one aspect of the present disclosure, a system implementing the adaptive optimization technique can, over a plurality of iterations, employ an adaptive effective learning rate while also ensuring that the effective learning rate is non-increasing.

Abstract: A system includes a rod, a nozzle, a camera, and a pressure control system. The rod has a conduit extending along a rod axis. The rod is disposed adjacent to the barrier within the orifice. The nozzle is connected to the rod and is in hydraulic communication with the conduit and the orifice. The camera is also connected to the rod. The nozzle and the camera are configured to protrude from the rod and rotate about the rod axis. The pressure control system is connected to the tubular body and is disposed around the rod. The nozzle is configured to cut away the portion of the barrier upon pumping of a cutting fluid through a conduit of the rod and out of the nozzle towards the barrier. The portion of the barrier is removable from the orifice of the tubular body once cut away.

Abstract: In order to provide a bipolar plate for a fuel cell, providing an anode plate with an anode side and a coolant side, wherein a first structuring for forming an anode flow field is formed on the anode side, and a cathode plate with a cathode side and a coolant side, wherein a second structuring for forming a cathode flow field is formed on the cathode side; wherein structural elements, which are contacted by the coolant sides of the anode plate and the cathode plate, for forming a coolant flow field, are arranged between the anode plate and the cathode plate, which bipolar plate has an optimized pressure distribution in a fuel cell stack and increased stability in comparison with the prior art, it is proposed that the structural elements may be made of an elastic material and that the structural elements have a different height in different regions of the coolant flow field. A fuel cell stack and a vehicle are also disclosed.

Abstract: A method for training a machine learning model, e.g., a neural network, using a regularization scheme is disclosed. The method includes generating regularized partial gradients of losses computed using an objective function for training the machine learning model.

Abstract: Provided are systems and methods which more efficiency train embedding models through the use of a cache of item embeddings for candidate items over a number of training iterations. The cached item embeddings can be “stale” embeddings that were generated by a previous version of the model at a previous training iteration. Specifically, at each iteration, the (potentially stale) item embeddings included in the cache can be used when generating similarity scores that are the basis for sampling a number of items to use as negatives in the current training iteration. For example, a Gumbel-Max sampling approach can be used to sample negative items that will enable an approximation of a true gradient. New embeddings can be generated for the sampled negative items and can be used to train the model at the current iteration.

Abstract: Implementations provide for use of spherical random features for polynomial kernels and large-scale learning. An example method includes receiving a polynomial kernel, approximating the polynomial kernel by generating a nonlinear randomized feature map, and storing the nonlinear feature map. Generating the nonlinear randomized feature map includes determining optimal coefficient values and standard deviation values for the polynomial kernel, determining an optimal probability distribution of vector values for the polynomial kernel based on a sum of Gaussian kernels that use the optimal coefficient values, selecting a sample of the vectors, and determining the nonlinear randomized feature map using the sampled vectors.

Abstract: An apparatus includes a gauge cutter, a lead impression block (LIB), a locking pin, and a locking device. The gauge cutter has a gauge cutter tubular body, a gauge cutter head, and an orifice. The LIB has an LIB body and an LIB head. The LIB is fixed in a first position within the orifice of the gauge cutter by being connected to an inner circumferential surface of the gauge cutter tubular body using a shear pin. The locking pin is disposed around the LIB body. The locking device is disposed around the inner circumferential surface of the gauge cutter tubular body and is configured to mate with the locking pin to hold the LIB in a second position within the gauge cutter. The second position includes a portion of the LIB body located within the orifice and the LIB head located outside of the orifice.