A system and method for handling processing with outliers. In some embodiments, the method includes: reading a first activation and a second activation, each including a least significant part and a most significant part, multiplying a first weight and a second weight by the respective activations, the multiplying of the first weight by the first activation including multiplying the first weight by the least significant part of the first activation in a first multiplier, the multiplying of the second weight by the second activation including: multiplying the second weight by the least significant part of the second activation in a second multiplier, and multiplying the second weight by the most significant part of the second activation in a shared multiplier, the shared multiplier being associated with a plurality of rows of an array of activations.

A skyrmion logic gate is provided. The logic gate comprises a first track configured for propagation of magnetic skyrmions and a second track configured for propagation of magnetic skyrmions. A junction links the first and second tracks. A continuous current flows through the logic gate, wherein skyrmions propagate due to the current.

A skyrmion logic gate is provided. The logic gate comprises a first track configured for propagation of magnetic skyrmions and a second track configured for propagation of magnetic skyrmions. A junction links the first and second tracks. A continuous current flows through the logic gate, wherein skyrmions propagate due to the current.

Aspects for Long Short-Term Memory (LSTM) blocks in a recurrent neural network (RNN) are described herein. As an example, the aspects may include one or more slave computation modules, an interconnection unit, and a master computation module collectively configured to calculate an activated input gate value, an activated forget gate value, a current cell status of the current computation period, an activated output gate value, and a forward pass result.

Aspects for Long Short-Term Memory (LSTM) blocks in a recurrent neural network (RNN) are described herein. As an example, the aspects may include one or more slave computation modules, an interconnection unit, and a master computation module collectively configured to calculate an activated input gate value, an activated forget gate value, a current cell status of the current computation period, an activated output gate value, and a forward pass result.

Aspects of the disclosure provide for methods, systems, and apparatuses, including computer-readable storage media, for sparse matrix multiplication. A system for matrix multiplication includes an array of sparse shards. Each sparse shard can be configured to receive an input sub-matrix and an input sub-vector, where the input sub-matrix has a number of non-zero values equal to or less than a predetermined maximum non-zero threshold. The sparse shard can, by a plurality of multiplier circuits, compute one or more products of vector values multiplied with respective non-zero values of the input sub-matrix. The sparse shard can generate, as output to the sparse shard and using the one or more products, a shard output vector that is the product of applying the shard input vector to the shard input matrix.

Aspects of the disclosure provide for methods, systems, and apparatuses, including computer-readable storage media, for sparse matrix multiplication. A system for matrix multiplication includes an array of sparse shards. Each sparse shard can be configured to receive an input sub-matrix and an input sub-vector, where the input sub-matrix has a number of non-zero values equal to or less than a predetermined maximum non-zero threshold. The sparse shard can, by a plurality of multiplier circuits, compute one or more products of vector values multiplied with respective non-zero values of the input sub-matrix. The sparse shard can generate, as output to the sparse shard and using the one or more products, a shard output vector that is the product of applying the shard input vector to the shard input matrix.

A module for processing a set of n input data comprising a reference datum, which is configured to a) calculate the sum of weights that are associated only with those input data other than the reference datum which have values strictly lower than the value of the reference datum; b) compare the sum calculated in step a) with a first threshold; c) calculate the sum of the weights that are associated only with those input data other than the reference datum which have values strictly higher than the value of the reference datum; d) compare the sum calculated in step c) with a second threshold; e) generate an output datum indicating whether the reference datum is a fractile of the set included in a predetermined subset of fractiles, only if: the sum calculated in step a) is lower than the first threshold and the sum calculated in step c) is lower than the second threshold.

A module for processing a set of n input data comprising a reference datum, which is configured to a) calculate the sum of weights that are associated only with those input data other than the reference datum which have values strictly lower than the value of the reference datum; b) compare the sum calculated in step a) with a first threshold; c) calculate the sum of the weights that are associated only with those input data other than the reference datum which have values strictly higher than the value of the reference datum; d) compare the sum calculated in step c) with a second threshold; e) generate an output datum indicating whether the reference datum is a fractile of the set included in a predetermined subset of fractiles, only if: the sum calculated in step a) is lower than the first threshold and the sum calculated in step c) is lower than the second threshold.

An artificial intelligence (AI) accelerator includes memory circuits configured to output weight data and vector data, a multiplication circuit/adder tree performing a multiplying/adding calculation on the weight data and the vector data to generate multiplication/addition result data, a first accumulator synchronized with an odd clock signal to perform an accumulative adding calculation on odd-numbered multiplication/addition result data of the multiplication/addition result data and a first latched data, and a second accumulator synchronized with an even clock signal to perform an accumulative adding calculation on even-numbered multiplication/addition result data of the multiplication/addition result data and a second latched data.