公开(公告)号：ZA202103206B

公开(公告)日：2022-08-31

申请号：ZA202103206

申请日：2021-05-12

Applicant: IBM

Inventor： LUUS FRANCOIS PIERRE ,  RIEGEL RYAN ,  AKHALWAYA ISMAIL YUNUS ,  KHAN NAWEED AGHMAD ,  VOS ETIENNE EBEN ,  MAKONDO NDIVHUWO

Abstract: Maximum expressivity can be received representing a ratio between maximum and minimum input weights to a neuron of a neural network implementing a weighted real-valued logic gate. Operator arity can be received associated with the neuron. Logical constraints associated with the weighted real-valued logic gate can be determined in terms of weights associated with inputs to the neuron, a threshold-of-truth, and a neuron threshold for activation. The threshold-of-truth can be determined as a parameter used in an activation function of the neuron, based on solving an activation optimization formulated based on the logical constraints, the activation optimization maximizing a product of expressivity representing a distribution width of input weights to the neuron and gradient quality for the neuron given the operator arity and the maximum expressivity. The neural network of logical neurons can be trained using the activation function at the neuron, the activation function using the determined threshold-of-truth.

公开(公告)号：ZA202001992B

公开(公告)日：2021-05-26

申请号：ZA202001992

申请日：2020-05-04

Applicant: IBM

Inventor： KURIEN TOBY ,  YOUNG RICHARD ,  MOLAPO MALETSABISA ,  AKHALWAYA ISMAIL YUNUS

Abstract: Language models may be run with an input set of words in a given sentence. Each of the language models can predict a set of next candidate words to follow the input set or words. Based on the sets of next candidate words predicted by the language models and an actual next word, language guidance can be provided.

公开(公告)号：ZA201904992B

公开(公告)日：2020-05-27

申请号：ZA201904992

申请日：2019-07-30

Applicant: IBM

Inventor： KURIEN TOBY ,  GRITZMAN ASHLEY D ,  AKHALWAYA ISMAIL YUNUS ,  KHAN NAWEED

Abstract: A method 30 and system 10 for facilitating safe handling of a physical medium of financial exchange, such as cash notes is provided to prevent cash-in-transit heists. The method 30 includes physically marking 33, using a first processing device 8, cash with a unique owner identifier which is associated with the owner of the cash, and physically invalidating 34 the cash through cutting, staining or stamping it. Once invalidated, the worthless cash can be transported risk-free to a desired cash processing facility. Upon arrival, the method includes verifying 44, using an authentication module 24, authenticity of the unique owner identifier and cash notes themselves. If verification fails, law enforcement is notified. If passed, the cash is destroyed 38, using a second processing device 9 which digitizes, using a digitizer 22, the cash notes and transfers a digital equivalent value, which is credited to the owner's account, to an electronic record keeping system 12.

公开(公告)号：ZA201802432B

公开(公告)日：2019-01-30

申请号：ZA201802432

申请日：2018-04-13

Applicant: IBM

Inventor： AKHALWAYA ISMAIL YUNUS

Abstract: An example embodiment of the invention provides a computer-implemented method of clustering data. The method comprises performing, by a clustering module, an nth pass of clustering on data to produce a plurality of nth clusters. A cluster analysis module calculates a plurality of components of each of the nth clusters and de-weights a largest component of each of the nth clusters to produce a plurality of de-weighted nth clusters. The clustering module performs an n+1th pass of clustering on the plurality of de-weighted nth clusters to produce a plurality of n+1th clusters. The method may be reiterated.