Abstract: Embodiments of the present disclosure relate to a method, system, and computer program product for efficient execution of a decision tree. According to the method, respective target values of a plurality of attributes of a target entity are obtained. Representations of a plurality of leaf nodes of a decision tree are obtained. Each of the representations indicates respective statistic values of a plurality of attributes of historical entities and a statistic prediction result determined from historical prediction results output at a respective one of the plurality of leaf nodes for the historical entities. Distance measures between the target entity and the plurality of leaf nodes are determined based on the target values and the statistic values indicated by the representations of the plurality of leaf nodes. A target prediction result for the target entity is determined based on the distance measures and the statistic prediction results of the historical entities.

Abstract: An approach is provided in which the approach generates anomaly score variables using multiple unsupervised models based on a set of data records. The approach normalizes the anomaly score variables into multiple normalized variables, and constructs at least one interaction based on a first one of the normalized variables and a second one of the normalized variables. The first normalized variable corresponds to a first one of the anomaly score variables and the second normalized variable corresponds to a second one of the anomaly score variables. The approach detects a set of anomalies based on the at least one interaction and transmits the set of anomalies to a user.

Abstract: A computer implemented method for machine learning model training. A number of processor units creates a cluster model comprising labeled samples and unlabeled samples. The number of processor units identifies cluster information for the labeled samples from the cluster model. The number of processor units adds a set of new features to a set of original features for the labeled samples using the cluster information to form an extended set of features for the labeled samples, wherein the labeled samples with the set of original features and the set of new features form a training data set for training a machine learning model.

Abstract: A method includes training a plurality of different types of machine learning models using a training dataset to produce a set of trained machine learning models and determining a lift of each trained machine learning model in the set of trained machine learning models using a validation dataset. The method also includes selecting a trained machine learning model from the set of trained machine learning models that has a highest lift of the set of trained machine learning models and predicting a likelihood that a person will perform an action by applying the selected trained machine learning model to data about the person.

Abstract: A computer-implemented method, system and computer program product for improving prediction accuracy in machine learning techniques. A teacher model is constructed, where the teacher model generates a weight for each data case. The current student model is then trained using training data and the weights generated by the teacher model. After training the current student model, the current student model generates state features, which are used by the teacher model to generate new weights. A candidate student model is then trained using training data and these new weights. A reward is generated by comparing the current student model with the candidate student model using training and testing data, which is used to update the teacher model if a stopping rule has not been satisfied. Upon a stopping rule being satisfied, the weights generated by the teacher model are deemed to be the “optimal” weights which are returned to the user.

Abstract: A computer-implemented method, system and computer program product for improving prediction accuracy in machine learning techniques. A teacher model is constructed, where the teacher model generates a weight for each data case. The current student model is then trained using training data and the weights generated by the teacher model. After training the current student model, the current student model generates state features, which are used by the teacher model to generate new weights. A candidate student model is then trained using training data and these new weights. A reward is generated by comparing the current student model with the candidate student model using training and testing data, which is used to update the teacher model if a stopping rule has not been satisfied. Upon a stopping rule being satisfied, the weights generated by the teacher model are deemed to be the “optimal” weights which are returned to the user.

Abstract: A computer-implemented method, system and computer program product for processing a data set is provided. In this method, an original data set including a plurality of data records is obtained. Each data record in the original data set has values of a first number of features. A representative data set having the plurality of representative data records is determined. Each representative data record has values of a second number of representatives. The second number of representatives are obtained by training an autoencoder neutral network with values of the first number of features as inputs, and the second number is smaller than the first number. The plurality of representative data records is segmented into two or more clusters based on the values of the second number of representatives. The representative data records in the two or more clusters are partitioned to form a predefined number of representative data subsets.

Abstract: Embodiments of the present disclosure relate to a method, system, and computer program product for efficient execution of a decision tree. According to the method, respective target values of a plurality of attributes of a target entity are obtained. Representations of a plurality of leaf nodes of a decision tree are obtained. Each of the representations indicates respective statistic values of a plurality of attributes of historical entities and a statistic prediction result determined from historical prediction results output at a respective one of the plurality of leaf nodes for the historical entities. Distance measures between the target entity and the plurality of leaf nodes are determined based on the target values and the statistic values indicated by the representations of the plurality of leaf nodes. A target prediction result for the target entity is determined based on the distance measures and the statistic prediction results of the historical entities.

Abstract: Techniques to ensemble machine learning (ML) models are provided. A plurality of residues is generated by processing a plurality of input records using a plurality of ML models. A plurality of data clusters is identified by evaluating, using a clustering model, the plurality of input records and the plurality of residues. A first ensemble is generated for a first data cluster of the plurality of data clusters, where the first ensemble comprises one or more of the plurality of ML models. Upon determining that a new input record corresponds to the first data cluster, the new input record is processed using the first ensemble.

Abstract: A computer-implemented method, system and computer program product for improving prediction accuracy in machine learning techniques. A teacher model is constructed, where the teacher model generates a weight for each data case. The current student model is then trained using training data and the weights generated by the teacher model. After training the current student model, the current student model generates state features, which are used by the teacher model to generate new weights. A candidate student model is then trained using training data and these new weights. A reward is generated by comparing the current student model with the candidate student model using training and testing data, which is used to update the teacher model if a stopping rule has not been satisfied. Upon a stopping rule being satisfied, the weights generated by the teacher model are deemed to be the “optimal” weights which are returned to the user.

Abstract: Evaluating data partition quality is provided. A historical data set is partitioned into a specified number of partitions. A quality of each partition in the specified number of partitions is evaluated by measuring a distribution similarity between variables from each data subset in a respective partition and the historical data set. A highest-quality partition in the specified number of partitions is recommended to build a supervised machine learning model based on the highest-quality partition having a highest variable distribution similarity measure with the historical data set.

Abstract: Embodiments of the invention may be used to produce a data mining signal by generating hybrid dataset representing data related to tools used during a semiconductor fabrication process. By selectively combining similar processes, the data mining signal strength of each tool used to perform the steps of the fabrication process may be increased. A combined process variable may be used to represent the group of tools and processes, collectively. A set of rules may be composed to determine which processes used in the semiconductor fabrication process should be combined in the hybrid dataset.

Abstract: Embodiments of the invention may be used to produce a data mining signal by generating hybrid dataset representing data related to tools used during a semiconductor fabrication process. By selectively combining similar processes, the data mining signal strength of each tool used to perform the steps of the fabrication process may be increased. A combined process variable may be used to represent the group of tools and processes, collectively. A set of rules may be composed to determine which processes used in the semiconductor fabrication process should be combined in the hybrid dataset.

Abstract: A method is described for fabricating a cloisonné structure, in which a top surface of a metal oxide layer is made coplanar with a top surface of a metallic structure formed on a substrate. A nitride layer is deposited on at least the top surface of the metallic structure, and the metal oxide layer is deposited over the metallic structure and the nitride layer. The metal oxide layer is then polished by a chemical-mechanical polishing (CMP) process using a slurry, to expose the nitride layer on the top surface of the metallic structure. Polishing of the nitride layer causes ammonia to be generated in the slurry. The ammonia is extracted as a gas from the slurry, and a signal is generated in accordance with the ammonia concentration. The CMP process is terminated in accordance with a change in the signal. In a preferred embodiment, the metal oxide is aluminum oxide, the nitride is aluminum nitride, and the nitride layer is deposited as a conformal layer on the substrate and the metallic structure.

Abstract: The change in thickness of a film on an underlying body such as a semiconductor substrate is monitored in situ by inducing a current in the film, and as the thickness of the film changes (either increase or decrease), the changes in the current are detected. With a conductive film, eddy currents are induced in the film by a generating an alternating electromagnetic field with a sensor which includes a capacitor and an inductor.

Abstract: A contactless method and apparatus for real-time in-situ monitoring of a chemical etching process during etching of at least one wafer in a wet chemical etchant bath are disclosed. The method comprises the steps of providing two conductive electrodes in the wet chemical bath, wherein the two electrodes are proximate to but not in contact with a wafer; monitoring an electrical characteristic between the two electrodes as a function of time in the etchant bath of the at least one wafer, wherein a prescribed change in the electrical characteristic is indicative of a prescribed condition of the etching process; and recording a plurality of values of the electrical characteristic as a function of time during etching. From the plurality of recorded values and corresponding times, instantaneous etch rates, average etch rates, and etching end points may be determined. Such a method and the apparatus therefor are particularly useful in a wet chemical etch station.

Abstract: An apparatus for rotary signal coupling in in-situ monitoring of a chemical-mechanical polishing process by a polisher is provided with a sensor fixed to a rotatable wafer carrier for creating a signal responsive to the chemical mechanical polishing process, a conductor coupled to the sensor for receiving the signal, the conductor fixed to the rotatable wafer carrier, a contact coupled to the conductor, the contact fixed to a stationary drive arm, and signal transfer means coupled to the contact for transferring the signal to a monitoring means.

Abstract: The change in thickness of a film on an underlying body such as a semiconductor substrate is monitored in situ by inducing a current in the film, and as the thickness of the film changes (either increase or decrease), the changes in the current are detected. With a conductive film, eddy currents are induced in the film by a generating an alternating electromagnetic field with a sensor which includes a capacitor and an inductor.

Abstract: An apparatus in a chemical vapor deposition (CVD) system monitors the actual wafer/substrate temperature during the deposition process. The apparatus makes possible the production of high quality aluminum oxide films with real-time wafer/substrate control. An infrared (IR) temperature monitoring device is used to control the actual wafer temperature to the process temperature setpoint. This eliminates all atmospheric temperature probing. The need for test runs and monitor wafers as well as the resources required to perform the operations is eliminated and operating cost are reduced. High quality, uniform films of aluminum oxide can be deposited on a silicon substrates with no need for additional photolithographic steps to simulate conformality that are present in a sputtered (PVD) type application. The result is a reduction in required process steps with subsequent anticipated savings in equipment, cycle time, chemicals, reduce handling, and increased yield of devices on the substrate.

Abstract: Rotary signal coupling in in-situ monitoring of a chemical-mechanical polishing process. A sensor fixed to a rotatable wafer carrier for creating a signal responsive to the chemical mechanical polishing process is coupled to a bottom half of a rotary transformer fixed to a rotating portion of the polisher. A top half of the rotary transformer, coupled to the bottom half of the rotary transformer, is fixed to a stationary portion of the polisher. The signal from the sensor is thus coupled through the rotary transformer to a process monitor.