Oracle International Corporation

Telenav Gmbh

Oracle International Corpration

Nipun Agarwal

Sam Idicula

Ravi Murthy

Eric Sedlar

Sivasankaran Chandrasekar

Thomas Baby

Namit Jain

Sabina Petride

Seema Sundara

Asha Tarachandani

Venkatanathan Varadarajan

Sandeep Agrawal

Felix Schmidt

Bhushan Khaladkar

Onur Kocberber

Balasubramanyam Sthanikam

Ashish Thusoo

Jagannathan Srinivasan

Pit Fender

Benjamin Schlegel

Tomas Karnagel

Jian Wen

Ning Zhang

Sanjay Jinturkar

Syam Sundar Pannala

Rajkumar Sen

Nitin Kunal

Abhyudaya Agrawal

Vikram Kapoor

Karoon Rashedi Nia

Farhan Tauheed

Tayler Hetherington

Viswanathan Krishnamurthy

Arun Raghavan

Cagri Balkesen

Zahra Zohrevand

Vijay Medi

Mayur Bency

Andrew Brownsword

Craig Schelp

Sreedhar Mukkamalla

Fei Ge

Anh-Tuan Tran

Boris Klots

Stuart Wray

Hesam Fathi Moghadam

Anatoly Yakovlev

Aarti Basant

Hamed Ahmadi

Vikas Aggarwal

Sanket Malde

Mohammed Irfan Rafiq

Nitin Kumar Gupta

Muralidhar Krishnaprasad

Dinesh Das

Thuvan Hoang

Matteo Casserini

Yasha Pushak

Arno Schneuwly

Negar Koochakzadeh

Ying Hu

Thomas Chang

Matthias Brantner

Thomas F Keefe

Vikram Reddy Pesati

Anil Nori

Adiel Yoaz

Anantha Kiran Kandukuri

Pravin Shinde

Gong Zhang

Krishna Kantikiran Pasupuleti

Saeid Allahdadian

Shrikumar Hariharasubrahmanian

Neema Jalali

Guang-Tong Zhou

Andrea Di Blas

Amin Suzani

Manel Fernandez Gomez

Georgios Giannikis

Milos Vasic

Subramanian Muralidhar

Zhen Hua Liu

Vikas Arora

James W Warner

Chon Hei Lei

Alan Paul Wood

Susan M Kotsovolos

Anand Manikutty

Hui Zhang

Tanvir Ahmed

Rohan Angrish

Suratna Budalakoti

Paul A Reilly

Tim Wing Yu

Kevin Moore

Brian Gold

Ashwini Surpur

Renata Khasanova

James Stenoish

Juan Fernandez Peinador

Herein are techniques for topic modeling and content perturbation that provide machine learning (ML) explainability (MLX) for natural language processing (NLP). A computer hosts an ML model that infers an original inference for each of many text documents that contain many distinct terms. To each text document (TD) is assigned, based on terms in the TD, a topic that contains a subset of the distinct terms. In a perturbed copy of each TD, a perturbed subset of the distinct terms is replaced. For the perturbed copy of each TD, the ML model infers a perturbed inference. For TDs of a topic, the computer detects that a difference between original inferences of the TDs of the topic and perturbed inferences of the TDs of the topic exceeds a threshold. Based on terms in the TDs of the topic, the topic is replaced with multiple, finer-grained new topics. After sufficient topic modeling, a regional explanation of the ML model is generated.

Efficient and accurate regional explanation technique for NLP models

End-to-end explanation techniques, which efficiently explain the behavior (feature importance) of any machine learning model on large tabular datasets, are disclosed. These techniques comprise two down-sampling methods to efficiently select a small set of representative samples of a high-dimensional dataset for explaining a machine learning model by making use of the characteristics of the dataset or of an explainer of a machine learning model to optimize the explanation quality. These techniques significantly improve the explanation speed while maintaining the explanation quality of a full dataset evaluation.

Efficient and scalable computation of global feature importance explanations

A systematic explainer is described herein, which comprises local, model-agnostic, surrogate ML model-based explanation techniques that faithfully explain predictions from any machine learning classifier or regressor. The systematic explainer systematically generates local data samples around a given target data sample, which improves on exhaustive or random data sample generation algorithms. Specifically, using principles of locality and approximation of local decision boundaries, techniques described herein identify a hypersphere (or data sample neighborhood) over which to train the surrogate ML model such that the surrogate ML model produces valuable, high-quality information explaining data samples in the neighborhood of the target data sample. Combining this systematic local data sample generation and a supervised neighborhood selection approach to weighting generated data samples relative to the target data sample achieves high explanation fidelity, locality, and repeatability when generating explanations for specific predictions from a given model.

Systematic approach for explaining machine learning predictions

Auto-parallel-load techniques are provided for automatically loading database objects from an on-disk database system into an in-memory database system. The auto-parallel-load techniques involve a pipeline that includes several components. In one implementation, each of the pipeline components is configured to receive, extract information from, and add information to, a 'state object'. One or more of the pipeline components include logic that is based on the output of a corresponding machine learning model. The machine learning models used by the pipeline components may be trained from training sets from which outliers have been excluded, and may be used as the basis for generating linear models that are used during runtime, to produce estimates that affect the parameters of the auto-parallel-load operation.

Automating data load operations for in-memory data warehouses

Embodiments implement a prediction-driven, rather than a trial-driven, approach to automatic data placement recommendations for partitioning data across multiple nodes in a database system. The system is configured to extract workload-specific features of a database workload running at a database system and dataset-specific features of a database running on the database system. The workload-specific features characterize utilization of the database workload. The dataset-specific features characterize how data is organized within the database. The system identifies a plurality of candidate keys for determining how to partition data stored in the database across nodes. Based at least in part on the workload-specific features, the dataset specific features, and the plurality of candidate keys, a set of candidate key combinations for partitioning data is generated. Using a machine learning model, determine a particular candidate key combination that optimizes query execution performance benefit based on the workload-specific features and the dataset specific features. Generate data placement commands to allocate the database tables across the nodes.

Workload-aware data placement advisor for OLAP database systems

In an embodiment, a computer-implemented method includes receiving a query from a client and determining a query plan for the query. The query plan comprises one or more query operators for executing at least a portion of the query on a database. The method also includes sending the one or more query operators to one or more computing nodes for the one or more computing nodes to execute the one or more query operators on one or more data fragments of the database. In this example, each computing node of the one or more computing nodes hosts a respective data fragment of the one or more data fragments. Further, the method includes detecting an error in executing a first query operator by a first computing node on a first data fragment, and sending, in response to detecting the error, the first query operator to a replacement computing node for executing on the first data fragment hosted by the spare computing node.

Fault tolerance in scale-out distributed query processing appliance

Techniques are described herein for performing efficient matrix multiplication in architectures with scratchpad memories or associative caches using asymmetric allocation of space for the different matrices. The system receives a left matrix and a right matrix. In an embodiment, the system allocates, in a scratchpad memory, asymmetric memory space for tiles for each of the two matrices as well as a dot product matrix. The system proceeds with then performing dot product matrix multiplication involving the tiles of the left and the right matrices, storing resulting dot product values in corresponding allocated dot product matrix tiles. The system then proceeds to write the stored dot product values from the scratchpad memory into main memory.

Asymmetric allocation of SRAM and data layout for efficient matrix-matrix multiplication

Herein is resource-constrained feature enrichment for analysis of parse trees such as suspicious database queries. In an embodiment, a computer receives a parse tree that contains many tree nodes. Each tree node is associated with a respective production rule that was used to generate the tree node. Extracted from the parse tree are many sequences of production rules having respective sequence lengths that satisfy a length constraint that accepts at least one fixed length that is greater than two. Each extracted sequence of production rules consists of respective production rules of a sequence of tree nodes in a respective directed tree path of the parse tree having a path length that satisfies that same length constraint. Based on the extracted sequences of production rules, a machine learning model generates an inference. In a bag of rules data structure, the extracted sequences of production rules are aggregated by distinct sequence and duplicates are counted.

Generalized production rules—N-gram feature extraction from abstract syntax trees (AST) for code vectorization

Techniques are provided for sparse ensembling of unsupervised machine learning models. In an embodiment, the proposed architecture is composed of multiple unsupervised machine learning models that each produce a score as output and a gating network that analyzes the inputs and outputs of the unsupervised machine learning models to select an optimal ensemble of unsupervised machine learning models. The gating network is trained to choose a minimal number of the multiple unsupervised machine learning models whose scores are combined to create a final score that matches or closely resembles a final score that is computed using all the scores of the multiple unsupervised machine learning models.

Sparse ensembling of unsupervised models

Herein are approaches for self-optimization of a database management system (DBMS) such as in real time. Adaptive just-in-time sampling techniques herein estimate database content statistics that a machine learning (ML) model may use to predict configuration settings that conserve computer resources such as execution time and storage space. In an embodiment, a computer repeatedly samples database content until a dynamic convergence criterion is satisfied. In each iteration of a series of sampling iterations, a subset of rows of a database table are sampled, and estimates of content statistics of the database table are adjusted based on the sampled subset of rows. Immediately or eventually after detecting dynamic convergence, a machine learning (ML) model predicts, based on the content statistic estimates, an optimal value for a configuration setting of the DBMS.

Enabling efficient machine learning model inference using adaptive sampling for autonomous database services

Herein, a computer generates and evaluates many preprocessor configurations for a window preprocessor that transforms a training timeseries dataset for an ML model. With each preprocessor configuration, the window preprocessor is configured. The window preprocessor then converts the training timeseries dataset into a configuration-specific point-based dataset that is based on the preprocessor configuration. The ML model is trained based on the configuration-specific point-based dataset to calculate a score for the preprocessor configuration. Based on the scores of the many preprocessor configurations, an optimal preprocessor configuration is selected for finally configuring the window preprocessor, after which, the window preprocessor can optimally transform a new timeseries dataset such as in an offline or online production environment such as for real-time processing of a live streaming timeseries.

Automated machine learning pipeline for timeseries datasets utilizing point-based algorithms

Techniques are described for executing machine learning models trained for specific operators with feature values that are based on the actual execution of a workload set. The machine learning models generate an estimate of benefit gain/cost for executing operations on data portions in the alternative encoding format. Such data potions may be sorted based on the estimated benefit, in an embodiment. Using cost estimation machine learning models for memory space, the data portions with the most benefits that comply with the existing memory space constraints are recommended and/or are automatically encoded into the alternative encoding format.

Workload-aware data encoding

Techniques are described herein for prediction of an buffer pool size (BPS). Before performing BPS prediction, gathered data are used to determine whether a target workload is in a steady state. Historical utilization data gathered while the workload is in a steady state are used to predict object-specific BPS components for database objects, accessed by the target workload, that are identified for BPS analysis based on shares of the total disk I/O requests, for the workload, that are attributed to the respective objects. Preference of analysis is given to objects that are associated with larger shares of disk I/O activity. An object-specific BPS component is determined based on a coverage function that returns a percentage of the database object size (on disk) that should be available in the buffer pool for that database object. The percentage is determined using either a heuristic-based or a machine learning-based approach.

Prediction of buffer pool size for transaction processing workloads

Herein are techniques that train regressor(s) to predict how effective would a machine learning model (MLM) be if trained with new hyperparameters and/or dataset. In an embodiment, for each training dataset, a computer derives, from the dataset, values for dataset metafeatures. The computer performs, for each hyperparameters configuration (HC) of a MLM, including landmark HCs: configuring the MLM based on the HC, training the MLM based on the dataset, and obtaining an empirical quality score that indicates how effective was said training the MLM when configured with the HC. A performance tuple is generated that contains: the HC, the values for the dataset metafeatures, the empirical quality score and, for each landmark configuration, the empirical quality score of the landmark configuration and/or the landmark configuration itself. Based on the performance tuples, a regressor is trained to predict an estimated quality score based on a given dataset and a given HC.

Using metamodeling for fast and accurate hyperparameter optimization of machine learning and deep learning models

Techniques described herein propose a ROWID Elimination Rewrite that uses functional dependencies to substitute and/or eliminate ROWID pseudo-columns referenced in a query in order to reduce memory pressure and speed up processing.

ROWID elimination rewrite

Growing community of inventors

Santa Clara, CA, United States of America

Nipun Agarwal