Oracle International Corporation

Renata Khasanova

Felix Schmidt

Matteo Casserini

Aneesh Dahiya

Nipun Agarwal

Stuart Wray

Craig Schelp

Arno Schneuwly

Kenyu Kobayashi

A computer-implemented method includes receiving an input for a deterministic model, and determining whether the input and an output from the model corresponding to the input are stored in a cache. The method further includes, in accordance with the input and the output not being stored in the cache, computing the output from the model based on the input, storing the input and the output as a cache element, and storing raw scores respectively indicating a usage frequency and order of usage of the element, and a computation time for computing the output. The method further includes calculating a caching score for the element; the caching score includes a sum of a normalized scores each corresponding to a product of a raw score and a scaling factor.

Caching strategy based on model execution time, frequency and input order with configurable priority

A computer-implemented method includes receiving a first input for a deterministic model and generating, in accordance with the model, a first output corresponding to the first input; the generating is performed in a first computation time. The method also includes storing the first input and the first output as a first input-output pair in a cache; the first input-output pair has a priority in the cache according to the first computation time. The method further includes subsequently receiving a second input for the model that is a duplicate of the first input; and generating a second output corresponding to the second input by retrieving the first output from the cache.

Caching strategy based on model execution time

Herein is a machine learning (ML) explainability (MLX) approach in which a natural language explanation is generated based on analysis of a parse tree such as for a suspicious database query or web browser JavaScript. In an embodiment, a computer selects, based on a respective relevance score for each non-leaf node in a parse tree of a statement, a relevant subset of non-leaf nodes. The non-leaf nodes are grouped in the parse tree into groups that represent respective portions of the statement. Based on a relevant subset of the groups that contain at least one non-leaf node in the relevant subset of non-leaf nodes, a natural language explanation of why the statement is anomalous is generated.

Textual explanations for abstract syntax trees with scored nodes

Techniques are described herein for using machine learning to learn vector representations of DNS requests such that the resulting embeddings represent the semantics of the DNS requests as a whole. Techniques described herein perform pre-processing of tokenized DNS request strings in which hashes, which are long and relatively random strings of characters, are detected in DNS request strings and each detected hash token is replaced with a placeholder token. A vectorizing ML model is trained using the pre-processed training dataset in which hash tokens have been replaced. Embeddings for the DNS tokens are derived from an intermediate layer of the vectorizing ML model. The encoding application creates final vector representations for each DNS request string by generating a weighted summation of the embeddings of all of the tokens in the DNS request string. Because of hash replacement, the resulting DNS request embeddings reflect semantics of the hashes as a group.

Growing community of inventors

Zurich, Switzerland

Renata Khasanova