Abstract: In an embodiment, a computer stores, in memory or storage, many explanation profiles, many log entries, and definitions of many features that log entries contain. Some features may contain a logic statement such as a database query, and these are specially aggregated based on similarity. Based on the entity specified by an explanation profile, statistics are materialized for some or all features. Statistics calculation may be based on scheduled batches of log entries or a stream of live log entries. At runtime, an inference that is based on a new log entry is received. Based on an entity specified in the new log entry, a particular explanation profile is dynamically selected. Based on the new log entry and statistics of features for the selected explanation profile, a local explanation of the inference is generated. In an embodiment, an explanation text template is used to generate the local explanation.

Abstract: Techniques for propagating scores in subgraphs are provided. In one technique, multiple path scores are stored, each path score associated with a path (or subgraph), of multiple paths, in a graph of nodes. The path scores may be generated by a machine-learned model. For each path score, a path that is associated with that path score is identified and nodes of that path are identified. For each identified node, a node score for that node is determined or computed based on the corresponding path score and the node score is stored in association with that node. Subsequently, for each node in a subset of the graph, multiple node scores that are associated with that node are identified and aggregated to generate a propagated score for that node. In a related technique, a propagated score of a node is used to compute a score for each leaf node of the node.

Abstract: 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.

Abstract: Herein are machine learning (ML) explainability (MLX) techniques for calculating and using a novel fidelity metric for assessing and comparing explainers that are based on feature attribution. In an embodiment, a computer generates many anomalous tuples from many non-anomalous tuples. Each anomalous tuple contains a perturbed value of a respective perturbed feature. For each anomalous tuple, a respective explanation is generated that identifies a respective identified feature as a cause of the anomalous tuple being anomalous. A fidelity metric is calculated by counting correct explanations for the anomalous tuples whose identified feature is the perturbed feature. Tuples may represent entries in an activity log such as structured query language (SQL) statements in a console output log of a database server. This approach herein may gauge the quality of a set of MLX explanations for why log entries or network packets are characterized as anomalous by an intrusion detector or other anomaly detector.

Abstract: The present invention relates to machine learning (ML) explainability (MLX). Herein are techniques for a novel relevance propagation rule in layer-wise relevance propagation (LRP) for feature attribution-based explanation (ABX) for a reconstructive autoencoder. In an embodiment, a reconstruction layer of a reconstructive neural network in a computer generates a reconstructed tuple that is based on an original tuple that contains many features. A reconstruction residual cost function calculates a reconstruction error that measures a difference between the original tuple and the reconstructed tuple. Applied to the reconstruction error is a novel reconstruction relevance propagation rule that assigns a respective reconstruction relevance to each reconstruction neuron in the reconstruction layer. Based on the reconstruction relevance of the reconstruction neurons, a respective feature relevance of each feature is determined, from which an ABX explanation may be automatically generated.

Abstract: Herein are machine learning (ML) explainability (MLX) techniques that perturb a non-anomalous tuple to generate an anomalous tuple as adversarial input to any explainer that is based on feature attribution. In an embodiment, a computer generates, from a non-anomalous tuple, an anomalous tuple that contains a perturbed value of a perturbed feature. In the anomalous tuple, the perturbed value of the perturbed feature is modified to cause a change in reconstruction error for the anomalous tuple. The change in reconstruction error includes a decrease in reconstruction error of the perturbed feature and/or an increase in a sum of reconstruction error of all features that are not the perturbed feature. After modifying the perturbed value, an attribution-based explainer automatically generates an explanation that identifies an identified feature as a cause of the anomalous tuple being anomalous. Whether the identified feature of the explanation is or is not the perturbed feature is detected.

Abstract: 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.

Abstract: The present invention avoids overfitting in deep neural network (DNN) training by using multitask learning (MTL) and self-supervised learning (SSL) techniques when training a multi-branch DNN to encode a sequence. In an embodiment, a computer first trains the DNN to perform a first task. The DNN contains: a first encoder in a first branch, a second encoder in a second branch, and an interpreter layer that combines data from the first branch and the second branch. The DNN second trains to perform a second task. After the first and second trainings, production encoding and inferencing occur. The first encoder encodes a sparse feature vector into a dense feature vector from which an inference is inferred. In an embodiment, a sequence of log messages is encoded into an encoded trace. An anomaly detector infers whether the sequence is anomalous. In an embodiment, the log messages are database commands.

Abstract: The present invention relates to threshold estimation and calibration for anomaly detection. Herein are machine learning (ML) and extreme value theory (EVT) techniques for normalizing and thresholding anomaly scores without presuming a values distribution. In an embodiment, a computer receives many unnormalized anomaly scores and, according to peak over threshold (POT), selects a highest subset of the unnormalized anomaly scores that exceed a tail threshold. Based on the highest subset of the unnormalized anomaly scores, parameters of a probability density function are trained according to EVT. After training and in a production environment, a normalized anomaly score is generated based on an unnormalized anomaly score and the trained parameters of the probability density function. Anomaly detection compares the normalized anomaly score to an optimized anomaly threshold.

Abstract: In an embodiment, a computer hosts a machine learning (ML) model that infers a particular inference for a particular tuple that is based on many features. The features are grouped into predefined super-features that each contain a disjoint (i.e. nonintersecting, mutually exclusive) subset of features. For each super-feature, the computer: a) randomly selects many permuted values from original values of the super-feature in original tuples, b) generates permuted tuples that are based on the particular tuple and a respective permuted value, and c) causes the ML model to infer a respective permuted inference for each permuted tuple. A surrogate model is trained based on the permuted inferences. For each super-feature, a respective importance of the super-feature is calculated based on the surrogate model. Super-feature importances may be used to rank super-features by influence and/or generate a local ML explainability (MLX) explanation.

Abstract: 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.

Abstract: 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.