Machine Learning Jobs
Fusion AI provides these job types to perform machine learning tasks.
Signals analysis
These jobs analyze a collection of signals in order to perform query rewriting, signals aggregation, or experiment analysis.
-
Estimate ground truth queries using click signals and query signals, with document relevance per query determined using a click/skip formula.
Query rewriting
These jobs produce data that can be used for query rewriting or to inform updates to the synonyms.txt file.
-
Perform head/tail analysis of queries from collections of raw or aggregated signals, to identify underperforming queries and the reasons. This information is valuable for improving overall conversions, Solr configurations, auto-suggest, product catalogs, and SEO/SEM strategies, in order to improve conversion rates.
-
Identify multi-word phrases in signals.
-
Token and Phrase Spell Correction
Detect misspellings in queries or documents using the numbers of occurrences of words and phrases.
Signals aggregation
-
A Spark SQL aggregation job where user-defined parameters are injected into a built-in SQL template at runtime.
Experiment analysis
-
Calculate relevance metrics (nDCG and so on) by replaying ground truth queries against catalog data using variants from an experiment.
-
SQL-Based Experiment Metric (deprecated)
This job is created by an experiment in order to calculate an objective.
SQL-Based Experiment Metric job is deprecated as of Fusion AI 4.0.2.
Collaborative recommenders
These jobs analyze signals and generate matrices used to provide collaborative recommendations.
-
Use this job when you want to compute user recommendations or item similarities using a collaborative filtering recommender. You can also implement a user-to-item recommender in the advanced section of this job’s configuration UI. This job uses SparkML’s Alternating Least Squares (ALS).
-
Train a collaborative filtering matrix decomposition recommender using SparkML’s Alternating Least Squares (ALS) to batch-compute query-query similarities. This can be used for items-for-query recommendations as well as queries-for-query recommendations.
Content-based recommenders
Content-based recommenders create matrices of similar items based on their content.
Content analysis
-
Use this job when you already have clusters or well-defined document categories, and you want to discover and attach keywords to see representative words within those existing clusters. (If you want to create new clusters, use the Document Clustering job.)
-
Use this job when you want to compute basic metrics about your collection, like average word length, phrase percentages, and outlier documents (with very many or very few documents).
-
Cluster a set of documents and attach cluster labels.
-
Logistic Regression Classifier Training
Train a regularized logistic regression model for text classification.
-
Use this job when you want to find outliers from a set of documents and attach labels for each outlier group.
-
Random Forest Classifier Training (deprecated)
Train a random forest classifier for text classification.
-
Word2Vec Model Training (Deprecated)
Train a shallow neural model, and project each document onto this vector embedding space.
Data ingest
-
The Parallel Bulk Loader (PBL) job enables bulk ingestion of structured and semi-structured data from big data systems, NoSQL databases, and common file formats like Parquet and Avro.
Legacy machine learning jobs
-
Compute user recommendations based on a pre-computed item similarity model.
-
Use this job when you only want to compute item-to-item similarities. This method is more lightweight than the generic Recommendations job.
Legacy Item Similarity job is deprecated as of Fusion AI 4.1.0. Use the ALS recommender job instead.