[ML] Regression Rescorer

[davidkyle]

Adds a rescorer utilising the regression models currently used in the Inference Processor, in fact the configuration shares many similarities.

GET interesting_index/_search
{
  "query": { ...  },
  "rescore": {
    "ml_rescore": {
      "model_id": "mustard",
      "field_mappings": {},
      "inference_config": { "regression": {} }
    },
  }
}

The rescorer loads the specified model (model_id) during rewrite then for each document extracts the fields required by the model and rescores the hit according to the regression result. The final score is computed from the search score and model score using the same score_mode operations as query rescore.

GET interesting_index/_search
{
  "query": { ...  },
  "rescore": {
    "ml_rescore": {
      "model_id": "mustard",
      "field_mappings": {},
      "inference_config": { "regression": {} },
      "model_weight": 2.0,
      "query_weight": 1.0,
      "score_mode": "total" 
    },
  }
}

As with all prototypes most of the work is left todo:

• It's assumed the fields doc value doubles. This may not be the case if preprocessors are used
• Loaded models are not cached
• Validation and error checking is largely missing
• Implement the explain API using feature importance
• documentation

[elasticmachine]

Pinging @elastic/ml-core (:ml)

[davidkyle]

retest this please

[benwtrent]

I like the simple implementation. Seems like a natural place to put inference support :)

[benwtrent]

Does a re-scorer run only on the coordinating node or down on the shards?

[benwtrent]

If it is down on the shards, It might be good to not inflate the definition until it is used, that way the query definition is smaller for serializing across the wire.

I am not sure the computation cost of inflating it once and serializing it vs inflating it X number of times on each shard.

[davidkyle]

👍 The rescorer only runs at the shard. The coordinating node receives the scores from each shard and is responsible for choosing which documents to return but doesn't actually do any rescoring

[benwtrent]

Caching is another consideration. I am not sure if we can cache the inflated object on the shard so that we know not to inflate it and just use the cached object.

The cost of writing the inflated model on the wire might not be THAT bad if you take into consideration all nodes having to gunzip + parse the JSON every time.

[joshdevins]

Loaded models are not cached

As in, on each rescore (so once per shard per query), we load the model again?

[davidkyle]

As in, on each rescore (so once per shard per query), we load the model again?

Well there is always room for improvement. The caching framework used by the inference ingest processor can be utilised and should be fairly easy to incorporate.

[benwtrent]

As in, on each rescore (so once per shard per query), we load the model again?

It is loaded once per query. Not once per shard since @davidkyle is loading it up in the coordinating node.

[joshdevins]

It is loaded once per query. Not once per shard since @davidkyle is loading it up in the coordinating node.

Right, and then it's serialized and sent to each shard?

I'll wait to hear back from the Elasticsearch team about caching. As a PoC, this seems totally fine to me, and beyond that we can iterate on caching strategies.

[davidkyle]

run elasticsearch-ci/2

[joshdevins]

run elasticsearch-ci/2

[davidkyle]

run elasticsearch-ci/2