IBM Watson™ Discovery Service Ideas

It would be nice if also passages can show the relevant keywords for the search. Normally in the UI I would just show the passages and would be nice to show where the search result come from

It would be handy to be able to delete documents from a collection by defining conditions. Something I would do in a sql-database like this: DELETE FROM collection_name WHERE publication_date < date('2018-01-01')

Our project would quite certainly benefit from using the bias parameter in querys: rather that using a hard filter based on some metadata value, we would like to weigh some metadata values more in certain cases. However, our metadata values are mo...

We have a JSON file containing all the needed natural language queries and examples for training WDS collection. It would be helpful if there is a way to upload these bulk of queries at the same time to WDS.

When iterating in Discovery, especially when experimenting with multiple and more complex queries it would be very useful to be able to save my query once I have it in suitable shape so I can iterate on another one without losing my work.

Annotation of semantic meaning of text within tables utilizing headers for context

Ability to connect OWL ontology to Discovery JSONs This to imply an instance of an ontology / knowledge graph where enrichments from WDS would be associated with nodes of an ontology and returned from an WDS query? Currently WDS only works with e...

Currently, FAQ documents (documents that are already structured in question and answer sections in pdf, word,) need to be pre processed into QnA json units and then imported into discovery. It would be better to have inbuilt feature to understand ...

The official Discovery API currently only supports document submission via multipart/form-dataPOST operations. I would like to post data directly in the ingestion request: curl -u "{username}":"{password}" \ -H "Content-Type: application/json"...

I have a use-case around call centers, it would be great if Watson Speech to Text was integrated as part of the Discovery pipeline