> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/sdgm/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/sdgm/_mcp/server.

# Does Kumo Engine support multitask or multi-objective learning?

> Distinguish joint multitask training from multilabel classification and weighted training

Kumo Engine does not currently natively support jointly training one model across distinct prediction targets with a combined loss. Depending on your objective, these approaches may address your use case:

* **Multilabel classification:** Predict several labels for an entity within one prediction task, such as different fraud categories. See the [Multilabel Classification example](/kumo-engine/user-code-guide/code-examples/linux-deployment/multilabel-classification).
* **Weighted training:** Combine behavioral signals into a unified training table and assign row weights to control their contribution, such as weighting different recommendation interactions. See [Weighting and Class Imbalance](/kumo-engine/user-code-guide/train-a-model/model-plan-guide#use-row-weights-and-balance-binary-classes) for supported tasks and settings.