fix bookmarks

Author: sdgilley

includes/machine-learning-compare-classic-aml.md

@@ -13,8 +13,8 @@ The following table summarizes the key differences between ML Studio (classic) a
 | Drag and drop interface | Classic experience | Updated experience - [Azure Machine Learning designer](../articles/machine-learning/concept-designer.md)| 
 | Code SDKs | Not supported | Fully integrated with [Azure Machine Learning Python](/python/api/overview/azure/ml/) and [R](https://github.com/Azure/azureml-sdk-for-r) SDKs |
 | Experiment | Scalable (10-GB training data limit) | Scale with compute target |
-| Training compute targets | Proprietary compute target, CPU support only | Wide range of customizable [training compute targets](../articles/machine-learning/concept-compute-target.md#train). Includes GPU and CPU support | 
-| Deployment compute targets | Proprietary web service format, not customizable | Wide range of customizable [deployment compute targets](../articles/machine-learning/concept-compute-target.md#deploy). Includes GPU and CPU support |
+| Training compute targets | Proprietary compute target, CPU support only | Wide range of customizable [training compute targets](../articles/machine-learning/concept-compute-target.md#training-compute-targets). Includes GPU and CPU support | 
+| Deployment compute targets | Proprietary web service format, not customizable | Wide range of customizable [deployment compute targets](../articles/machine-learning/concept-compute-target.md#compute-targets-for-inference). Includes GPU and CPU support |
 | ML Pipeline | Not supported | Build flexible, modular [pipelines](../articles/machine-learning/concept-ml-pipelines.md) to automate workflows |
 | MLOps | Basic model management and deployment; CPU only deployments | Entity versioning (model, data, workflows), workflow automation, integration with CICD tooling, CPU and GPU deployments [and more](../articles/machine-learning/concept-model-management-and-deployment.md) |
 | Model format | Proprietary format, Studio (classic) only | Multiple supported formats depending on training job type |