## Changelog since v1.15.3

### Other notable changes

* Restores compatibility with kubectl client-side validation when structural schemas are present in a CRD which sets spec.preserveUnknownFields=true ([#82666](https://github.com/kubernetes/kubernetes/pull/82666), [@liggitt](https://github.com/liggitt))

- Kubectl: restores `--dry-run`, `--dry-run=true`, and `--dry-run=false` for compatibility with pre-1.23 invocations. ([#107021](https://github.com/kubernetes/kubernetes/pull/107021), [@liggitt](https://github.com/liggitt)) [SIG CLI and Testing]

- Kubectl: restores `--dry-run`, `--dry-run=true`, and `--dry-run=false` for compatibility with pre-1.23 invocations. ([#107003](https://github.com/kubernetes/kubernetes/pull/107003), [@julianvmodesto](https://github.com/julianvmodesto)) [SIG CLI and Testing]

        Specifically, for the type of interruption with advance notice, it
        automatically saves a checkpoint, exits the program without raising an
        unrecoverable error, and restores the progress when training restarts.

*   `tf.math`:

    *   Added `tf.math.approx_max_k` and `tf.math.approx_min_k` which are the
        optimized alternatives to `tf.math.top_k` on TPU. The performance