# Meeting notes from 9/29/2022

**URL:** <https://discuss.ai.google.dev/t/meeting-notes-from-9-29-2022/26481>\
**Category:** SIG Federated\
**Created:** [October 13, 2022, 7:09pm UTC](https://discuss.ai.google.dev/t/meeting-notes-from-9-29-2022/26481 "2022-10-13T19:09:44Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![krzys\_ostrowski](https://d1dlmcr85iqnpo.cloudfront.net/user_avatar/discuss.ai.google.dev/krzys_ostrowski/32/8649_2.png) [@krzys\_ostrowski](https://discuss.ai.google.dev/u/krzys_ostrowski)\
**Post date:** [October 13, 2022, 7:09pm UTC](https://discuss.ai.google.dev/t/meeting-notes-from-9-29-2022/26481/1 "2022-10-13T19:09:44Z")

</div>

We welcomed [OpenMined](https://www.openmined.org/) as a new member of the SIG, introduced our [collaboration on PySyTFF](https://blog.tensorflow.org/2022/09/bridging-communities-tensorflow-federated-tff-and-openmined.html), and discussed commonalities with platforms under development by others - in particular, the relevance of the logical partitioning of data to cross-silo scenarios (where individual silos may be hosting data that’s naturally partitioned by some notion of a “user”). We also continued the discussion of LinkedIn’s performance benchmarks and the need for overlapping data ingestion and training. Following the meeting, we open-sourced [components](https://github.com/tensorflow/federated/blob/main/tensorflow_federated/python/program/prefetching_data_source.py) to support this, and continued the discussion [on Discord](https://discord.com/channels/991185649401081876/992618367120851054). Notes are [posted on GitHub](https://github.com/tensorflow/federated/blob/main/docs/collaborations/notes/2022-09-29.md). Thanks!
