09:37:56 From John Healy To Everyone:
	mutual reachability distance
09:38:03 From Trevor Karn To Everyone:
	Thanks for a great talk! Is there any way to use the difference between Lan/Ran to measure uncertainty for classification?
09:47:57 From Ilir Dema To Everyone:
	sorry, can you let me in organizers
09:49:20 From Steve Reagan, USA, Mi To Everyone:
	Thank you. Where would you suggest getting further info and examples on Kan extensions for data science, particularly predictive models?
11:10:30 From Ilir Dema To Everyone:
	Any examples of kernel density functions you have experimented with?
11:17:28 From Rolando Kindelan To Everyone:
	Have you faced scalability problems?? if yes, how to deal with them?
11:19:23 From Rolando Kindelan To Everyone:
	Ohh I see, It should be the RIVET job
13:56:55 From John Healy To Everyone:
	https://bit.ly/3MiR2cH
14:02:40 From John Healy To Everyone:
	Leland’s notebook also includes code for downloading these various datasets.
14:22:35 From Rolando Kindelan To Everyone:
	you can select a subcomplex and use the link operation to get neighborhoods instead of fixing a k value
14:23:25 From Davi Sidarta To Everyone:
	the number of neighbors should be small enough in relation with the sampling number so that we can approximate neighborhoods as euclidean, do they not?
14:41:07 From Rolando Kindelan To Everyone:
	https://arxiv.org/abs/2111.05214
14:41:13 From Luis Scoccola To Everyone:
	Elder-Rule-Staircodes for Augmented Metric Spaces: https://arxiv.org/abs/2003.04523
15:16:17 From Davi Sidarta To Everyone:
	I think UMAP handles this by using a multi-component spectral layout initialization by default
15:32:12 From Davi Sidarta To Everyone:
	Do you have anything we can read about you work, Luis? Very cool!
15:37:53 From Luis Scoccola To Everyone:
	luis.scoccola@gmail.com
15:43:02 From Leland McInnes To Everyone:
	PacMAP
15:43:43 From Luis Scoccola To Everyone:
	https://jmlr.org/papers/volume22/20-1061/20-1061.pdf
15:48:11 From Ilir Dema To Everyone:
	Thank you