November 2026
M	T	W	T	F	S	S
	1
2	3	4	5	6	7	8
9	10	11	12	13	14	15
16	17	18	19	20	21	22
23	24	25	26	27	28	29
30

Sheharyar Raza: Managing Randomness in Artificial Intelligence for Transfusion Medicine

Sheharyar Raza/LinkedIn;Ruchika Goel/LinkedIn

Sep 15, 2026, 15:17

Sheharyar Raza: Managing Randomness in Artificial Intelligence for Transfusion Medicine

Sheharyar Raza, Transfusion Medicine Specialist at Sunnybrook Health Sciences Center and Internal Medicine Physician at Unity Health Toronto, shared a post on LinkedIn about a recent article he and Ruchika Goel co-authored, published in Vox Sanguinis, adding:

“Brief response to a letter published about our recent article by Ruchika Goel and I.

Our core point is that the distinction between newer generative models and ‘traditional’ machine learning models isn’t that one is stochastic and one is deterministic.

Depending on your specific use and pipeline, a language model architecture can be effectively deterministic (e.g. embedding models), while a traditional ML model can behave highly non-deterministically (e.g. clustering with a non-convex objective or bootstrap-based methods with small n).

Methods for evaluating the two are also roughly analogous, even if they appear superficially different.

What matters is to look at steps where, and how much, randomness comes into your model, and how it is being handled.

Thanks Aman for helpful discussion.”

Title: Response to comment on ‘Harnessing big data and artificial intelligence in transfusion medicine: Opportunities for precision, safety and efficiency’

Authors: Sheharyar Raza, Ruchika Goel

Sheharyar Raza: Managing Randomness in Artificial Intelligence for Transfusion Medicine