Differences

This shows you the differences between two versions of the page.

--- atmos:researchmeetingnotes:researchmeetingnotes:home:christiannairy [2025/06/11 18:31] – delene
+++ atmos:researchmeetingnotes:researchmeetingnotes:home:christiannairy [2025/08/07 14:47] (current) – delene
@@ Line 726: / Line 726: @@
   * Add Neural Network.
+===== July 15, 2025 =====
+  Attending: David Delene, Christian Nairy, Andrew Detwiler.
+==== Discussion ====
+  * Talked about the UK Meeting.  Talked about Navy project and paper.
+  * XGBoots - When going to another flight, go away down.  Random sample generation. Had to make more images up, parameters up. Need a range values. The generated data was not help, it was hurting.  No doing too well to get chain-ag, good as non-chain. Data in balance is an issue. Overlaps.
+  * Only have 1,050 chains.  Can do things to the images to create more images.
+  * Talk about CNN. Works well on Unseen Data. Look at false positives.  Some single crystal.
+  * Get 256 feature parameters.  Like stuff that goes into XGBoots, Get lots of different parameters. The 256 does pretty well.
+  * Get large dataset to got XGBoots. ConvXGB. Get much better.
+  * The idea is to have the same number of different types of particles.
+  * Composition time.
+  * Suggest - See what for area of chains identified.
+  * JGL - Paper, Talk about electric field separating charges.
+==== Items ====
+  * What is the overall lesson? What does this mean for all crystals?  What about the science?
+  * Science, we want to identify area of lot of chain-ags.
+===== August 06, 2025 =====
+  Attending: David Delene, Christian Nairy
+==== Discussion ====
+  * Saudi Paper
+  * Setting up for the Chamber Experiment.
+  * Work on White-paper.
+  * Visa - Application, UK.