Previous conference records:

• 230925_3rd_DDS@Osaka
• 230926_3rd_DDS@Osaka

Conference Records 2024-01-19:

Interesting sessions

Session name: Mathematics

Details and Comments

• Infer trajectory and dynamics of single-cell by RNA-seq data
• Image and video could also be probability distribution: in the case of RNA-seq, trajectory could be viewed as single group of high-dimensional point cloud divide to two
• Cell populations should differentiate while forming sub-population → cluster-to-cluster OT → Cell population ~ Gaussian → entropic Gaussian mixture OT
• Cluster-to-cluster OT: Using entropic transport plan, construct entropic barycenter and barycentric projection map
• Need to know: Waddington's landscape: $-\mathrm{\Delta }\varphi =\mathbf{div}{v}_{cell}$
• Comments: this method could be used in eye-tracking data

Session name: Material Sci group

Details and Comments

• Sub Session 01 Evaluation of CFRP by TDA: X-CT images → pre-process of images → death of PD1: the threshold of void → birth of PD0: the center of void → scoring
• Sub Session 02 quantitative analysis of SEM images of Fe: TDA to find PD0 and get multiple structural features from a single persistence diagram
• Sub Session 03 imaging manifolds of mine: voids in mine lead to myths → experiments to figure out the features of voids in $NaAlS{i}_{2}O$ → cracking leads tree-like voids → how to describe and analyze the mechanism of manifolds leading voids
• Comments on session 1: how to describe the cracking under ultra-high-pressure?

Session name: Recruit Session 07

Details and Comments

• Cooperative, distributed localization yields high accuracy
• Q: Observation biases are not considered
• Observation model: $y_{i,j}(t)=r_{i,j}+b_i+e_{i,j}(t)$
• Algorithm: log-likelihood → augmented Lagrange → calculate the gradient
• Comments: high accuracy but heavy updating

Session name: Recruit Session 08

Details and Comments

• Concepts: Node → modules | Edges → interaction between modules
• Stimulating gene expression dynamics on LPS
• Single-cell level to organism level with SBM
• Comments: It might be useful to infer the interaction between mental states

Session name: Recruit Session 09

Details and Comments

• It is difficult to classify the physical properties of amorphous
• Plot PD to predict the distribution of $Si$ amorphous

Session name: Recruit Session 10

Details and Comments

• Functional GLM: on BMI and survival interval → identify weight trajectories related to mortality
• Comparison: mixed model vs. functional vs. joint model

Session name: Recruit Session 11

Details and Comments

• Application on Antibody design: it is important to recognize the structure of protein
• Using GAN to generate the image of protein might be useful
• Challenge: design protein to pump out virus

Session name: Recruit Session 13

Details and Comments

• Rhythm of biology matters
• Goodwin model could fit (describe) bio-rhythm
• Model-free analysis?

Session name: Recruit Session 14

Details and Comments

• Shape optimization problem → shape derivative
• Steepest descent method: set an initial → find some shape size $ϵ$ until convergence
• Need to know: shape flow

Session name: Recruit Session 15

Details and Comments

• Network of interacting rhythms → Full dynamical system → phase oscillator model
• Experiment: operate the frog's sound to test the interaction within natural states

Session name: Recruit Session 16

Details and Comments

• Tracking signal in cortex: mechanism of information path to function
• Need to know: Granger causality, LiNGAM (Linear non-G acyclic Model)
• Find critical brain area on spinal cord injury recovery

Session name: Recruit Session 17

Details and Comments

• Self-propelled particles: physical variable space → collective variable space
• Set up different parameters to create different patterns of herds: Snapshots → PD → Persistent Images → ML
• Comments: TRY this flow to our application and visualize it.

Daily Summary

What's new

• Interesting work flow suggested by Suetani (Recruit Session 17)
• Some new model or model-free methods
• General usage of PD

What should I check

References

Yachimura et al., 2023, scEGOT: Single-cell trajectory inference framework based on entropic Gaussian mixture optimal transport https://www.biorxiv.org/content/10.1101/2023.09.11.557102v1
Wang et al., 2023, Elastic Shape Analysis of Tree-like 3D Objects using Extended SRVF Representation https://arxiv.org/abs/2110.08693
Stolz et al., 2023, Relational persistent homology for multispecies data with application to the tumor microenvironment https://arxiv.org/abs/2308.06205