Obsidian Source: Notes / Probabilistic ML for MM Data
Summary
Pending synthesis from local Obsidian source.
Original source title: Sample sensor readings with two overlapping patterns
Extracted Preview
Aalto Research Program - Overview
Need to prepare a motivation letter to let them know about the work that is there.
There are two branches where they want to work - one is ==uncertainty quantification== in large scale vision/language models and ==combining semantic understanding== with scene reconstruction.
Broader initiative in multi-modal modelling. Tractable modelling, Uncertainty Quantification and MM modelling.
Transformers are go-to, so highlight that you have worked with them before.
Probabilistic modelling. Keen interest in ==Language Models and Vision==.
The idea is not to show that you know too much, just that you are capable enough to take this work and can do it justice.
- Highlighting specific skills and interests in application which aligns with their needs. ==Enthusiasm should be there.
Probabilistic models are mathematical descriptions of how likely different patterns are given in the "evidence" you have. Examples include weather forecasting, spam filtering.
Integration Notes
- Source folder:
/home/yashs/Documents/Docs/Obsidian/Research-Notes - Local source:
/home/yashs/Documents/Docs/Obsidian/Research-Notes/Notes/Probabilistic ML for MM Data.md - Raw copy:
raw/obsidian/research-notes/Notes/Probabilistic ML for MM Data.md