AI4All Day 5: Data Science, Computational Biology, and a Thrilling First Week
Nidhi Parthasarathy, Friday, July 1st, 2022
Linear Algebra in ML
The two morning sessions on Day 5 focused on more cohort-level introduction to programming and data science. We started with learning about linear algebra, and about how vectors and matrices worked. We learned about visualization, and various matrix/vector concepts (dimensionality, operations). We talked about how matrices could represent images and form the basis for ML models.
We discussed images of rib cages and how vectors could be used to distinguish between normal and abnormal health conditions (putting features into matrices, and using matrix/vector operations like dot products to track directionality and similarity). We also learned about weight matrices to use during machine learning. We also did some code labs on numpy going deep into various exercises on matrices, etc.
Learning about Computational Biology in the Field!
The afternoon session featured a talk from Marina Sirota (Professor at UCSF) on computational biology. She started off by introducing herself and her journey in bridging her interests in biology/genetics and computation/CS. She discussed her research in comparative genetics and her work at Pfizer and bioinformatics at Stanford. She mentioned Watson/Crick’s “Central Dogma” on the structure of DNA and the concepts of transcription and translation. She discussed measuring genetic expression using microarrays (the small glass chips with thousands of probes that she mentioned sounded really cool). She went through all the steps involved in this process: RNA extraction, fluorescent labeling, hybridization, scanning and signal processing, and normalization and analysis.
She talked about the long lead time and challenges with research in her area (15 years and 800M to bring a drug to pharmacies; 90% of drugs fail in early development!) and how drug repurposing could help reduce costs and improve the process. She concluded by discussing clinical science and the use of data and visualization and computation in understanding patterns in science and medicine.
Many of the biology concepts were new to me and it was fascinating to understand all the interesting aspects of how the world of bio-computation works. I felt inspired to try doing some work in this space! It will be cool if I get a chance to do some research like Prof. Sirota in these areas!
Projects in AI
The next session featured four demos from Salesforce. The two that stood out to me were the Slack-AI demo and the CodeGen demo. The Slack-AI demo was about a new feature that hasn’t gone out yet and uses AI on Slack channels using NLP. The CodeGen demo showed how the user could talk to an AI in natural language and it would then write the code for them (“Can you write a program that sorts an array?” “Can you make it more efficient?”)
Virtual Social
We ended the week with another social session, this time focused on getting to know the people in our cohort more. We had a fun discussion on creating an AI-powered coffee house for us to chill out!
A Summary of the Week
All in all, a great first week. I can’t believe how much we have learned in one week. From learning about AI and its four subunits (computer vision, natural language processing, robotics, and computational biology) to deep-diving into medical AI and the intricacies of classification and machine learning in our cohort, this week has been very educational. At the same time, it has also been very inspiring with all the different talks and demos that we had about a variety of topics from vision and robotics to ethics and design. I am looking forward to next week to learn more new AI concepts and hear from other speakers! And, it has also been fun to learn about research. Maybe I can get to help out on some research project someday!
Continue to the next blog to see what happened in week 2.
