Open-sourcing Twitter’s algorithm isn’t what most people think it is. I don’t think even Elon Musk or most people at Twitter really understand where this process goes.
1/10 #DataScience#MachineLearning#Twitter
The code is not very insightful. The model itself is too complex for people to understand and interact with. So, what does open-sourcing the algorithm look like?
2/10 #DataScience#MachineLearning#Twitter
It’s the ability to click on a Tweet in your timeline and get a detailed explanation of why it was served to you. There are levels of model explainability.
3/10 #DataScience#MachineLearning#Twitter
I follow the topics ‘data science’ and ‘sneakers,’ so I get served a lot of content from those topics. Diving deeper is much more difficult. Out of all the sneakers-related posts, why was I served this one?
5/10 #DataScience#MachineLearning#Twitter
I’ve never engaged with that specific account, and I don’t follow it closely, so why did the recommender algorithm serve me that tweet? Diving deeper is much more difficult.
6/10 #DataScience#MachineLearning#Twitter
Out of all the sneakers-related posts, why was I served this one?
Does the recommender love this tweet because it was popular, or is this tweet popular because the recommender loved it? Socrates asked, ‘Whose bias do ya’ll seek?’
7/10 #DataScience#MachineLearning#Twitter
This will be a tougher conversation regarding political content and topics far more sensitive than sneakers and data science. I hope Elon and Twitter take the hard road instead of a superficial solution.
8/10 #DataScience#MachineLearning#Twitter
Euthyphro’s Dilemma is one of our field’s biggest challenges. When models become actors in human systems, they change the governing dynamics.
9/10 #DataScience#MachineLearning#Twitter
We must create a framework of acceptable influences, or people will never trust models enough to adopt them for high-value use cases.
10/10 #DataScience#MachineLearning#Twitter
• • •
Missing some Tweet in this thread? You can try to
force a refresh
How will companies move into the Metaverse? Most platform-based businesses are already there. Google, Amazon, and Facebook are all platform native companies so they have a clear lane into the Metaverse. 1/7 #Metaverse#Strategy
Their businesses have always been digital-first and built on a platform with access to a business ecosystem or marketplace. Building an increasingly capable platform grew their accessible ecosystems. 2/7 #Metaverse#Strategy
Platforms remove barriers to scale so a company like Amazon could disrupt and rapidly take market share from retail incumbents. Google and Facebook entered emerging, very small ecosystems-Google for search and Facebook for social. 3/7 #Metaverse#Strategy
Coaching and mentoring are learned capabilities. Businesses must invest in training leaders and senior technical individual contributors.
Coaching builds a farm system of talent. Here are some coaching lessons from my 15yrs in technical #leadership.
1/10 #careeradvice
1. Part of mentoring is being a career therapist. People seek out mentorship when they hit barriers they don't know how to break past. There's usually a lot of built up frustration to work through first.
2/10 #leadership#careeradvice
BUT coaching sessions must focus on improvement. I work through the emotions first but always spend the last 15-20 minutes on tangible next steps. Career therapy only works if they make progress towards long term goals.
3/10 #leadership#careeradvice
If you're a Data Scientist who wants to be a better developer or builder, here's a thread on how to do it. There's so much bad advice out there, and I hope this helps clear things up. 1/8 #DataScience#MachineLearning#Programming
1. Spend a year coding as part of a team. Have people review your code and participate in code reviews. This will help you unlearn many bad habits. You'll also get exposure to different styles and best practices. 2/8 #DataScience#MachineLearning#Programming
2. Build traditional software engineering type projects. Services and Web Apps are great because you'll learn fundamental coding skills.
Most Data Strategies are missing a critical component. It's a Data Monetization Catalog, and they are not difficult to build. Here's my process: 1/8 #DataScience#MachineLearning#Data#Strategy
The process starts with the question, what use cases is this data used for? Use cases have business value, and it's a straight-line connection. 2/8 #DataScience#MachineLearning#Data#Strategy
I walk clients through this exercise, and it reveals excellent insights because data catalogs and dictionaries are connected to technical use cases but rarely to business use cases.
Data Science introduces a new model or architecture weekly, and it can be tough to keep up. Here are some of the basics and recent releases with resources to help you quickly understand each one.
1/15 #DataScience#MachineLearning#DeepLearning
Let's start with DALL E2. Here's a python implementation. Sometimes the easiest way to learn about it is to use it.
Google recently released an overview of PaLM. It's one of a growing list of large scale language models improving on the capabilities of earlier models like GPT-3. Deep learning is going big.
The Data Science learning path today is different than it was 3 years ago and looks nothing like it did 7 years ago. This thread has the main layers and example resources covering the basics, assuming you've got basic math covered.
1/18 #DataScience#MachineLearning
1. Research Methods. We do a lot of research and experimentation now. Data Scientists used to be model-centric but that's changed because our work must meet higher reliability requirements. I wrote an intro post: vinvashishta.substack.com/p/a-basic-intr…
2/18 #DataScience#MachineLearning
2. Causal Inference. Data Science has taken a hard turn towards causal inference, again to meet increasing model reliability requirements. An education on CI always starts with Pearl. ftp.cs.ucla.edu/pub/stat_ser/r…
3/18 #DataScience#MachineLearning