Applied ML | Recommender systems
Subscribe
Sign in
Home
Notes
Archive
About
Latest
Top
Discussions
Ranking model calibration in recommender systems
We define calibration of ranking models in calibration, the benefit it can bring to prioritize calibration and how to achieve it without affecting…
Jun 9
•
Gaurav Chakravorty
and
Marc Ferradou
14
Share this post
Ranking model calibration in recommender systems
recsysml.substack.com
Copy link
Facebook
Email
Note
Other
May 2024
Entrypoint retention modeling in recommender systems
Choose/rank items at the entrypoint of a recommended feed to drive retention and not just consumption
May 24
•
Gaurav Chakravorty
,
Vish Sangale
, and
Nimit Desai
5
Share this post
Entrypoint retention modeling in recommender systems
recsysml.substack.com
Copy link
Facebook
Email
Note
Other
Optimal whole page ranking = reward / risk
We show how tech can learn from finance in using risk models for better feed construction of recommender systems.
May 11
•
Gaurav Chakravorty
and
Vish Sangale
5
Share this post
Optimal whole page ranking = reward / risk
recsysml.substack.com
Copy link
Facebook
Email
Note
Other
7
User representation in a recommender system | memorization vs generalization
We look at memorization, generalization and mixture of representations based implementations for user preference representation in a recommender system
May 3
•
Gaurav Chakravorty
,
Hong Chen
, and
Saurabh Gupta
15
Share this post
User representation in a recommender system | memorization vs generalization
recsysml.substack.com
Copy link
Facebook
Email
Note
Other
January 2024
Reducing selection bias / popularity bias in ranking
Through the post, code and videos we show how to make your multi-task ranking model unbiased
Jan 20
•
Gaurav Chakravorty
and
Ameya Raul
9
Share this post
Reducing selection bias / popularity bias in ranking
recsysml.substack.com
Copy link
Facebook
Email
Note
Other
How to reduce cost of ranking by knowledge distillation
Using knowledge distillation can make your early ranking model more aligned with final ranker + Sample Code + Video walkthrough
Jan 6
•
Gaurav Chakravorty
11
Share this post
How to reduce cost of ranking by knowledge distillation
recsysml.substack.com
Copy link
Facebook
Email
Note
Other
December 2023
Does your model get better at task T when you rank by estimated probability p(T) ?
To understand what to optimize in a ranking model
Dec 22, 2023
•
Gaurav Chakravorty
6
Share this post
Does your model get better at task T when you rank by estimated probability p(T) ?
recsysml.substack.com
Copy link
Facebook
Email
Note
Other
System design of an Early Ranker
Towards greater user satisfaction, recommendation quality, reduced latency and compute capacity savings + Sample code
Dec 15, 2023
•
Gaurav Chakravorty
9
Share this post
System design of an Early Ranker
recsysml.substack.com
Copy link
Facebook
Email
Note
Other
November 2023
Early (Stage) Ranking in recommender systems
We look at "early (stage) ranking", why it is needed, what is success for it and end with a baseline implementation of it.
Nov 26, 2023
•
Gaurav Chakravorty
17
Share this post
Early (Stage) Ranking in recommender systems
recsysml.substack.com
Copy link
Facebook
Email
Note
Other
2
August 2022
Using ML to build a social networking notification system
System design of an ML driven notification system of a social network
Aug 26, 2022
•
Gaurav Chakravorty
3
Share this post
Using ML to build a social networking notification system
recsysml.substack.com
Copy link
Facebook
Email
Note
Other
July 2022
Life of a search query
Presentation of how search is today and how that is very different from the mental model users have of search in early 2000s
Jul 30, 2022
•
Gaurav Chakravorty
12
Share this post
Life of a search query
recsysml.substack.com
Copy link
Facebook
Email
Note
Other
June 2022
How to build ML-first products
Problems companies face in adopting ML and how to solve them
Jun 24, 2022
•
Gaurav Chakravorty
22
Share this post
How to build ML-first products
recsysml.substack.com
Copy link
Facebook
Email
Note
Other
Share
Copy link
Facebook
Email
Note
Other
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts