Just ask: Designing intent-driven algos
4 likes266 views
The document discusses the development of intent-driven algorithms in personalization, particularly within the context of clothing recommendation systems at Stitch Fix. It emphasizes the integration of user intent and preferences into recommendation models to enhance user experience and optimize stock management. The approach aims to improve decision-making for both stylists and customers by allowing for adaptability and responsiveness to changing preferences.
![Activity on your site
Activity on other sites
Inferences youve made
about them
Implicit Explicit
GeneralSpeci?c
20 years and [Amazon] still thinks I need
more than one vacuum.
https://news.ycombinator.com/item?id=14705752
Problem:
Preferences change over time
Collaborative ?ltering
(a la Amazon)](https://image.slidesharecdn.com/justaskpydatalondon-190714121705/85/Just-ask-Designing-intent-driven-algos-8-320.jpg)
Just ask: Designing intent-driven algos
- 1. Just ask: Designing intent-driven algos Anna Schneider Data Science Manager @windupanna Stitch Fix @stitch?x_algo
- 2. Womens, Mens, Kids Plus, Maternity, Petite, Big&Tall US, and now UK! Personal styling for everybody Keep what you love, send back the rest
- 3. Personal styling for everybody Art + science = personalization recommenders 100+ on Algorithms (not just recommendersalso demand forecasting, warehouse layout, experimentation, etc) Recommender = algo results that guide human decisions
- 4. 4 Clients (aka customers) Buyers What to stock for all clients? Stylists What to send to this client?
- 5. 5 Clients (aka customers) Buyers What to stock for all clients? Stylists What to send to this client? Lets build some fake styling recommenders!
- 6. Me A friend Im in the market for a new backpack, what should I get? Hmm, well what are you looking for? I love my old messenger bag, but I want a regular backpack this time. What do you like about your current bag? I like that its waterproof, has good pockets, ?ts a 15 laptop, and has lasted for years. It looks unique too :) Why get a new bag then? Well it ?ts pretty well, but I want something that ?ts even better. And it would be fun to try a new brand. Cool, I think youd like this one! General Speci?c Speci?c Implicit Explicit
- 7. Activity on your site Activity on other sites Inferences youve made about them Amazon https://pdfs.semanticscholar.org/0f06/d328f6deb44e5e67408e0c16a8c7356330d1.pdf Implicit Explicit GeneralSpeci?c Collaborative ?ltering (a la Amazon) 1 0 ... 1 1 1 ... 0 ... ... ... ... 0 1 ... 1 ... ... People who bought this also bought
- 8. Activity on your site Activity on other sites Inferences youve made about them Implicit Explicit GeneralSpeci?c 20 years and [Amazon] still thinks I need more than one vacuum. https://news.ycombinator.com/item?id=14705752 Problem: Preferences change over time Collaborative ?ltering (a la Amazon)
- 9. Activity on your site Activity on other sites Inferences youve made about them Ratings Did you want this Pro?le questions Implicit Explicit GeneralSpeci?c Content-based (a la OKCupid) OKCupid https://help.okcupid.com/article/128-how-is-match-calculated Youre a match! Preppy? Goth? Likes/is green? 1 1 1 1 0 1 0 0 0
- 10. Activity on your site Activity on other sites Inferences youve made about them Ratings Did you want this Pro?le questions Implicit Explicit GeneralSpeci?c Problem: Preferences may be incompatible Content-based (a la OKCupid) Preppy? Goth? Likes/is green? 1 1 1 1 0 1 0 0 0
- 11. Activity on your site Activity on other sites Inferences youve made about them Ratings Did you want this Pro?le questions Location Time Implicit Explicit GeneralSpeci?c Context-aware CF (a la Net?ix, Spotify) Recommender Systems Handbook (2011) http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.423.4220&rep=rep1&type=pdf Mobasher (KDD 2014) https://www.kdd.org/kdd2014/tutorials/KDD-%20The%20RecommenderProblemRevisited-Part2.pdf Net?ix https://help.net?ix.com/en/node/100639 People who bought this in your situation also bought 1 0 ... 1 US Jan 1 1 ... 0 UK Jul ... ... ... ... ... ... 0 1 ... 1 US Sep ... ...
- 12. Activity on your site Activity on other sites Inferences youve made about them Ratings Did you want this Pro?le questions Location Time Implicit Explicit GeneralSpeci?c Problem: People like you are boring Context-aware CF (a la Net?ix, Spotify) The more vanilla the release, the better it works for Spotify. If its challenging music? Nah. https://thebaf?er.com/salvos/the-problem-with- muzak-pelly
- 13. Activity on your site Activity on other sites Inferences youve made about them Ratings Did you want this Pro?le questions Location Time Search terms Con?g params NLP Selecting ?lters/tags Implicit Explicit GeneralSpeci?c Solution: Just ask! Intent-driven (via request note) I loved the distressed jeans in the last ?x! This time Id like an upscale out?t for my cousins wedding.
- 14. Intent-driven I loved the distressed jeans in the last ?x! This time Id like an upscale out?t for my cousins wedding. Use with classic recommenders or other algos Allows growth mindset for your model of user Enables stylist + algo to make good decisions on behalf of client Smarter explore-exploit
- 15. 15 Clients (aka customers) Buyers What to stock for all clients? Stylists What to send to this client? Lets build a real merch recommender!
- 16. 16 Old-school buying ++ + - +++ -- Hunt and peck Vendor offerings Ranked by performance
- 17. 17 Back-endFront-end Data Filter to category: Shirts View recs API Filters For each item Predicted performance Predicted demand Dimensions (?t, color, etc) Predictions Architecture: before Filtered preds ++ + - +++ --
- 18. 18 Back-endFront-end Data Filter to category: Shirts View recs API Filters For each item Predicted performance Predicted demand Dimensions (?t, color, etc) Predictions Challenge: wrong problem? Filtered preds ++ + - +++ --
- 19. 19 Does this meet our revenue and margin objectives? Can we sell through what were stocking without missing opportunity? Welcome to the present, were running a real business Is there the right diversity across price points, aesthetics, ?ts, etc? Is it seasonally relevant for next season?
- 20. 20 Reframing: Constrained optimization Jeff Schecter's Multithreaded blog post Choose merch to stock our decision Maximize predicted performance what we care about (objective function) in order to Receipts cover a breadth of merch and clients Financial targets are achieved Supply is matched to demand while ensuring constraints
- 21. 21 Back-endFront-end Data Target margin? Target % blue? Target % slim-?t? $xx yy% zz% Get recs API User intent Solver engine Optimizer constraints Optimal solution Q2 performance ++ Margin on target Great for all body types Intent- driven recs For each item Predicted performance Predicted demand Dimensions (?t, color, etc) Preds Architecture: after
- 22. 22 Back-endFront-end Data Target margin? Target % blue? Target % slim-?t? $xx yy% zz% Get recs API Solver engine Q2 performance ++ Margin on target Great for all body types Challenge: speed For each item Predicted performance Predicted demand Dimensions (?t, color, etc) Batch ETLs for model training, of?ine prediction This might be slow, so rest should be fast
- 23. 23 Back-endFront-end Data Target margin? Target % blue? Target % slim-?t? $xx yy% zz% Get recs API Solver engine Q2 performance ++ Margin on target Great for all body types Challenge: UX For each item Predicted performance Predicted demand Dimensions (?t, color, etc) Help user encode intent Build trust that recs satisfy intent Encourage iteration to re?ne intent
- 24. 24 Back-endFront-end Data API Parallel: Maps Graph nodes and edges Predicted travel times User settings/prefs start destination Get route Routing engine Best route Batch ETLs for model training, of?ine prediction User intent Optimizer constraints Preds Optimal solution Intent- driven recs Iteration
- 25. Intent-driven algos enable DS to separate prediction task from recommendation task App Rec engine Batch preds modular and composable algos, teams, etc Batch preds Batch preds Batch preds Rec engine App App
- 26. Intent-driven algos require DS to process user input during online computation use task queues (eg celery), sanitize input, handle novel data App Rec engine Batch preds Batch preds Batch preds Batch preds Rec engine App App
- 27. Intent-driven algos enable users to ask different questions and get different answers Intent iterative, conversation-like Recs AlgosUser
- 28. Intent-driven algos require users to understand how to have a conversation with algos Intent support structured thinking, connect recs to intent Recs AlgosUser
- 29. Preferences change over time Preferences may be incompatible Why classical recommenders fail Intent-driven algos Collect speci?c as well as general preferences Enable many outcomes for the same user Algos guide user behavior Users guide algo behavior (human in the loop)
- 30. Anna Schneider Thanks!Data Science Manager @windupanna Stitch Fix @stitch?x_algo