Pinterest's FORTE framework safely finds and removes unneeded features in production recommenders
_reachsumit · x · 2026-10-12
Pinterest presents FORTE, a model-agnostic production framework for discovering, validating and removing features in industrial recommender systems. The core challenge is feature-lifecycle asymmetry: adding a feature must produce gains, while removal must not regress any metric — complicated by feature interactions, daily importance shifts, A/B test contention and multi-model consumption.
Key techniques: an efficient perturbation implementation enabling daily feature evaluation; grouped ablation over candidate sets built from all low-rank features to handle interactions; offline results predictive enough of online A/B tests to confidently advance deprecations; and lineage tracing that identifies all downstream consumer models to realize full training-set savings.
Validated over two large-scale recommender systems.
More from Research
- 1,200 AI answers show Reddit citations are nearly all from Google engines: Gemini 52.4%, ChatGPT 0% — drrickio · 2026-10-12
- Open experiment log pushes integer multiplication bounds down to κ≈0.001467, invites agent collaboration — IgorCarron · 2026-10-12
- LLMs don't guess, they optimize: a statistical mechanics take claims 3,418 gradient steps to coherence — TinfoilTricorn · 2026-10-12
- A critical look at Lean4: why no one can prove the proof assistant is consistent — thomasahle · 2026-10-12
- ThinkDrag: mmlab-ntu adds visual reasoning to drag-based image editing, claims SOTA — mmlab-ntu · 2026-10-12
- Dex-One2Many learns dexterous manipulation from one human video, +71% on unseen setups — Jusuk Lee · 2026-10-12