LoKA enables practical FP8 use in numerically sensitive large recommendation models via online profiling of activations, reusable model modifications for stability, and dynamic kernel dispatching.
On the embedding collapse when scaling up recommendation models
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
citation-role summary
background 2
citation-polarity summary
years
2026 3verdicts
UNVERDICTED 3roles
background 2polarities
background 2representative citing papers
RankUp raises effective rank of representations in deep MetaFormer recommenders via randomized splitting and multi-embeddings, delivering 2-5% GMV gains in production deployments at Weixin.
MPZCH applies multi-probe linear hashing plus eviction policies to achieve zero collisions on user embeddings and higher freshness on item embeddings while keeping training and inference speeds comparable to standard methods.
citing papers explorer
-
LoKA: Low-precision Kernel Applications for Recommendation Models At Scale
LoKA enables practical FP8 use in numerically sensitive large recommendation models via online profiling of activations, reusable model modifications for stability, and dynamic kernel dispatching.
-
RankUp: Towards High-rank Representations for Large Scale Advertising Recommender Systems
RankUp raises effective rank of representations in deep MetaFormer recommenders via randomized splitting and multi-embeddings, delivering 2-5% GMV gains in production deployments at Weixin.
-
Multi-Probe Zero Collision Hash (MPZCH): Mitigating Embedding Collisions and Enhancing Model Freshness in Large-Scale Recommenders
MPZCH applies multi-probe linear hashing plus eviction policies to achieve zero collisions on user embeddings and higher freshness on item embeddings while keeping training and inference speeds comparable to standard methods.