Pith. sign in

REVIEW 10 cited by

On the Embedding Collapse when Scaling up Recommendation Models

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2310.04400 v2 pith:5XAWQHPF submitted 2023-10-06 cs.LG cs.IR

On the Embedding Collapse when Scaling up Recommendation Models

classification cs.LG cs.IR
keywords modelscollapseembeddingrecommendationscalabilityinteractionanalysisavailable
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Recent advances in foundation models have led to a promising trend of developing large recommendation models to leverage vast amounts of available data. Still, mainstream models remain embarrassingly small in size and na\"ive enlarging does not lead to sufficient performance gain, suggesting a deficiency in the model scalability. In this paper, we identify the embedding collapse phenomenon as the inhibition of scalability, wherein the embedding matrix tends to occupy a low-dimensional subspace. Through empirical and theoretical analysis, we demonstrate a \emph{two-sided effect} of feature interaction specific to recommendation models. On the one hand, interacting with collapsed embeddings restricts embedding learning and exacerbates the collapse issue. On the other hand, interaction is crucial in mitigating the fitting of spurious features as a scalability guarantee. Based on our analysis, we propose a simple yet effective multi-embedding design incorporating embedding-set-specific interaction modules to learn embedding sets with large diversity and thus reduce collapse. Extensive experiments demonstrate that this proposed design provides consistent scalability and effective collapse mitigation for various recommendation models. Code is available at this repository: https://github.com/thuml/Multi-Embedding.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 10 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. ROCS: Request-Oriented Compute Sharing for Efficient Large-Scale Recommendation

    cs.LG 2026-07 conditional novelty 7.0

    ROCS restructures recommendation models so user-side computation is shared across all candidate items, yielding up to 3x serving throughput at equal or better prediction quality.

  2. The Pitfall of Scaling Up: Uncovering and Mitigating Popularity Bias Amplification in Scaling Transformer-based Recommenders

    cs.IR 2026-06 unverdicted novelty 7.0

    Transformer recommenders amplify popularity bias via spectral collapse when scaled; SPRINT constrains attention column-sums and feed-forward spectral norms to improve fairness and scaling behavior.

  3. SpecFormer: Mitigating Embedding and Attention Collapse via Spectral-Aware Transformer for Recommendation

    cs.IR 2026-07 conditional novelty 6.0

    A spectral-softening Transformer that flattens embedding singular values reduces attention collapse and improves CTR prediction accuracy in recommender systems.

  4. DiffoR: A Unified Continuous Generative Framework for Universal Ordinal Regression

    cs.LG 2026-05 unverdicted novelty 6.0

    DiffOR reformulates ordinal regression as continuous generative modeling using diffusion models with dual-decoupling to capture soft semantic transitions.

  5. LoKA: Low-precision Kernel Applications for Recommendation Models At Scale

    cs.LG 2026-05 unverdicted novelty 6.0

    LoKA enables practical FP8 use in numerically sensitive large recommendation models via profiling, model adaptations, and runtime kernel orchestration.

  6. LoKA: Low-precision Kernel Applications for Recommendation Models At Scale

    cs.LG 2026-05 unverdicted novelty 6.0

    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.

  7. RankUp: Towards High-rank Representations for Large Scale Advertising Recommender Systems

    cs.IR 2026-04 unverdicted novelty 6.0

    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.

  8. Multi-Probe Zero Collision Hash (MPZCH): Mitigating Embedding Collisions and Enhancing Model Freshness in Large-Scale Recommenders

    cs.LG 2026-02 unverdicted novelty 6.0

    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.

  9. LoKA: Low-precision Kernel Applications for Recommendation Models At Scale

    cs.LG 2026-05 unverdicted novelty 5.0

    LoKA claims to make FP8 practical for large recommendation models via statistical probing, model adaptations, and accuracy-aware kernel dispatch.

  10. RankUp: Towards High-rank Representations for Large Scale Advertising Recommender Systems

    cs.IR 2026-04 unverdicted novelty 5.0

    RankUp enhances representation capacity in deep MetaFormer recommenders via permutation splitting and multi-embeddings, achieving GMV improvements of 2-5% in Weixin production systems.