Pith. sign in

REVIEW 4 cited by

ScaleBiO: Scalable Bilevel Optimization for LLM Data Reweighting

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 2406.19976 v2 pith:QD64J2SY submitted 2024-06-28 cs.LG math.OC

ScaleBiO: Scalable Bilevel Optimization for LLM Data Reweighting

classification cs.LG math.OC
keywords bileveloptimizationdataparadigmscalebiollmsreweightingalgorithms
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Bilevel optimization has shown its utility across various machine learning settings, yet most algorithms in practice require second-order information, making it challenging to scale them up. Only recently, a paradigm of first-order algorithms has emerged in the theoretical literature, capable of effectively addressing bilevel optimization problems. Nevertheless, the practical efficiency of this paradigm remains unverified, particularly in the context of large language models (LLMs). This paper introduces the first scalable instantiation of this paradigm called ScaleBiO, focusing on bilevel optimization for large-scale LLM data reweighting. By combining with a recently proposed memory-efficient training technique called LISA, our novel algorithm allows the paradigm to scale to $\sim$30B-sized LLMs on $8\times$H100 GPUs, marking the first successful application of bilevel optimization under practical scenarios for large-sized LLMs. Empirically, extensive experiments on data reweighting verify the effectiveness of ScaleBiO for different-scaled models, including Llama-3-8B, Gemma-2-9B, Qwen-2-7B, and Qwen-2.5-32B, where bilevel optimization succeeds in instruction-following and math reasoning tasks, outperforming several popular baselines, including uniform sampling, influence-aware data filtering, and reference-model-based sampling methods. Theoretically, ScaleBiO ensures the optimality of the learned data weights, along with a convergence guarantee matching the conventional first-order bilevel optimization paradigm on smooth and strongly convex objectives.

discussion (0)

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

Forward citations

Cited by 4 Pith papers

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

  1. Distributionally Robust Multi-Task Reinforcement Learning via Adaptive Task Sampling

    cs.LG 2026-05 unverdicted novelty 7.0

    DRATS derives a minimax objective from a feasibility formulation of MTRL to adaptively sample tasks with the largest return gaps, leading to better worst-task performance on MetaWorld benchmarks.

  2. On the Condition Number Dependency in Bilevel Optimization

    math.OC 2025-11 conditional novelty 7.0

    NC-SC bilevel optimization provably needs Ω(κ_y^2 ε^-2) first-order oracle calls in the worst case, beating the minimax lower bound; a faster O~(κ_y^{7/2} ε^-2) fully first-order method is also given.

  3. BLISS: A Lightweight Bilevel Influence Scoring Method for Data Selection in Language Model Pretraining

    cs.LG 2025-10 conditional novelty 6.0

    A bilevel optimization method ranks pretraining data by training a small proxy model on weighted samples, yielding modest downstream-task gains without external pretrained models.

  4. TANDEM: Bi-Level Data Mixture Optimization with Twin Networks

    cs.LG 2026-06 unverdicted novelty 5.0

    TANDEM solves bi-level data mixture optimization for LLMs via twin proxy and reference networks that measure domain efficacy by model difference and up-weight beneficial domains, with claimed theoretical guarantees an...