Evaluating Federated Pre-Training: On the Reliability of Downstream Fine-Tuning and Intrinsic Evaluation

• URL: http://arxiv.org/abs/2607.28658v1
• Date: Mon, 20 Jul 2026 15:14:12 GMT
• Title: Evaluating Federated Pre-Training: On the Reliability of Downstream Fine-Tuning and Intrinsic Evaluation
• Abstract summary: Federated pre-training offers a way to train foundation models on private or distributed data without centralizing the underlying datasets.<n>We study which evaluation protocol more reliably reflects federated pre-training quality.
• Score: 3.488204158798457
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Federated pre-training offers a way to train foundation models on private or distributed data without centralizing the underlying datasets. However, evaluating federated pre-training remains challenging because differences in client participation and local data availability can make directly comparable evaluation difficult. Moreover, pre-training test perplexity is tied to the pre-training distribution, while downstream benchmarks introduce task-specific adaptation that may not faithfully reflect the test perplexity established during pre-training. In this work, we study which evaluation protocol more reliably reflects federated pre-training quality. Using a controlled set of centralized and federated-trained models of a 16M parameter transformer model trained on identical client data, we assess evaluation protocols by whether they preserve a reference ranking established on the same pre-training testset. We compare downstream fine-tuning on GLUE, including full, head-only, and reduced-data variants, with next-token prediction on GLUE text as an intrinsic evaluation signal. Our results show that downstream fine-tuning does not reliably preserve the pre-training ranking, whereas direct next-token prediction exhibits a strong correspondence with the pre-training test perplexity. These findings suggest that downstream fine-tuning alone can be misleading when comparing federated pre-trained models, and that evaluation signals closer to the original pre-training objective deserve greater attention.

Related papers

• Expected Gain-based Escalation in Vertical Federated Learning [6.43987569868938]
  Collaborative inference can improve predictive performance by integrating complementary information across agents.<n>Applying collaborative fusion to every sample can incur unnecessary communication and computational overhead.<n>We study selective escalation in a two-round VFL inference protocol.
  arXiv  Detail & Related papers  (2026-06-30T08:32:53Z)
• Early Data Exposure Improves Robustness to Subsequent Fine-Tuning [26.380638571847815]
  We study how upstream training choices shape how robustly that capability is retained.<n>Early exposure - mixing post-training data into pretraining - consistently improves the frontier between retained upstream performance and downstream performance.<n>Our findings suggest that robustness to subsequent fine-tuning should be treated as a first-class objective of upstream training.
  arXiv  Detail & Related papers  (2026-05-12T20:08:00Z)
• Online Bayesian Imbalanced Learning with Bregman-Calibrated Deep Networks [0.7106986689736825]
  We present textitOnline Bayesian Imbalanced Learning (OBIL), a principled framework that decouples likelihood-ratio estimation from class-prior assumptions.<n>Our approach builds on the established connection between Bregman divergences and proper scoring rules to show that deep networks trained with such losses produce posterior probability estimates.<n>We prove that these likelihood-ratio estimates remain valid under arbitrary changes in class priors and cost structures, requiring only a threshold adjustment for optimal Bayes decisions.
  arXiv  Detail & Related papers  (2026-02-08T21:23:00Z)
• Can Small Training Runs Reliably Guide Data Curation? Rethinking Proxy-Model Practice [109.9635246405237]
  We show that the experiment conclusions about data quality can flip with even minor adjustments to training hyper parameters.<n>We introduce a simple patch to the evaluation protocol: using reduced learning rates for proxy model training.<n> Empirically, we validate this approach across 23 data recipes covering four critical dimensions of data curation.
  arXiv  Detail & Related papers  (2025-12-30T23:02:44Z)
• Provably Reliable Conformal Prediction Sets in the Presence of Data Poisoning [53.42244686183879]
  Conformal prediction provides model-agnostic and distribution-free uncertainty quantification.<n>Yet, conformal prediction is not reliable under poisoning attacks where adversaries manipulate both training and calibration data.<n>We propose reliable prediction sets (RPS): the first efficient method for constructing conformal prediction sets with provable reliability guarantees under poisoning.
  arXiv  Detail & Related papers  (2024-10-13T15:37:11Z)
• TeLeS: Temporal Lexeme Similarity Score to Estimate Confidence in End-to-End ASR [1.8477401359673709]
  Class-probability-based confidence scores do not accurately represent quality of overconfident ASR predictions. We propose a novel Temporal-Lexeme Similarity (TeLeS) confidence score to train Confidence Estimation Model (CEM) We conduct experiments with ASR models trained in three languages, namely Hindi, Tamil, and Kannada, with varying training data sizes.
  arXiv  Detail & Related papers  (2024-01-06T16:29:13Z)
• Unsupervised Dense Retrieval with Relevance-Aware Contrastive Pre-Training [81.3781338418574]
  We propose relevance-aware contrastive learning. We consistently improve the SOTA unsupervised Contriever model on the BEIR and open-domain QA retrieval benchmarks. Our method can not only beat BM25 after further pre-training on the target corpus but also serves as a good few-shot learner.
  arXiv  Detail & Related papers  (2023-06-05T18:20:27Z)
• Self-Distillation for Further Pre-training of Transformers [83.84227016847096]
  We propose self-distillation as a regularization for a further pre-training stage. We empirically validate the efficacy of self-distillation on a variety of benchmark datasets for image and text classification tasks.
  arXiv  Detail & Related papers  (2022-09-30T02:25:12Z)
• Semantic Self-adaptation: Enhancing Generalization with a Single Sample [45.111358665370524]
  We propose a self-adaptive approach for semantic segmentation. It fine-tunes the parameters of convolutional layers to the input image using consistency regularization. Our empirical study suggests that self-adaptation may complement the established practice of model regularization at training time.
  arXiv  Detail & Related papers  (2022-08-10T12:29:01Z)
• Robust Flow-based Conformal Inference (FCI) with Statistical Guarantee [4.821312633849745]
  We develop a series of conformal inference methods, including building predictive sets and inferring outliers for complex and high-dimensional data. We evaluate our method, robust flow-based conformal inference, on benchmark datasets.
  arXiv  Detail & Related papers  (2022-05-22T04:17:30Z)
• Unsupervised Domain Adaptation for Speech Recognition via Uncertainty Driven Self-Training [55.824641135682725]
  Domain adaptation experiments using WSJ as a source domain and TED-LIUM 3 as well as SWITCHBOARD show that up to 80% of the performance of a system trained on ground-truth data can be recovered.
  arXiv  Detail & Related papers  (2020-11-26T18:51:26Z)
• Pre-training Is (Almost) All You Need: An Application to Commonsense Reasoning [61.32992639292889]
  Fine-tuning of pre-trained transformer models has become the standard approach for solving common NLP tasks. We introduce a new scoring method that casts a plausibility ranking task in a full-text format. We show that our method provides a much more stable training phase across random restarts.
  arXiv  Detail & Related papers  (2020-04-29T10:54:40Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.