idSCD: Identifying Training Datasets through Semantic Correlation Descriptors

• URL: http://arxiv.org/abs/2605.30462v1
• Date: Thu, 28 May 2026 18:38:15 GMT
• Title: idSCD: Identifying Training Datasets through Semantic Correlation Descriptors
• Abstract summary: We argue that datasets leave dataset-specific traces in a model's learned semantic correlation structure.<n>We introduce a white-box semantic fingerprinting approach based on semantic correlation descriptors (SCDs)<n>We propose a practical SCD-based membership score that tests whether a target dataset is part of a model's training mixture.
• Score: 5.756453494739546
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Can a dataset be recognized from the spurious correlations it induces during training? We argue that datasets leave dataset-specific traces in a model's learned semantic correlation structure: incidental regularities that are predictive within a dataset, but not causal for the underlying task, can be internalized during training. We use this insight to study dataset-level membership inference, moving beyond existing methods that rely on behavioral or distributional evidence such as confidence scores, losses, margins, generated samples, or query responses. We introduce a white-box semantic fingerprinting approach based on semantic correlation descriptors (SCDs), which capture the semantic correlation structure learned by a model and make it comparable across dataset mixtures. In a controlled leave-one-dataset-out diagnostic, SCDs recover dataset-specific changes and perfectly separate matching from non-matching dataset pairs. We then propose a practical SCD-based membership score that tests whether a target dataset is part of a model's training mixture using only the model's SCD and the target dataset's standalone SCD, without requiring leave-one-dataset-out models. Across three diverse experimental settings, with dataset groups for natural language inference, emotion classification, and medical text classification, we test both the advantages and limitations of SCD-based membership inference with different degrees of semantic separation and keyword support between dataset splits. On average, the classifier based on this score achieves the highest performance and the lowest std, outperforming black-box baselines RMIA, Attack-P, and LiRA, as well as the white-box SIF baseline. These results show that dataset membership can be traced through internal semantic correlations, with the largest relative gain exceeding 60% in ROC-AUC when dataset groups expose distinct semantic particularities.

Related papers

• Benchmarking External Generalization of SPD Matrix Learning for Resting-State fMRI Connectome Prediction [35.50001465263556]
  We ask whether within-dataset performance remains when the test data come from an entirely held-out rs-fMRI dataset.<n>We introduce a reproducible age-prediction benchmark across six rs-fMRI datasets.
  arXiv  Detail & Related papers  (2026-08-31T08:15:08Z)
• Statistical Embeddings for Similarity, Retrieval, and Interpretable Alignment of Numeric Tabular Datasets [0.0]
  The proposed methodology characterizes numeric datasets through structured exploratory data analysis descriptors.<n>A penalized formulation of Canonical Correlation Analysis (CCA) is applied to recover sparse, interpretable variable-level correspondences.<n>The methodology is evaluated across 15 datasets spanning general-purpose benchmarks, materials informatics, and nuclear-grade graphite characterization.
  arXiv  Detail & Related papers  (2026-05-28T17:40:42Z)
• SAS: Semantic-aware Sampling for Generative Dataset Distillation [55.27114962330541]
  We introduce a semantic-aware perspective for dataset distillation by leveraging Contrastive Language-Image Pretraining (CLIP) as a semantic prior for post-sampling.<n>Our goal is to obtain distilled datasets that are not only compact but also semantically class-discriminative and diverse.
  arXiv  Detail & Related papers  (2026-05-18T08:05:46Z)
• Bridging the Semantic Gap for Categorical Data Clustering via Large Language Models [64.58262227709842]
  ARISE (Attention-weighted Representation with Integrated Semantic Embeddings) is presented.<n>It builds semantic-aware representations that complement the metric space of categorical data for accurate clustering.<n>Experiments on eight benchmark datasets demonstrate consistent improvements over seven representative counterparts.
  arXiv  Detail & Related papers  (2026-01-03T11:37:46Z)
• AssayMatch: Learning to Select Data for Molecular Activity Models [21.367469322467418]
  AssayMatch is a framework for data selection that builds smaller, more homogenous training sets attuned to the test set of interest.<n>We demonstrate that models trained on data selected by AssayMatch are able to surpass the performance of the model trained on the complete dataset.
  arXiv  Detail & Related papers  (2025-11-20T06:25:51Z)
• CBW: Towards Dataset Ownership Verification for Speaker Verification via Clustering-based Backdoor Watermarking [85.68235482145091]
  Large-scale speech datasets have become valuable intellectual property.<n>We propose a novel dataset ownership verification method.<n>Our approach introduces a clustering-based backdoor watermark (CBW)<n>We conduct extensive experiments on benchmark datasets, verifying the effectiveness and robustness of our method against potential adaptive attacks.
  arXiv  Detail & Related papers  (2025-03-02T02:02:57Z)
• Measuring Bias of Web-filtered Text Datasets and Bias Propagation Through Training [22.53813258871828]
  We investigate biases in pretraining datasets for large language models (LLMs) through dataset classification experiments.<n>We find that neural networks can classify surprisingly well which dataset a single text sequence belongs to, significantly better than a human can.
  arXiv  Detail & Related papers  (2024-12-03T21:43:58Z)
• Dataset Distillation-based Hybrid Federated Learning on Non-IID Data [18.226454363903446]
  We propose a hybrid federated learning framework called HFLDD, which integrates dataset distillation to generate independent and equally distributed (IID) data.<n>In particular, we partition the clients into heterogeneous clusters, where the data labels among different clients within a cluster are unbalanced.<n>This training process is like traditional federated learning on IID data, and hence effectively alleviates the impact of non-IID data on model training.
  arXiv  Detail & Related papers  (2024-09-26T03:52:41Z)
• Downstream-Pretext Domain Knowledge Traceback for Active Learning [138.02530777915362]
  We propose a downstream-pretext domain knowledge traceback (DOKT) method that traces the data interactions of downstream knowledge and pre-training guidance. DOKT consists of a traceback diversity indicator and a domain-based uncertainty estimator. Experiments conducted on ten datasets show that our model outperforms other state-of-the-art methods.
  arXiv  Detail & Related papers  (2024-07-20T01:34:13Z)
• Affinity Clustering Framework for Data Debiasing Using Pairwise Distribution Discrepancy [10.184056098238765]
  Group imbalance, resulting from inadequate or unrepresentative data collection methods, is a primary cause of representation bias in datasets. This paper presents MASC, a data augmentation approach that leverages affinity clustering to balance the representation of non-protected and protected groups of a target dataset.
  arXiv  Detail & Related papers  (2023-06-02T17:18:20Z)
• Data-SUITE: Data-centric identification of in-distribution incongruous examples [81.21462458089142]
  Data-SUITE is a data-centric framework to identify incongruous regions of in-distribution (ID) data. We empirically validate Data-SUITE's performance and coverage guarantees.
  arXiv  Detail & Related papers  (2022-02-17T18:58:31Z)
• Learning to Model and Ignore Dataset Bias with Mixed Capacity Ensembles [66.15398165275926]
  We propose a method that can automatically detect and ignore dataset-specific patterns, which we call dataset biases. Our method trains a lower capacity model in an ensemble with a higher capacity model. We show improvement in all settings, including a 10 point gain on the visual question answering dataset.
  arXiv  Detail & Related papers  (2020-11-07T22:20:03Z)

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.