Pith. sign in

REVIEW 10 cited by

Generalizing to unseen domains via distribution matching

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

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 1911.00804 v6 pith:CFGVQJNG submitted 2019-11-03 cs.LG stat.ML

Isabela Albuquerque , João Monteiro , Mohammad Darvishi , Tiago H. Falk , Ioannis Mitliagkas This is my paper · ORCID
classification cs.LGstat.ML
keywords domainsdatadomaingeneralizationassumptionsbounddiscrepancydistributions
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

Supervised learning results typically rely on assumptions of i.i.d. data. Unfortunately, those assumptions are commonly violated in practice. In this work, we tackle such problem by focusing on domain generalization: a formalization where the data generating process at test time may yield samples from never-before-seen domains (distributions). Our work relies on the following lemma: by minimizing a notion of discrepancy between all pairs from a set of given domains, we also minimize the discrepancy between any pairs of mixtures of domains. Using this result, we derive a generalization bound for our setting. We then show that low risk over unseen domains can be achieved by representing the data in a space where (i) the training distributions are indistinguishable, and (ii) relevant information for the task at hand is preserved. Minimizing the terms in our bound yields an adversarial formulation which estimates and minimizes pairwise discrepancies. We validate our proposed strategy on standard domain generalization benchmarks, outperforming a number of recently introduced methods. Notably, we tackle a real-world application where the underlying data corresponds to multi-channel electroencephalography time series from different subjects, each considered as a distinct domain.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 10 Pith papers

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

  1. ACE and Diverse Generalization via Selective Disagreement

    cs.LG 2025-09 conditional novelty 6.0 of 10

    ACE learns an ensemble of classifiers that agree on labeled data but confidently and selectively disagree on target-distribution data, recovering diverse human-interpretable concepts under complete spurious correlation.

  2. Prototypical Progressive Alignment and Reweighting for Generalizable Semantic Segmentation

    cs.CV 2025-07 conditional novelty 6.0 of 10

    PPAR aligns segmentation features with CLIP text prototypes from easy to hard and reweights source pixels to improve generalization to unseen domains, achieving top average mIoU on several urban scene benchmarks.

  3. A General Method for Detecting Information Generated by Large Language Models

    cs.CL 2025-06 conditional novelty 6.0 of 10

    A new LLM-text detector built from twin memory networks and domain-generalization losses outperforms prior detectors on held-out LLMs and domains in the paper's benchmark.

  4. Multi-Domain Graph Foundation Models: Robust Knowledge Transfer via Topology Alignment

    cs.SI 2025-02 conditional novelty 6.0 of 10

    MDGFM aligns graph topologies across domains with learnable tokens and structure refinement, improving few-shot transfer to unseen graphs.

  5. Towards Understanding Extrapolation: a Causal Lens

    cs.LG 2025-01 conditional novelty 6.0 of 10

    Under a minimal-change latent model, the invariant part of a representation is identifiable from one off-support target sample, under bounded dense shifts or arbitrary sparse shifts.

  6. Open-Set Heterogeneous Domain Adaptation: Theoretical Analysis and Algorithm

    cs.LG 2024-12 conditional novelty 6.0 of 10

    RL-OSHeDA, a two-stage representation learning method with pseudo-labeling, outperforms existing domain adaptation baselines on 56 open-set heterogeneous domain adaptation tasks.

  7. Cross-Receiver Generalization for RF Fingerprint Identification via Feature Disentanglement and Adversarial Training

    cs.LG 2025-10 conditional novelty 5.0 of 10

    DRIFT disentangles transmitter and receiver features and uses adversarial alignment plus receiver-clustering regularization to keep RF fingerprint identification accurate on unseen receivers.

  8. Domain Generalization via Pareto Optimal Gradient Matching

    cs.LG 2025-07 conditional novelty 5.0 of 10

    POGM combines pairwise gradient inner product maximization with a ball constraint around the ERM gradient, using a meta-learning update that avoids second-order derivatives and achieves competitive DomainBed accuracy.

  9. Weight Averaging for Out-of-Distribution Generalization and Few-Shot Domain Adaptation

    cs.CV 2025-01 reject novelty 4.0 of 10

    Gradient-similarity-regularized weight averaging and WA+SAM fine-tuning are tested on OOD and few-shot domain adaptation benchmarks, with mixed results that do not support the claimed improvements.

  10. Epistemic Artificial Intelligence is Essential for Machine Learning Models to Truly 'Know When They Do Not Know'

    cs.AI 2025-05 conditional novelty 3.0 of 10

    Machine learning should use second-order uncertainty measures, such as credal sets and random sets, so models can explicitly represent ignorance and avoid overconfident predictions on unfamiliar data.

Pith tools