How robust are pre-trained models to distribution shift?
- URL: http://arxiv.org/abs/2206.08871v1
- Date: Fri, 17 Jun 2022 16:18:28 GMT
- Title: How robust are pre-trained models to distribution shift?
- Authors: Yuge Shi, Imant Daunhawer, Julia E. Vogt, Philip H.S. Torr, Amartya
Sanyal
- Abstract summary: We show how spurious correlations affect the performance of popular self-supervised learning (SSL) and auto-encoder based models (AE)
We develop a novel evaluation scheme with the linear head trained on out-of-distribution (OOD) data, to isolate the performance of the pre-trained models from a potential bias of the linear head used for evaluation.
- Score: 82.08946007821184
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: The vulnerability of machine learning models to spurious correlations has
mostly been discussed in the context of supervised learning (SL). However,
there is a lack of insight on how spurious correlations affect the performance
of popular self-supervised learning (SSL) and auto-encoder based models (AE).
In this work, we shed light on this by evaluating the performance of these
models on both real world and synthetic distribution shift datasets. Following
observations that the linear head itself can be susceptible to spurious
correlations, we develop a novel evaluation scheme with the linear head trained
on out-of-distribution (OOD) data, to isolate the performance of the
pre-trained models from a potential bias of the linear head used for
evaluation. With this new methodology, we show that SSL models are consistently
more robust to distribution shifts and thus better at OOD generalisation than
AE and SL models.
Related papers
- Revisiting Spurious Correlation in Domain Generalization [12.745076668687748]
We build a structural causal model (SCM) to describe the causality within data generation process.
We further conduct a thorough analysis of the mechanisms underlying spurious correlation.
In this regard, we propose to control confounding bias in OOD generalization by introducing a propensity score weighted estimator.
arXiv Detail & Related papers (2024-06-17T13:22:00Z) - Low-rank finetuning for LLMs: A fairness perspective [54.13240282850982]
Low-rank approximation techniques have become the de facto standard for fine-tuning Large Language Models.
This paper investigates the effectiveness of these methods in capturing the shift of fine-tuning datasets from the initial pre-trained data distribution.
We show that low-rank fine-tuning inadvertently preserves undesirable biases and toxic behaviors.
arXiv Detail & Related papers (2024-05-28T20:43:53Z) - Evaluating Fairness in Self-supervised and Supervised Models for
Sequential Data [10.626503137418636]
Self-supervised learning (SSL) has become the de facto training paradigm of large models.
This study explores the impact of pre-training and fine-tuning strategies on fairness.
arXiv Detail & Related papers (2024-01-03T09:31:43Z) - A PAC-Bayesian Perspective on the Interpolating Information Criterion [54.548058449535155]
We show how a PAC-Bayes bound is obtained for a general class of models, characterizing factors which influence performance in the interpolating regime.
We quantify how the test error for overparameterized models achieving effectively zero training error depends on the quality of the implicit regularization imposed by e.g. the combination of model, parameter-initialization scheme.
arXiv Detail & Related papers (2023-11-13T01:48:08Z) - Guide the Learner: Controlling Product of Experts Debiasing Method Based
on Token Attribution Similarities [17.082695183953486]
A popular workaround is to train a robust model by re-weighting training examples based on a secondary biased model.
Here, the underlying assumption is that the biased model resorts to shortcut features.
We introduce a fine-tuning strategy that incorporates the similarity between the main and biased model attribution scores in a Product of Experts loss function.
arXiv Detail & Related papers (2023-02-06T15:21:41Z) - On the Generalization and Adaption Performance of Causal Models [99.64022680811281]
Differentiable causal discovery has proposed to factorize the data generating process into a set of modules.
We study the generalization and adaption performance of such modular neural causal models.
Our analysis shows that the modular neural causal models outperform other models on both zero and few-shot adaptation in low data regimes.
arXiv Detail & Related papers (2022-06-09T17:12:32Z) - Using Explainable Boosting Machine to Compare Idiographic and Nomothetic
Approaches for Ecological Momentary Assessment Data [2.0824228840987447]
This paper explores the use of non-linear interpretable machine learning (ML) models in classification problems.
Various ensembles of trees are compared to linear models using imbalanced synthetic and real-world datasets.
In one of the two real-world datasets, knowledge distillation method achieves improved AUC scores.
arXiv Detail & Related papers (2022-04-04T17:56:37Z) - Influence Tuning: Demoting Spurious Correlations via Instance
Attribution and Instance-Driven Updates [26.527311287924995]
influence tuning can help deconfounding the model from spurious patterns in data.
We show that in a controlled setup, influence tuning can help deconfounding the model from spurious patterns in data.
arXiv Detail & Related papers (2021-10-07T06:59:46Z) - Self-Damaging Contrastive Learning [92.34124578823977]
Unlabeled data in reality is commonly imbalanced and shows a long-tail distribution.
This paper proposes a principled framework called Self-Damaging Contrastive Learning to automatically balance the representation learning without knowing the classes.
Our experiments show that SDCLR significantly improves not only overall accuracies but also balancedness.
arXiv Detail & Related papers (2021-06-06T00:04:49Z) - How Training Data Impacts Performance in Learning-based Control [67.7875109298865]
This paper derives an analytical relationship between the density of the training data and the control performance.
We formulate a quality measure for the data set, which we refer to as $rho$-gap.
We show how the $rho$-gap can be applied to a feedback linearizing control law.
arXiv Detail & Related papers (2020-05-25T12:13:49Z)
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.