User profiles for Shiori Sagawa
Shiori SagawaPhD Student, Stanford University Verified email at stanford.edu Cited by 7000 |
An investigation of why overparameterization exacerbates spurious correlations
S Sagawa, A Raghunathan… - … on Machine Learning, 2020 - proceedings.mlr.press
We study why overparameterization—increasing model size well beyond the point of zero
training error—can hurt test error on minority groups despite improving average test error …
training error—can hurt test error on minority groups despite improving average test error …
Distributionally robust neural networks for group shifts: On the importance of regularization for worst-case generalization
Overparameterized neural networks can be highly accurate on average on an iid test set yet
consistently fail on atypical groups of the data (eg, by learning spurious correlations that …
consistently fail on atypical groups of the data (eg, by learning spurious correlations that …
Distributionally robust language modeling
Language models are generally trained on data spanning a wide range of topics (eg, news,
reviews, fiction), but they might be applied to an a priori unknown target distribution (eg, …
reviews, fiction), but they might be applied to an a priori unknown target distribution (eg, …
On the opportunities and risks of foundation models
AI is undergoing a paradigm shift with the rise of models (eg, BERT, DALL-E, GPT-3) that are
trained on broad data at scale and are adaptable to a wide range of downstream tasks. We …
trained on broad data at scale and are adaptable to a wide range of downstream tasks. We …
Just train twice: Improving group robustness without training group information
Standard training via empirical risk minimization (ERM) can produce models that achieve low
error on average but high error on minority groups, especially in the presence of spurious …
error on average but high error on minority groups, especially in the presence of spurious …
Wilds: A benchmark of in-the-wild distribution shifts
Distribution shifts—where the training distribution differs from the test distribution—can
substantially degrade the accuracy of machine learning (ML) systems deployed in the wild. …
substantially degrade the accuracy of machine learning (ML) systems deployed in the wild. …
Accuracy on the line: on the strong correlation between out-of-distribution and in-distribution generalization
For machine learning systems to be reliable, we must understand their performance in
unseen, out-of-distribution environments. In this paper, we empirically show that out-of-distribution …
unseen, out-of-distribution environments. In this paper, we empirically show that out-of-distribution …
Openflamingo: An open-source framework for training large autoregressive vision-language models
We introduce OpenFlamingo, a family of autoregressive vision-language models ranging
from 3B to 9B parameters. OpenFlamingo is an ongoing effort to produce an open-source …
from 3B to 9B parameters. OpenFlamingo is an ongoing effort to produce an open-source …
Extending the wilds benchmark for unsupervised adaptation
… The project was initiated by Shiori Sagawa, Pang Wei Koh, and Percy Liang. Shiori Sagawa
and Pang Wei Koh led the project and coordinated the activities below. Tony Lee developed …
and Pang Wei Koh led the project and coordinated the activities below. Tony Lee developed …
Out-of-domain robustness via targeted augmentations
Abstract Models trained on one set of domains often suffer performance drops on unseen
domains, eg, when wildlife monitoring models are deployed in new camera locations. In this …
domains, eg, when wildlife monitoring models are deployed in new camera locations. In this …