SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2210.14011 · EMNLP · 2022

Are All Spurious Features in Natural Language Alike? An Analysis through a Causal Lens

He He, Nitish Joshi, Xiang Pan

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 7 functions out of this paper's own repositories and ran 5 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
joshinh/spurious-correlations-nlp canonical 5 of 7
FunctionStatusWhere it lives
build_probe Ran joshinh/spurious-correlations-nlp/nlp/scripts/probing.py
code served (permissive licence) · get_code("83b8335aa82f90c4")
compute_metrics_default Ran joshinh/spurious-correlations-nlp/nlp/scripts/training.py
code served (permissive licence) · get_code("323fa71e27b117eb")
get_projection_to_intersection_of_nullspaces Ran joshinh/spurious-correlations-nlp/nlp/scripts/debias.py
code served (permissive licence) · get_code("6603d34965799b5d")
get_rowspace_projection Ran joshinh/spurious-correlations-nlp/nlp/scripts/debias.py
code served (permissive licence) · get_code("109a64c49d7364c5")
per_class_accuracy_with_names Ran joshinh/spurious-correlations-nlp/nlp/scripts/training.py
code served (permissive licence) · get_code("02436469ae12f271")
debias_by_specific_directions Not yet run joshinh/spurious-correlations-nlp/nlp/scripts/debias.py
code served (permissive licence) · get_code("3b1da49bcea7e00a")
paired_accuracy Not yet run joshinh/spurious-correlations-nlp/nlp/scripts/training.py
code served (permissive licence) · get_code("b2a4cae875a44fdc")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

The term 'spurious correlations' has been used in NLP to informally denote any undesirable feature-label correlations. However, a correlation can be undesirable because (i) the feature is irrelevant to the label (e.g. punctuation in a review), or (ii) the feature's effect on the label depends on the context (e.g. negation words in a review), which is ubiquitous in language tasks. In case (i), we want the model to be invariant to the feature, which is neither necessary nor sufficient for prediction. But in case (ii), even an ideal model (e.g. humans) must rely on the feature, since it is necessary (but not sufficient) for prediction. Therefore, a more fine-grained treatment of spurious features is needed to specify the desired model behavior. We formalize this distinction using a causal model and probabilities of necessity and sufficiency, which delineates the causal relations between a feature and a label. We then show that this distinction helps explain results of existing debiasing methods on different spurious features, and demystifies surprising results such as the encoding of spurious features in model representations after debiasing.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2210.14011")
get_code_for_paper("2210.14011")
have("2210.14011")

Connect an agent — have() is free.