SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1904.03310 · 2019

Gender Bias in Contextualized Word Embeddings

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 12 functions out of this paper's own repositories and ran 10 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
dadebias/genda-lens pwc_unofficial 10 of 12
FunctionStatusWhere it lives
get_coref_predictions Ran dadebias/genda-lens/genda_lens/coref_tasks/abc_utils.py
code served (permissive licence) · get_code("00d3520c5093929f")
idx_pron Ran dadebias/genda-lens/genda_lens/lm_tasks/wino_utils.py
code served (permissive licence) · get_code("5019b056504a4da4")
load_sents Ran dadebias/genda-lens/genda_lens/lm_tasks/abc_utils.py
code served (permissive licence) · get_code("2ffc9b1a540aa3c7")
load_texts Ran dadebias/genda-lens/genda_lens/coref_tasks/wino_utils.py
code served (permissive licence) · get_code("a3c6c64ab86bc773")
load_texts Ran dadebias/genda-lens/genda_lens/lm_tasks/wino_utils.py
code served (permissive licence) · get_code("3f30ce652b7903ea")
logratio Ran dadebias/genda-lens/genda_lens/coref_tasks/wino_utils.py
code served (permissive licence) · get_code("d63e61c158ee7c53")
remove_duplicates Ran dadebias/genda-lens/genda_lens/ner_tasks/process_names.py
code served (permissive licence) · get_code("fe6f242ddb1d59bb")
remove_sq_br Ran dadebias/genda-lens/genda_lens/coref_tasks/wino_utils.py
code served (permissive licence) · get_code("82a5a069748bdc70")
remove_sq_br Ran dadebias/genda-lens/genda_lens/lm_tasks/wino_utils.py
code served (permissive licence) · get_code("4f9c8f00673b995a")
tokenize_sentence Ran dadebias/genda-lens/genda_lens/lm_tasks/abc_utils.py
code served (permissive licence) · get_code("754ad37cfcc6370c")
load_mdl Not yet run dadebias/genda-lens/genda_lens/lm_tasks/load_model.py
code served (permissive licence) · get_code("1b4325216f88a7d3")
load_mdl Not yet run dadebias/genda-lens/genda_lens/lm_tasks/abc_utils.py
code served (permissive licence) · get_code("e77d690f3880683f")

Repositories linked to this paper

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

Abstract

In this paper, we quantify, analyze and mitigate gender bias exhibited in ELMo's contextualized word vectors. First, we conduct several intrinsic analyses and find that (1) training data for ELMo contains significantly more male than female entities, (2) the trained ELMo embeddings systematically encode gender information and (3) ELMo unequally encodes gender information about male and female entities. Then, we show that a state-of-the-art coreference system that depends on ELMo inherits its bias and demonstrates significant bias on the WinoBias probing corpus. Finally, we explore two methods to mitigate such gender bias and show that the bias demonstrated on WinoBias can be eliminated.

For agents

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

get_harvested_code_for_paper("1904.03310")
get_code_for_paper("1904.03310")
have("1904.03310")

Connect an agent — have() is free.