SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2005.00699 · ACL · 2020

Gender Bias in Multilingual Embeddings and Cross-Lingual Transfer

Kai-Wei Chang, Subhabrata Mukherjee, Ahmed Awadallah, Saghar Hosseini, Jieyu Zhao

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 12 functions out of this paper's own repositories and ran 8 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
msr-lit/multilingualbias pwc_unofficial 8 of 12
FunctionStatusWhere it lives
dedup_middle Ran msr-lit/multilingualbias/extrinsic/replicateMLBs/DE/preprocess_de.py
pointer only (licence: MIT) · get_code("34530fbdde2cb8fe")
get_sequences_lengths Ran msr-lit/multilingualbias/extrinsic/bios_codes/utils.py
pointer only (licence: MIT) · get_code("d308860a48dbe40d")
infer_gender Ran msr-lit/multilingualbias/extrinsic/replicateMLBs/DE/down_de_bios.py
pointer only (licence: MIT) · get_code("c2394da354f9fe76")
infer_gender Ran msr-lit/multilingualbias/extrinsic/replicateMLBs/ES/download_es_bios.py
pointer only (licence: MIT) · get_code("ee32fe610b551817")
infer_gender Ran msr-lit/multilingualbias/extrinsic/replicateMLBs/FR/download_fr_bios.py
pointer only (licence: MIT) · get_code("be8c2c987a279534")
parse_args Ran msr-lit/multilingualbias/extrinsic/bios_codes/config.py
pointer only (licence: MIT) · get_code("afbe2e1ae156e07e")
softmax_masked Ran msr-lit/multilingualbias/extrinsic/bios_codes/utils.py
pointer only (licence: MIT) · get_code("58ffc257d9f94fd1")
to_device Ran msr-lit/multilingualbias/extrinsic/bios_codes/utils.py
pointer only (licence: MIT) · get_code("6a750bc656206858")
extract_name Not yet run msr-lit/multilingualbias/extrinsic/replicateMLBs/ES/download_es_bios.py
pointer only (licence: MIT) · get_code("5bc8cc37fb144264")
group_by Not yet run msr-lit/multilingualbias/extrinsic/replicateMLBs/DE/preprocess_de.py
pointer only (licence: MIT) · get_code("a726daef874a3d4e")
load_pkl Not yet run msr-lit/multilingualbias/extrinsic/replicateMLBs/DE/preprocess_de.py
pointer only (licence: MIT) · get_code("019de8c1d879acd4")
re_escape_title Not yet run msr-lit/multilingualbias/extrinsic/replicateMLBs/DE/down_de_bios.py
pointer only (licence: MIT) · get_code("f7b10b80fcb10ccf")

Repositories linked to this paper

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

Abstract

Multilingual representations embed words from many languages into a single semantic space such that words with similar meanings are close to each other regardless of the language. These embeddings have been widely used in various settings, such as cross-lingual transfer, where a natural language processing (NLP) model trained on one language is deployed to another language. While the crosslingual transfer techniques are powerful, they carry gender bias from the source to target languages. In this paper, we study gender bias in multilingual embeddings and how it affects transfer learning for NLP applications. We create a multilingual dataset for bias analysis and propose several ways for quantifying bias in multilingual representations from both the intrinsic and extrinsic perspectives. Experimental results show that the magnitude of bias in the multilingual representations changes differently when we align the embeddings to different target spaces and that the alignment direction can also have an influence on the bias in transfer learning. We further provide recommendations for using the multilingual word representations for downstream tasks.

For agents

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

get_harvested_code_for_paper("2005.00699")
get_code_for_paper("2005.00699")
have("2005.00699")

Connect an agent — have() is free.