[ { "id": "authors:e65gr-rdk79", "collection": "authors", "collection_id": "e65gr-rdk79", "cite_using_url": "https://resolver.caltech.edu/CaltechAUTHORS:20230316-153717662", "type": "monograph", "title": "AutoBiasTest: Controllable Sentence Generation for Automated and Open-Ended Social Bias Testing in Language Models", "author": [ { "family_name": "Kocielnik", "given_name": "Rafal", "orcid": "0000-0001-5602-6056", "clpid": "Kocielnik-Rafal" }, { "family_name": "Prabhumoye", "given_name": "Shrimai", "clpid": "Prabhumoye-Shrimai" }, { "family_name": "Zhang", "given_name": "Vivian", "clpid": "Zhang-Vivian" }, { "family_name": "Alvarez", "given_name": "R. Michael", "orcid": "0000-0002-8113-4451", "clpid": "Alvarez-R-M" }, { "family_name": "Anandkumar", "given_name": "Anima", "orcid": "0000-0002-6974-6797", "clpid": "Anandkumar-A" } ], "abstract": "Social bias in Pretrained Language Models (PLMs) affects text generation and other downstream NLP tasks. Existing bias testing methods rely predominantly on manual templates or on expensive crowd-sourced data. We propose a novel AutoBiasTest method that automatically generates sentences for testing bias in PLMs, hence providing a flexible and low-cost alternative. Our approach uses another PLM for generation and controls the generation of sentences by conditioning on social group and attribute terms. We show that generated sentences are natural and similar to human-produced content in terms of word length and diversity. We illustrate that larger models used for generation produce estimates of social bias with lower variance. We find that our bias scores are well correlated with manual templates, but AutoBiasTest highlights biases not captured by these templates due to more diverse and realistic test sentences. By automating large-scale test sentence generation, we enable better estimation of underlying bias distributions.", "publisher": "arXiv", "publication_date": "2023-02-14" }, { "id": "authors:7k8cy-29x80", "collection": "authors", "collection_id": "7k8cy-29x80", "cite_using_url": "https://resolver.caltech.edu/CaltechAUTHORS:20221221-004733367", "type": "monograph", "title": "Can You Label Less by Using Out-of-Domain Data? Active & Transfer Learning with Few-shot Instructions", "author": [ { "family_name": "Kocielnik", "given_name": "Rafal", "orcid": "0000-0001-5602-6056", "clpid": "Kocielnik-Rafal" }, { "family_name": "Kangaslahti", "given_name": "Sara", "clpid": "Kangaslahti-Sara" }, { "family_name": "Prabhumoye", "given_name": "Shrimai", "clpid": "Prabhumoye-Shrimal" }, { "family_name": "Hari", "given_name": "Meena", "clpid": "Hari-Meena" }, { "family_name": "Alvarez", "given_name": "R. Michael", "orcid": "0000-0002-8113-4451", "clpid": "Alvarez-R-M" }, { "family_name": "Anandkumar", "given_name": "Anima", "orcid": "0000-0002-6974-6797", "clpid": "Anandkumar-A" } ], "abstract": "Labeling social-media data for custom dimensions of toxicity and social bias is challenging and labor-intensive. Existing transfer and active learning approaches meant to reduce annotation effort require fine-tuning, which suffers from over-fitting to noise and can cause domain shift with small sample sizes. In this work, we propose a novel Active Transfer Few-shot Instructions (ATF) approach which requires no fine-tuning. ATF leverages the internal linguistic knowledge of pre-trained language models (PLMs) to facilitate the transfer of information from existing pre-labeled datasets (source-domain task) with minimum labeling effort on unlabeled target data (target-domain task). Our strategy can yield positive transfer achieving a mean AUC gain of 10.5% compared to no transfer with a large 22b parameter PLM. We further show that annotation of just a few target-domain samples via active learning can be beneficial for transfer, but the impact diminishes with more annotation effort (26% drop in gain between 100 and 2000 annotated examples). Finally, we find that not all transfer scenarios yield a positive gain, which seems related to the PLMs initial performance on the target-domain task.", "doi": "10.48550/arXiv.2211.11798", "publisher": "arXiv", "publication_date": "2022-11-21" }, { "id": "authors:7552k-s1m43", "collection": "authors", "collection_id": "7552k-s1m43", "cite_using_url": "https://resolver.caltech.edu/CaltechAUTHORS:20220714-224624824", "type": "monograph", "title": "Few-shot Instruction Prompts for Pretrained Language Models to Detect Social Biases", "author": [ { "family_name": "Prabhumoye", "given_name": "Shrimai", "clpid": "Prabhumoye-Shrimai" }, { "family_name": "Kocielnik", "given_name": "Rafal", "clpid": "Kocielnik-Rafal" }, { "family_name": "Shoeybi", "given_name": "Mohammad", "clpid": "Shoeybi-Mohammad" }, { "family_name": "Anandkumar", "given_name": "Anima", "orcid": "0000-0002-6974-6797", "clpid": "Anandkumar-A" }, { "family_name": "Catanzaro", "given_name": "Bryan", "clpid": "Catanzaro-Bryan" } ], "abstract": "Detecting social bias in text is challenging due to nuance, subjectivity, and difficulty in obtaining good quality labeled datasets at scale, especially given the evolving nature of social biases and society. To address these challenges, we propose a few-shot instruction-based method for prompting pre-trained language models (LMs). We select a few class-balanced exemplars from a small support repository that are closest to the query to be labeled in the embedding space. We then provide the LM with instruction that consists of this subset of labeled exemplars, the query text to be classified, a definition of bias, and prompt it to make a decision. We demonstrate that large LMs used in a few-shot context can detect different types of fine-grained biases with similar and sometimes superior accuracy to fine-tuned models. We observe that the largest 530B parameter model is significantly more effective in detecting social bias compared to smaller models (achieving at least 13% improvement in AUC metric compared to other models). It also maintains a high AUC (dropping less than 2%) when the labeled repository is reduced to as few as 100 samples. Large pretrained language models thus make it easier and quicker to build new bias detectors.", "doi": "10.48550/arXiv.arXiv.2112.07868", "publisher": "arXiv", "publication_date": "2021-12-15" } ]