2020/12/08 by Nishtha Madaan, Inkit Padhi, Madaan, Nishtha +5 · 7 citations
Computer Science · #Artificial Intelligence (cs.AI) #Computation and Language (cs.CL) #Explainable Artificial Intelligence (XAI) #FOS: Computer and information sciences #Imbalanced Data Classification Techniques #Machine Learning (cs.LG) #Software Engineering Research #Topic Modeling
paper · pdf · doi:10.48550/arxiv.2012.04698
openalex publication_date 2020/12/08 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Machine Learning has seen tremendous growth recently, which has led to larger\nadoption of ML systems for educational assessments, credit risk, healthcare,\nemployment, criminal justice, to name a few. The trustworthiness of ML and NLP\nsystems is a crucial aspect and requires a guarantee that the decisions they\nmake are fair and robust. Aligned with this, we propose a framework GYC, to\ngenerate a set of counterfactual text samples, which are crucial for testing\nthese ML systems. Our main contributions include a) We introduce GYC, a\nframework to generate counterfactual samples such that the generation is\nplausible, diverse, goal-oriented, and effective, b) We generate counterfactual\nsamples, that can direct the generation towards a corresponding condition such\nas named-entity tag, semantic role label, or sentiment. Our experimental\nresults on various domains show that GYC generates counterfactual text samples\nexhibiting the above four properties. GYC generates counterfactuals that can\nact as test cases to evaluate a model and any text debiasing algorithm.\n