Search for: All records

This paper investigates the performance of a diverse set of large language models (LLMs) including leading closed-source (GPT-4, GPT-4o mini, Claude 3.5 Haiku) and open-source (Llama 3.1 70B, Llama 3.1 8B) models, alongside the earlier GPT-3.5 within the context of U.S. tax resolutions. AI-driven solutions like these have made substantial inroads into legal-critical systems with significant socio-economic implications. However, their accuracy and reliability have not been assessed in some legal domains, such as tax. Using the Volunteer Income Tax Assistance (VITA) certification tests—endorsed by the US Internal Revenue Service (IRS) for tax volunteering—this study compares these LLMs to evaluate their potential utility in assisting both tax volunteers as well as taxpayers, particularly those with low and moderate income. Since the answers to these questions are not publicly available, we first analyze 130 questions with the tax domain experts and develop the ground truths for each question. We then benchmarked these diverse LLMs against the ground truths using both the original VITA questions and syntactically perturbed versions (a total of 390 questions) to assess genuine understanding versus memorization/hallucinations. Our comparative analysis reveals distinct performance differences: closed-source models (GPT-4, Claude 3.5 Haiku, GPT-4o mini) generally demonstrated higher accuracy and robustness compared to GPT-3.5 and the open-source Llama models. For instance, on basic multiple-choice questions, top models like GPT-4 and Claude 3.5 Haiku achieved 83.33% accuracy, surpassing GPT-3.5 (54.17%) and the open-source Llama 3.1 8B (50.00%). These findings generally hold across both original and perturbed questions. However, the paper acknowledges that these developments are initial indicators, and further research is necessary to fully understand the implications of deploying LLMs in this domain. A critical limitation observed across all evaluated models was significant difficulty with open-ended questions, which require accurate numerical calculation and application of tax rules. We hope that this paper provides a means and a standard to evaluate the efficacy of current and future LLMs in the tax domain. 

Data-driven software is increasingly being used as a critical component of automated decision-support systems. Since this class of software learns its logic from historical data, it can encode or amplify discriminatory practices. Previous research on algorithmic fairness has focused on improving “average-case” fairness. On the other hand, fairness at the extreme ends of the spectrum, which often signifies lasting and impactful shifts in societal attitudes, has received significantly less emphasis. Leveraging the statistics of extreme value theory (EVT), we propose a novel fairness criterion called extreme counterfactual discrimination (ECD). This criterion estimates the worst-case amounts of disadvantage in outcomes for individuals solely based on their memberships in a protected group. Utilizing tools from search-based software engineering and generative AI, we present a randomized algorithm that samples a statistically significant set of points from the tail of ML outcome distributions even if the input dataset lacks a sufficient number of relevant samples. We conducted several experiments on four ML models (deep neural networks, logistic regression, and random forests) over 10 socially relevant tasks from the literature on algorithmic fairness. First, we evaluate the generative AI methods and find that they generate sufficient samples to infer valid EVT distribution in 95% of cases. Remarkably, we found that the prevalent bias mitigators reduce the average-case discrimination but increase the worst-case discrimination significantly in 35% of cases. We also observed that even the tail-aware mitigation algorithm—MiniMax-Fairness—increased the worst-case discrimination in 30% of cases. We propose a novel ECD-based mitigator that improves fairness in the tail in 90% of cases with no degradation of the average-case discrimination. We hope that the EVT framework serves as a robust tool for evaluating fairness in both average-case and worst-case discrimination. 

As the US tax law evolves to adapt to ever-changing politico-economic realities, tax preparation software plays a significant role in helping taxpayers navigate these complexities. The dynamic nature of tax regulations poses a significant challenge to accurately and timely maintaining tax software artifacts. The state-of-the-art in maintaining tax prep software is time-consuming and error-prone as it involves manual code analysis combined with an expert interpretation of tax law amendments. We posit that the rigor and formality of tax amendment language, as expressed in IRS publications, makes it amenable to automatic translation to executable specifications (code). Our research efforts focus on identifying, understanding, and tackling technical challenges in leveraging Large Language Models (LLMs), such as ChatGPT and Llama, to faithfully extract code differentials from IRS publications and automatically integrate them with the prior version of the code to automate tax prep software maintenance. 

Not AvailableMachine learning (ML) is increasingly used in high-stakes areas like autonomous driving, finance, and criminal justice. However, it often unintentionally perpetuates biases against marginalized groups. To address this, the software engineering community has developed fairness testing and debugging methods, establishing best practices for fair ML software. These practices focus on training model design, including the selection of sensitive and non-sensitive attributes and hyperparameter configuration. However, the application of these practices across different socio-economic and cultural contexts is challenging, as societal constraints vary. Our study proposes a search-based software engineering approach to evaluate the robustness of these fairness practices. We formulate these practices as the first-order logic properties and search for two neighborhood datasets where the practice satisfies in one dataset, but fail in the other one. Our key observation is that these practices should be general and robust to various uncertainty such as noise, faulty labeling, and demographic shifts. To generate datasets, we sift to the causal graph representations of datasets and apply perturbations over the causal graphs to generate neighborhood datasets. In this short paper, we show our methodology using an example of predicting risks in the car insurance application. 

Due to the ever-increasing complexity of in- come tax laws in the United States, the num- ber of US taxpayers filing their taxes using tax preparation software (henceforth, tax soft- ware) continues to increase. According to the U.S. Internal Revenue Service (IRS), in FY22, nearly 50% of taxpayers filed their individual income taxes using tax software. Given the legal consequences of incorrectly filing taxes for the taxpayer, ensuring the correctness of tax software is of paramount importance. Meta- morphic testing has emerged as a leading solu- tion to test and debug legal-critical tax software due to the absence of correctness requirements and trustworthy datasets. The key idea behind metamorphic testing is to express the proper- ties of a system in terms of the relationship between one input and its slightly metamor- phosed twinned input. Extracting metamor- phic properties from IRS tax publications is a tedious and time-consuming process. As a response, this paper formulates the task of gen- erating metamorphic specifications as a transla- tion task between properties extracted from tax documents—expressed in natural language—to a contrastive first-order logic form. We per- form a systematic analysis on the potential and limitations of in-context learning with Large Language Models (LLMs) for this task, and outline a research agenda towards automating the generation of metamorphic specifications for tax preparation software.