Publications

2024

1. CCF BigData

   Evaluation of Retrieval-Augmented Generation: A Survey

   Hao Yu, Aoran Gan, Kai Zhang, Shiwei Tong, Qi Liu, and Zhaofeng Liu

   Jun 2024

   Abs arXiv DOI

   Retrieval-Augmented Generation (RAG) has recently gained traction in natural language processing. Numerous studies and real-world applications are leveraging its ability to enhance generative models through external information retrieval. Evaluating these RAG systems, however, poses unique challenges due to their hybrid structure and reliance on dynamic knowledge sources. To better understand these challenges, we conduct A Unified Evaluation Process of RAG (Auepora) and aim to provide a comprehensive overview of the evaluation and benchmarks of RAG systems. Specifically, we examine and compare several quantifiable metrics of the Retrieval and Generation components, such as relevance, accuracy, and faithfulness, within the current RAG benchmarks, encompassing the possible output and ground truth pairs. We then analyze the various datasets and metrics, discuss the limitations of current benchmarks, and suggest potential directions to advance the field of RAG benchmarks.

2. COLM 2024

   Web Retrieval Agents for Evidence-Based Misinformation Detection

   Jacob-Junqi Tian, Hao Yu, Yury Orlovskiy, Mauricio Rivera, Zachary Yang, Jean-François Godbout, Reihaneh Rabbany, and Kellin Pelrine

   Jun 2024

   Abs

   This paper develops an agent-based automated fact-checking approach for detecting misinformation. We demonstrate that combining a powerful LLM agent, which does not have access to the internet for searches, with an online web search agent yields better results than when each tool is used independently. Our approach is robust across multiple models, outperforming alternatives and increasing the macro F1 of misinformation detection by as much as 20 percent compared to LLMs without search. We also conduct extensive analyses on the sources our system leverages and their biases, decisions in the construction of the system like the search tool and the knowledge base, the type of evidence needed and its impact on the results, and other parts of the overall process. By combining strong performance with in-depth understanding, we hope to provide building blocks for future search-enabled misinformation mitigation systems.

3. LREC EURALI

   An Evaluation of Language Models for Hyperpartisan Ideology Detection in Persian Twitter

   Sahar Omidi Shayegan, Isar Nejadgholi, Kellin Pelrine, Hao Yu, Sacha Levy, Zachary Yang, Jean-François Godbout, and Reihaneh Rabbany

   2nd Workshop on Resources and Technologies for Indigenous, Endangered and Lesser-resourced Languages in Eurasia @ LREC-COLING 2024, Jun 2024

   Abs

   Large Language Models (LLMs) are now capable of successfully identifying the political beliefs of English-speaking social media users from their posts. However, assessing how LLMs perform in non-English languages remains difficult. In this work, we contribute to this area of research by determining the extent to which LLMs can predict the political ideologies of users on Persian social media. We begin by discussing the challenges associated with defining political parties within the Persian context and propose a solution based on a technique designed for the detection of hyper-partisan ideologies on social media. We create a new benchmark and show the potential and limitations of both open-source and commercial LLMs in classifying the hyper-partisan ideologies of users. We compare these models with smaller fine-tuned ones, both on the Persian language (ParsBERT) and translated data (RoBERTa), and confirm that they considerably outperform generative LLMs in this task. We further demonstrate that the performance of the generative LLMs degrades when classifying users based on their tweets instead of their bios, even if tweets are added as additional information; whereas the smaller fine-tuned models are more robust and achieve similar performance for all input settings. This study represents a first step toward political ideology detection in Persian social media, with implications for future research to understand the dynamics of political conflicts in Iran.

2023

1. arXiv

   Open, Closed, or Small Language Models for Text Classification?

   Hao Yu*, Zachary Yang*, Kellin Pelrine, Jean Francois Godbout, and Reihaneh Rabbany

   arXiv preprint arXiv:2308.10092, Jun 2023

   Abs arXiv

   Recent advancements in large language models have demonstrated remarkable capabilities across various NLP tasks. But many questions remain, including whether open-source models match closed ones, why these models excel or struggle with certain tasks, and what types of practical procedures can improve performance. We address these questions in the context of classification by evaluating three classes of models using eight datasets across three distinct tasks: named entity recognition, political party prediction, and misinformation detection. While larger LLMs often lead to improved performance, open-source models can rival their closed-source counterparts by fine-tuning. Moreover, supervised smaller models, like RoBERTa, can achieve similar or even greater performance in many datasets compared to generative LLMs. On the other hand, closed models maintain an advantage in hard tasks that demand the most generalizability. This study underscores the importance of model selection based on task requirements.

2. EMNLP 2023

   SWEET - Weakly Supervised Person Name Extraction for Fighting Human Trafficking

   Javin Liu*, Hao Yu*, Vidya Sujaya*, Pratheeksha Nair, Kellin Pelrine, and Reihaneh Rabbany

   In Findings of the Association for Computational Linguistics: EMNLP 2023, Dec 2023

   Abs DOI

   In this work, we propose a weak supervision pipeline SWEET: Supervise Weakly for Entity Extraction to fight Trafficking for extracting person names from noisy escort advertisements. Our method combines the simplicity of rule-matching (through antirules, i.e., negated rules) and the generalizability of large language models fine-tuned on benchmark, domain-specific and synthetic datasets, treating them as weak labels. One of the major challenges in this domain is limited labeled data. SWEET addresses this by obtaining multiple weak labels through labeling functions and effectively aggregating them. SWEET outperforms the previous supervised SOTA method for this task by 9% F1 score on domain data and better generalizes to common benchmark datasets. Furthermore, we also release HTGEN, a synthetically generated dataset of escort advertisements (built using ChatGPT) to facilitate further research within the community.

3. Quantum

   TensorCircuit: a Quantum Software Framework for the NISQ Era

   Shi-Xin Zhang, Jonathan Allcock, Zhou-Quan Wan, Shuo Liu, Jiace Sun, Hao Yu, Xing-Han Yang, Jiezhong Qiu, Zhaofeng Ye, Yu-Qin Chen, Chee-Kong Lee, Yi-Cong Zheng, Shao-Kai Jian, Hong Yao, Chang-Yu Hsieh, and Shengyu Zhang

   Quantum, Feb 2023

   DOI