United States

Conversation forecasting tasks a model with predicting the outcome of an unfolding conversation. For instance, it can be applied in social media moderation to predict harmful user behaviors before they occur, allowing for preventative interventions. While large language models (LLMs) have recently been proposed as an effective tool for conversation forecasting, it&#39;s unclear what biases they may have, especially against forecasting the (potentially harmful) outcomes we request them to predict during moderation. This paper explores to what extent model uncertainty can be used as a tool to mitigate potential biases. Specifically, we ask three primary research questions: 1) how does LLM forecasting accuracy change when we ask models to represent their uncertainty; 2) how does LLM bias change when we ask models to represent their uncertainty; 3) how can we use uncertainty representations to reduce or completely mitigate biases without many training data points. We address these questions for 5 open-source language models tested on 2 datasets designed to evaluate conversation forecasting for social media moderation.

EMNLP 2024

Eliciting Uncertainty in Chain-of-Thought to Mitigate Bias against Forecasting Harmful User Behaviors

uncertainty; chain of thought; bias mitigation; social media moderation

### Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

Message from the General Chair: 
*It is my great pleasure and honor to welcome you to the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025), held in beautiful Vienna, Austria, from July 27 to August 1, 2025. ACL2025continues our field’s tradition of excellence in scholarship, innovation, and inclusivity, and I am deeply grateful to the many volunteers who have worked tirelessly to bring this event to life.* 
[Read more](https://drive.google.com/file/d/1GI_hvOpjswAuYdUTromfeDiPpCcqidwg/view?usp=sharing)

To access this event page, you need to log in with the **email address you registered with**. Access credentials will be sent to your email from Underline - subject line "Welcome to ACL 2025". Please be sure to check your spam email folder if you do not see an email confirmation right away.

Please log in

To access this event page, you need to log in with the **email address you registered with**. 

ACL 2025

Welcome to The 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2025)

Conversation forecasting tasks a model with predicting the outcome of an unfolding conversation. For instance, it can be applied in social media moderation to predict harmful user behaviors before they occur, allowing for preventative interventions. While large language models (LLMs) have recently been proposed as an effective tool for conversation forecasting, it's unclear what biases they may have, especially against forecasting the (potentially harmful) outcomes we request them to predict during moderation. This paper explores to what extent model uncertainty can be used as a tool to mitigate potential biases. Specifically, we ask three primary research questions: 1) how does LLM forecasting accuracy change when we ask models to represent their uncertainty; 2) how does LLM bias change when we ask models to represent their uncertainty; 3) how can we use uncertainty representations to reduce or completely mitigate biases without many training data points. We address these questions for 5 open-source language models tested on 2 datasets designed to evaluate conversation forecasting for social media moderation.

## Welcome to EMNLP 2024! 
We are excited to welcome you to one of the most prominent conferences in the field of Natural Language Processing. This year, EMNLP 2024 is being held in a hybrid format,
offering both virtual and in-person participation in beautiful Miami. Due to a record-breaking number of submissions, we've expanded the total number of accepted papers to accommodate more cutting-edge research from around the globe.
### [Conference Handbook](https://drive.google.com/file/d/1WPROgxjLAC96AJL7Ugy0tEnYm7dkrbHt/view?usp=sharing)

You are required to register for this event. **Please register [here](https://2024.emnlp.org/registration/).** The EMNLP 2024 event page on Underline will be open to public one week prior to the event.

Please register!

EMNLP 2024 will take place in Miami, Florida from Nov 12th to Nov 16th, 2024, at the Hyatt Regency Miami Hote and on Underline for remote participants.

Due to the recent rise in digital misinformation, there has been great interest shown in using LLMs for fact-checking and claim verification. In this paper, we answer the question: Do LLMs know multilingual facts and can they use this knowledge for effective fact-checking? To this end, we create a benchmark by filtering multilingual claims from the X-fact dataset and evaluating the multilingual fact-checking capabilities of five LLMs across five diverse languages: Spanish, Italian, Portuguese, Turkish, and Tamil on our benchmark. We employ three different prompting techniques: Zero-Shot, English Chain-of-Thought, and Cross-Lingual Prompting, using both greedy and self-consistency decoding. We extensively analyze our results and find that GPT-4o achieves the highest accuracy, but zero-shot prompting with self-consistency was the most effective overall. We also show that techniques like Chain-of-Thought and Cross-Lingual Prompting, which are designed to improve reasoning abilities, do not necessarily improve the fact-checking abilities of LLMs. Interestingly, we find a strong negative correlation between model accuracy and the amount of internet content for a given language. This suggests that LLMs are better at fact-checking from knowledge in low-resource languages. We hope that this study will encourage more work on multilingual fact-checking using LLMs.

Multilingual Fact-Checking using LLMs

Islamophobic language on online platforms fosters intolerance, making detection and elimination crucial for promoting harmony. Traditional hate speech detection models rely on NLP techniques like tokenization, part-of-speech tagging, and encoder-decoder models. However, Graph Neural Networks (GNNs), with their ability to utilize relationships between data points, offer more effective detection and greater explainability. In this work, we represent speeches as nodes and connect them with edges based on their context and similarity to develop the graph. This study introduces a novel paradigm using GNNs to identify and explain hate speech towards Islam. Our model leverages GNNs to understand the context and patterns of hate speech by connecting texts via pretrained NLP-generated word embeddings, achieving state-of-the-art performance and enhancing detection accuracy while providing valuable explanations. This highlights the potential of GNNs in combating online hate speech and fostering a safer, more inclusive online environment.

Explainable Identification of Hate Speech towards Islam using Graph Neural Networks

Epidemiological datasets are essential for public health analysis and decision-making, yet they remain scarce and often difficult to compile due to inconsistent data formats, language barriers, and evolving political boundaries. Traditional methods of creating such datasets involve extensive manual effort and are prone to errors in accurate location extraction. To address these challenges, we propose utilizing large language models (LLMs) to automate the extraction and geotagging of epidemiological data from textual documents. Our approach significantly reduces the manual effort required, limiting human intervention to validating a subset of records against text snippets and verifying the geotagging reasoning, as opposed to reviewing multiple entire documents manually to extract, clean, and geotag. Additionally, the LLMs identify information often overlooked by human annotators, further enhancing the dataset's completeness. Our findings demonstrate that LLMs can be effectively used to semi-automate the extraction and geotagging of epidemiological data, offering several key advantages: (1) comprehensive information extraction with minimal risk of missing critical details; (2) minimal human intervention; (3) higher-resolution data with more precise geotagging; and (4) significantly reduced resource demands compared to traditional methods.

From Text to Maps: LLM-Driven Extraction and Geotagging of Epidemiological Data

Training supervised machine learning systems with a fairness loss can improve prediction fairness across different demographic groups. However, doing so requires demographic annotations for training data, without which we cannot produce debiased classifiers for most tasks. Drawing inspiration from transfer learning methods, we investigate whether we can utilize demographic data from a related task to improve the fairness of a target task. We adapt a single-task fairness loss to a multi-task setting to exploit demographic labels from a related task in debiasing a target task, and demonstrate that demographic fairness objectives transfer fairness within a multi-task framework. Additionally, we show that this approach enables intersectional fairness by transferring between two datasets with different single-axis demographics. We explore different data domains to show how our loss can improve fairness domains and tasks.

Transferring Fairness using Multi-Task Learning with Limited Demographic Information

In the era dominated by information overload and its facilitation with Large Language Models (LLMs), the prevalence of misinformation poses a significant threat to public discourse and societal well-being. A critical concern at present involves the identification of machine-generated news. In this work, we take a significant step by introducing a benchmark dataset designed for neural news detection in four languages: English, Turkish, Hungarian, and Persian. The dataset incorporates outputs from multiple multilingual generators (in both, zero-shot and fine-tuned setups) such as BloomZ, LLaMa-2, Mistral, Mixtral, and GPT-4. Next, we experiment with a variety of classifiers, ranging from those based on linguistic features to advanced Transformer-based models and LLMs prompting. We present the detection results aiming to delve into the interpretablity and robustness of machine-generated texts detectors across all target languages.

Crafting Tomorrow's Headlines: Neural News Generation and Detection in English, Turkish, Hungarian, and Persian

Climate change (CC) has attracted increasing attention in NLP in recent years. However, detecting the stance on CC in multimodal data is understudied and remains challenging due to a lack of reliable datasets. To improve the understanding of public opinions and communication strategies, this paper presents MultiClimate, the first open-source manually-annotated stance detection dataset with $100$ CC-related YouTube videos and $4,209$ frame-transcript pairs. We deploy state-of-the-art vision and language models, as well as multimodal models for MultiClimate stance detection. Results show that text-only BERT significantly outperforms image-only ResNet50 and ViT. Combining both modalities achieves state-of-the-art, $0.747$/$0.749$ in accuracy/F1. Our 100M-sized fusion models also beat CLIP and BLIP, as well as the much larger 9B-sized multimodal IDEFICS and text-only Llama3 and Gemma2, indicating that multimodal stance detection remains challenging for large language models. Our code, dataset, as well as supplementary materials, are available at https://github.com/werywjw/MultiClimate.

MultiClimate: Multimodal Stance Detection on Climate Change Videos

Detecting biases in natural language understanding (NLU) for African American Vernacular English (AAVE) is crucial to developing inclusive natural language processing (NLP) systems. To address dialect-induced performance discrepancies, we introduce AAVENUE (AAVE Natural Language Understanding Evaluation), a benchmark for evaluating large language model (LLM) performance on NLU tasks in AAVE and Standard American English (SAE). AAVENUE builds upon and extends existing benchmarks like VALUE, replacing deterministic syntactic and morphological transformations with a more flexible methodology leveraging LLM-based translation with few-shot prompting, improving performance across our evaluation metrics when translating key tasks from the GLUE and SuperGLUE benchmarks. We compare AAVENUE and VALUE translations using five popular LLMs and a comprehensive set of metrics including fluency, BARTScore, quality, coherence, and understandability. Additionally, we recruit fluent AAVE speakers to validate our translations for authenticity. Our evaluations reveal that LLMs consistently perform better on SAE tasks than AAVE-translated versions, underscoring inherent biases and highlighting the need for more inclusive NLP models.

AAVENUE: Detecting LLM Biases on NLU Tasks in AAVE via a Novel Benchmark

As large language models (LLMs) gain traction in healthcare, concerns about their susceptibility to demographic biases are growing. We introduce DiversityMedQA, a novel benchmark designed to assess LLM responses to medical queries across diverse patient demographics, such as gender and ethnicity. By perturbing questions from the MedQA dataset, which comprises of medical board exam questions, we created a benchmark that captures the nuanced differences in medical diagnosis across varying patient profiles. To ensure that our perturbations did not alter the clinical outcomes, we implemented a filtering strategy to validate each perturbation, so that any performance discrepancies would be indicative of bias. Our findings reveal notable discrepancies in model performance when tested against these demographic variations. By releasing DiversityMedQA, we provide a resource for evaluating and mitigating demographic bias in LLM medical diagnoses.

DiversityMedQA: A Benchmark for Assessing Demographic Biases in Medical Diagnosis using Large Language Models

In this paper, we propose a novel application to improve industrial safety by generating preventive recommendations using LLMs. Using a dataset of 275 incidents representing 11 different incident types sampled from real-life OSHA incidents, we compare three different LLMs to evaluate the quality of preventive recommendations generated by them. We also show that LLMs are not a panacea for the preventive recommendation generation task. They have limitations and can produce responses that are incorrect or irrelevant. We found that about 65% of the output from Vicuna model was not acceptable at all at the basic readability and other sanity checks level. Mistral and Phi_3 are better than Vicuna, but not all of their recommendations are of similar quality. We find that for a given safety incident case, the generated recommendations can be categorized as specific, generic, or irrelevant. This helps us to better quantify and compare the performance of the models. This paper is among the initial and novel work for the preventive recommendation generation problem. We believe it will pave way for use of NLP to positively impact the industrial safety.

Improving Industrial Safety by Auto-Generating Case-specific Preventive Recommendations

Recently, work in NLP has shifted to few-shot (in-context) learning, with large language models (LLMs) performing well across a range of tasks. However, while fairness evaluations have become a standard for supervised methods, little is known about the fairness of LLMs as prediction systems. Further, common standard methods for fairness involve access to model weights or are applied during finetuning, which are not applicable in few-shot learning. Do LLMs exhibit prediction biases when used for standard NLP tasks?

In this work, we analyze the effect of shots, which directly affect the performance of models, on the fairness of LLMs as NLP classification systems. We consider how different shot selection strategies, both existing and new demographically sensitive methods, affect model fairness across three standard fairness datasets. We find that overall the performance of LLMs is not indicative of their fairness, and there is not a single method that fits all scenarios. In light of these facts, we discuss how future work can include LLM fairness in evaluations.

Downloads

Next from EMNLP 2024

Multilingual Fact-Checking using LLMs

Similar lecture

Evaluating Theory of (an uncertain) Mind: Predicting the Uncertain Beliefs of Others from Conversational Cues

.css-70qvj9{display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}Downloads

Next from EMNLP 2024

Multilingual Fact-Checking using LLMs

Similar lecture

Evaluating Theory of (an uncertain) Mind: Predicting the Uncertain Beliefs of Others from Conversational Cues

Downloads