While Large Language Models (LLMs) have been increasingly used to generate synthetic data for various downstream tasks, researchers overlook the causal direction in the data synthesis process. A natural causal direction should contain two steps: diverse raw data are generated first, and subsequently annotated for downstream tasks. However, most LLM-based methods adopt an anticausal direction: embedding label information in the prompt to force LLMs to generate targeted data. This reversal raises a critical question: How does the direction of data synthesis impact the quality and utility of the synthetic data? In this work, we empirically study the impact of causal and anticausal data synthesis. To do so, we first design simple yet effective prompting strategies to control the causal direction of LLM-based data synthesis. Using GPT-5 as the data generator, we construct synthetic datasets for three distinct machine learning tasks. We then fine-tune BERT-base and LLaMA-3.2-1B models on these datasets and evaluate them against human-curated benchmarks. Our experiments reveal consistent patterns: (1) models trained on anticausal synthetic data suffer larger performance drops across all tasks and model families --- Accuracy declines range from 13.7%-59.1% for BERT and 4.9%-54.3% for LLaMA, and (2) distributional analysis shows that anticausal synthetic datasets deviate further from human data. Our findings provide practical guidance on how to generate better synthetic data and make good use of it.

Traditional disaster analysis and modelling tools for assessing the severity of a disaster are predictive in nature. Based on the past observational data, these tools prescribe how the current input state (e.g., environmental conditions, situation reports) results in a severity assessment. However, these systems are not meant to be interventional in the causal sense, where the user can modify the current input state to simulate counterfactual ''what-if'' scenarios. In this work, we provide an alternative interventional tool that complements traditional disaster modelling tools by leveraging real-time data sources like satellite imagery, news, and social media. Our tool also helps understand the causal attribution of different factors on the estimated severity, over any given region of interest. In addition, we provide actionable recourses that would enable easier mitigation planning. Our source code is publicly available.

Dawei Li, Bohan Jiang, Liangjie Huang, Alimohammad Beigi, Chengshuai Zhao, Zhen Tan, Amrita Bhattacharjee, Yuxuan Jiang, Canyu Chen, Tianhao Wu, Kai Shu, Lu Cheng, Huan Liu. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing. 2025.

Affective polarization, the emotional divide between ideological groups marked by in-group love and out-group hate, has intensified in the United States, driving contentious issues like masking and lockdowns during the COVID-19 pandemic. Despite its societal impact, existing models of opinion change fail to account for emotional dynamics nor offer methods to quantify affective polarization robustly and in real-time. In this paper, we introduce a discrete choice model that captures decision-making within affectively polarized social networks and propose a statistical inference method estimate key parameters---in-group love and out-group hate---from social media data. Through empirical validation from online discussions about the COVID-19 pandemic, we demonstrate that our approach accurately captures real-world polarization dynamics and explains the rapid emergence of a partisan gap in attitudes towards masking and lockdowns. This framework allows for tracking affective polarization across contentious issues has broad implications for fostering constructive online dialogues in digital spaces.

Visual-textual retrieval, as a link between computer vision and natural language processing, aims at jointly learning visual-semantic relevance to bridge the heterogeneity gap across visual and textual spaces. Existing methods conduct retrieval only relying on the ranking of pairwise similarities, but they cannot self-evaluate the uncertainty of retrieved results, resulting in unreliable retrieval and hindering interpretability. To address this problem, we propose a novel Trust-Consistent Learning framework (TCL) to endow visual-textual retrieval with uncertainty evaluation for trustworthy retrieval. More specifically, TCL first models the matching evidence according to cross-modal similarity to estimate the uncertainty for cross-modal uncertainty-aware learning. Second, a simple yet effective consistency module is presented to enforce the subjective opinions of bidirectional learning to be consistent for high reliability and accuracy. Finally, extensive experiments are conducted to demonstrate the superiority and generalizability of TCL on six widely-used benchmark datasets, i.e., Flickr30K, MS-COCO, MSVD, MSR-VTT, ActivityNet, and DiDeMo. Furthermore, some qualitative experiments are carried out to provide comprehensive and insightful analyses for trustworthy visual-textual retrieval, verifying the reliability and interoperability of TCL. The code is available in https://github.com/QinYang79/TCL.

We present a modular pipeline for summarizing broadcast news videos using large language and vision models, specifically integrating Whisper for ASR, TransNetV2 for shot segmentation, LLaVA for image captioning, and LLaMA for generating structured summaries. Implemented within the CLAMS platform using the Multimedia Interchange Format (MMIF) for component interoperability, our approach combines ASR transcriptions and image captions to enhance metadata extraction. We evaluated our pipeline with automated metrics based on user-generated Youtube video descriptons as well as human assessments. Our analysis highlights challenges with automated metrics and emphasizes the value of human evaluation for nuanced assessment. This work demonstrates the effectiveness of multimodal summarization for video metadata extraction and paves the way for enhanced video accessibility.

News media has been frequently utilized as a political tool to stray from facts, making biased statements and claims without evidence. During the COVID-19 vaccine campaign, politically biased news (PBN) has significantly undermined public trust in vaccines. Despite medical evidence showing the benefits of these vaccines, the misperceptions of the vaccine's safety, risks, and efficacy have led to a non-negligible fraction of the population resistant to receiving the vaccine. In this paper, we analyze: (i) how inherent vaccine stances subtly influence individuals' selection of news sources and participation in social media discussions; and (ii) the impact of exposure to PBN on users' attitudes toward vaccines. In doing so, we first curate a comprehensive dataset that connects PBN with related social media discourse. Utilizing advanced deep learning and causal inference techniques, we reveal distinct user behaviors between social media groups with various vaccine stances. Moreover, we observe that individuals with moderate stances, particularly the vaccine-hesitant majority, are more vulnerable to the influence of PBN compared to those with extreme views. Our findings provide critical insights to foster this line of research.

Zhen Tan, Dawei Li, Song Wang, Alimohammad Beigi, Bohan Jiang, Amrita Bhattacharjee, Mansooreh Karami, Jundong Li, Lu Cheng, Huan Liu. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.

Wildfires have become the primary mechanism for disturbing boreal forests. An adequate understanding of the restoration trend of forests after wildfires can help to develop forest recovery and protection measures. In this study, we employed Landsat 5 TM and MODIS data with vegetation area-series data to compare and assess the restoration of vegetation in the last 20 years after wildfires in the Daxinganling. The results show that (1) all vegetation types showed significant recovery trends in both burned zones (BZ) and unburned zones (UNBZ); (2) the rates of deciduous needleleaf forest (DNF) and mixed forest (MF) were slightly higher in UNBZ than in BZ, but the recovery rate of deciduous broadleaf forest (DBF) in BZ reached 156.5km2/a, which was higher than its rate in UNBZ; (3) DNF and MF had the same main period and existed in the same time scale with similar recovery periods, DBF has a 4-year period in BZ at a 6-year time scale, which represents DBF’s short recovery period and sharp fluctuations in this region. These results have important reference value for monitoring the restoration of boreal forest vegetation and ecological environmental protection after wildfires.

The advent of generative Large Language Models (LLMs) such as ChatGPT has catalyzed transformative advancements across multiple domains. However, alongside these advancements, they have also introduced potential threats. One critical concern is the misuse of LLMs by disinformation spreaders, leveraging these models to generate highly persuasive yet misleading content that challenges the disinformation detection system. This work aims to address this issue by answering three research questions: (1) To what extent can the current disinformation detection technique reliably detect LLM-generated disinformation? (2) If traditional techniques prove less effective, can LLMs themself be exploited to serve as a robust defense against advanced disinformation? and, (3) Should both these strategies falter, what novel approaches can be proposed to counter this burgeoning threat effectively? A holistic exploration for the formation and detection of disinformation is conducted to foster this line of research.

The COVID-19 epidemic is considered as the global health crisis of the whole society and the greatest challenge mankind faced since World War Two. Unfortunately, the fake news about COVID-19 is spreading as fast as the virus itself. The incorrect health measurements, anxiety, and hate speeches will have bad consequences on people's physical health, as well as their mental health in the whole world. To help better combat the COVID-19 fake news, we propose a new fake news detection dataset MM-COVID1(Multilingual and Multidimensional COVID-19 Fake News Data Repository). This dataset provides the multilingual fake news and the relevant social context. We collect 3981 pieces of fake news content and 7192 trustworthy information from English, Spanish, Portuguese, Hindi, French and Italian, 6 different languages. We present a detailed and exploratory analysis of MM-COVID from different perspectives.