Giuseppe Carenini

Wen Xiao, Giuseppe Carenini

Despite the success of recent abstractive summarizers on automatic evaluation metrics, the generated summaries still present factual inconsistencies with the source document. In this paper, we focus on entity-level factual inconsistency, i.e. reducing the mismatched entities between the generated summaries and the source documents. We therefore propose a novel entity-based SpanCopy mechanism, and explore its extension with a Global Relevance component. Experiment results on four summarization datasets show that SpanCopy can effectively improve the entity-level factual consistency with essentially no change in the word-level and entity-level saliency. The code is available at https://github.com/Wendy-Xiao/Entity-based-SpanCopy

Rindranirina Ramamonjison, Timothy T. Yu, Raymond Li et al. · tsinghua

The Natural Language for Optimization (NL4Opt) Competition was created to investigate methods of extracting the meaning and formulation of an optimization problem based on its text description. Specifically, the goal of the competition is to increase the accessibility and usability of optimization solvers by allowing non-experts to interface with them using natural language. We separate this challenging goal into two sub-tasks: (1) recognize and label the semantic entities that correspond to the components of the optimization problem; (2) generate a meaning representation (i.e., a logical form) of the problem from its detected problem entities. The first task aims to reduce ambiguity by detecting and tagging the entities of the optimization problems. The second task creates an intermediate representation of the linear programming (LP) problem that is converted into a format that can be used by commercial solvers. In this report, we present the LP word problem dataset and shared tasks for the NeurIPS 2022 competition. Furthermore, we investigate and compare the performance of the ChatGPT large language model against the winning solutions. Through this competition, we hope to bring interest towards the development of novel machine learning applications and datasets for optimization modeling.

Maria Laricheva, Chiyu Zhang, Yan Liu et al.

Conversational data is essential in psychology because it can help researchers understand individuals cognitive processes, emotions, and behaviors. Utterance labelling is a common strategy for analyzing this type of data. The development of NLP algorithms allows researchers to automate this task. However, psychological conversational data present some challenges to NLP researchers, including multilabel classification, a large number of classes, and limited available data. This study explored how automated labels generated by NLP methods are comparable to human labels in the context of conversations on adulthood transition. We proposed strategies to handle three common challenges raised in psychological studies. Our findings showed that the deep learning method with domain adaptation (RoBERTa-CON) outperformed all other machine learning methods; and the hierarchical labelling system that we proposed was shown to help researchers strategically analyze conversational data. Our Python code and NLP model are available at https://github.com/mlaricheva/automated_labeling.

Yan Liu, Maria Laricheva, Chiyu Zhang et al.

Transition to Adulthood is an essential life stage for many families. The prior research has shown that young people with intellectual or development disabil-ities (IDD) have more challenges than their peers. This study is to explore how to use natural language processing (NLP) methods, especially unsupervised machine learning, to assist psychologists to analyze emotions and sentiments and to use topic modeling to identify common issues and challenges that young people with IDD and their families have. Additionally, the results were compared to those obtained from young people without IDD who were in tran-sition to adulthood. The findings showed that NLP methods can be very useful for psychologists to analyze emotions, conduct cross-case analysis, and sum-marize key topics from conversational data. Our Python code is available at https://github.com/mlaricheva/emotion_topic_modeling.

Wen Xiao, Lesly Miculicich, Yang Liu et al. · microsoft-research

Content-Controllable Summarization generates summaries focused on the given controlling signals. Due to the lack of large-scale training corpora for the task, we propose a plug-and-play module RelAttn to adapt any general summarizers to the content-controllable summarization task. RelAttn first identifies the relevant content in the source documents, and then makes the model attend to the right context by directly steering the attention weight. We further apply an unsupervised online adaptive parameter searching algorithm to determine the degree of control in the zero-shot setting, while such parameters are learned in the few-shot setting. By applying the module to three backbone summarization models, experiments show that our method effectively improves all the summarizers, and outperforms the prefix-based method and a widely used plug-and-play model in both zero- and few-shot settings. Tellingly, more benefit is observed in the scenarios when more control is needed.

Amirhossein Dabiriaghdam, Shayan Vassef, Mohammadreza Bakhtiari et al.

Multimodal large language models are increasingly capable of complex reasoning, yet their performance often degrades when they must externalize a problem through a tool and then reason over the tool's output, specifically when they rely on visual aids. This gap is especially important because real engineering and scientific workflows often rely on visualization tools for analysis, validation, and decision-making. To study this discrepancy, we introduce VAMPS (Visual-Assisted Mathematical Problem Solving), a benchmark for graph-assisted mathematics. VAMPS contains 1,168 multimodal, bilingual multiple-choice question-answer pairs drawn from Iranian University Entrance Exam algebra and calculus problems and expanded with human-reviewed LLM-generated synthetic variants, all selected so that plotting provides a natural solution strategy by revealing intersections, extrema, asymptotes, etc. Designed for both benchmarking and diagnosis, VAMPS goes beyond prior multimodal benchmarks that primarily evaluate reasoning over fixed visual inputs by testing whether a model can benefit from constructing a useful graph and grounding its answer in the resulting visualization. Overall, we found that across a diverse set of models, direct analytical solving surprisingly outperforms tool-enabled visual solving, even on problems where plotting is a natural strategy.

Raymond Li, Ruixin Yang, Wen Xiao et al. · gatech

While transformer-based models have achieved state-of-the-art results in a variety of classification and generation tasks, their black-box nature makes them challenging for interpretability. In this work, we present a novel visual analytical framework to support the analysis of transformer-based generative networks. In contrast to previous work, which has mainly focused on encoder-based models, our framework is one of the first dedicated to supporting the analysis of transformer-based encoder-decoder models and decoder-only models for generative and classification tasks. Hence, we offer an intuitive overview that allows the user to explore different facets of the model through interactive visualization. To demonstrate the feasibility and usefulness of our framework, we present three detailed case studies based on real-world NLP research problems.

Chuyuan Li, Patrick Huber, Wen Xiao et al.

Discourse processing suffers from data sparsity, especially for dialogues. As a result, we explore approaches to build discourse structures for dialogues, based on attention matrices from Pre-trained Language Models (PLMs). We investigate multiple tasks for fine-tuning and show that the dialogue-tailored Sentence Ordering task performs best. To locate and exploit discourse information in PLMs, we propose an unsupervised and a semi-supervised method. Our proposals achieve encouraging results on the STAC corpus, with F1 scores of 57.2 and 59.3 for unsupervised and semi-supervised methods, respectively. When restricted to projective trees, our scores improved to 63.3 and 68.1.

Patrick Huber, Giuseppe Carenini

With a growing number of BERTology work analyzing different components of pre-trained language models, we extend this line of research through an in-depth analysis of discourse information in pre-trained and fine-tuned language models. We move beyond prior work along three dimensions: First, we describe a novel approach to infer discourse structures from arbitrarily long documents. Second, we propose a new type of analysis to explore where and how accurately intrinsic discourse is captured in the BERT and BART models. Finally, we assess how similar the generated structures are to a variety of baselines as well as their distribution within and between models.

Linzi Xing, Patrick Huber, Giuseppe Carenini

Recent neural supervised topic segmentation models achieve distinguished superior effectiveness over unsupervised methods, with the availability of large-scale training corpora sampled from Wikipedia. These models may, however, suffer from limited robustness and transferability caused by exploiting simple linguistic cues for prediction, but overlooking more important inter-sentential topical consistency. To address this issue, we present a discourse-aware neural topic segmentation model with the injection of above-sentence discourse dependency structures to encourage the model make topic boundary prediction based more on the topical consistency between sentences. Our empirical study on English evaluation datasets shows that injecting above-sentence discourse structures to a neural topic segmenter with our proposed strategy can substantially improve its performances on intra-domain and out-of-domain data, with little increase of model's complexity.

Raymond Li, Gabriel Murray, Giuseppe Carenini

In this work, we propose a method that combines two popular research areas by injecting linguistic structures into pre-trained language models in the parameter-efficient fine-tuning (PEFT) setting. In our approach, parallel adapter modules encoding different linguistic structures are combined using a novel Mixture-of-Linguistic-Experts architecture, where Gumbel-Softmax gates are used to determine the importance of these modules at each layer of the model. To reduce the number of parameters, we first train the model for a fixed small number of steps before pruning the experts based on their importance scores. Our experiment results with three different pre-trained models show that our approach can outperform state-of-the-art PEFT methods with a comparable number of parameters. In addition, we provide additional analysis to examine the experts selected by each model at each layer to provide insights for future studies.

Linzi Xing, Quan Tran, Fabian Caba et al.

Video topic segmentation unveils the coarse-grained semantic structure underlying videos and is essential for other video understanding tasks. Given the recent surge in multi-modal, relying solely on a single modality is arguably insufficient. On the other hand, prior solutions for similar tasks like video scene/shot segmentation cater to short videos with clear visual shifts but falter for long videos with subtle changes, such as livestreams. In this paper, we introduce a multi-modal video topic segmenter that utilizes both video transcripts and frames, bolstered by a cross-modal attention mechanism. Furthermore, we propose a dual-contrastive learning framework adhering to the unsupervised domain adaptation paradigm, enhancing our model's adaptability to longer, more semantically complex videos. Experiments on short and long video corpora demonstrate that our proposed solution, significantly surpasses baseline methods in terms of both accuracy and transferability, in both intra- and cross-domain settings.

Linzi Xing, Brad Hackinen, Giuseppe Carenini

U.S. Federal Regulators receive over one million comment letters each year from businesses, interest groups, and members of the public, all advocating for changes to proposed regulations. These comments are believed to have wide-ranging impacts on public policy. However, measuring the impact of specific comments is challenging because regulators are required to respond to comments but they do not have to specify which comments they are addressing. In this paper, we propose a simple yet effective solution to this problem by using an iterative contrastive method to train a neural model aiming for matching text from public comments to responses written by regulators. We demonstrate that our proposal substantially outperforms a set of selected text-matching baselines on a human-annotated test set. Furthermore, it delivers performance comparable to the most advanced gigantic language model (i.e., GPT-4), and is more cost-effective when handling comments and regulator responses matching in larger scale.

Patrick Huber, Giuseppe Carenini

Discourse analysis and discourse parsing have shown great impact on many important problems in the field of Natural Language Processing (NLP). Given the direct impact of discourse annotations on model performance and interpretability, robustly extracting discourse structures from arbitrary documents is a key task to further improve computational models in NLP. To this end, we present a new, supervised paradigm directly tackling the domain adaptation issue in discourse parsing. Specifically, we introduce the first fully supervised discourse parser designed to alleviate the domain dependency through a staged model of weak classifiers by introducing the gradient boosting framework.

Patrick Huber, Giuseppe Carenini

Discourse parsing is an essential upstream task in Natural Language Processing with strong implications for many real-world applications. Despite its widely recognized role, most recent discourse parsers (and consequently downstream tasks) still rely on small-scale human-annotated discourse treebanks, trying to infer general-purpose discourse structures from very limited data in a few narrow domains. To overcome this dire situation and allow discourse parsers to be trained on larger, more diverse and domain-independent datasets, we propose a framework to generate "silver-standard" discourse trees from distant supervision on the auxiliary task of sentiment analysis.

Patrick Huber, Giuseppe Carenini

With a growing need for robust and general discourse structures in many downstream tasks and real-world applications, the current lack of high-quality, high-quantity discourse trees poses a severe shortcoming. In order the alleviate this limitation, we propose a new strategy to generate tree structures in a task-agnostic, unsupervised fashion by extending a latent tree induction framework with an auto-encoding objective. The proposed approach can be applied to any tree-structured objective, such as syntactic parsing, discourse parsing and others. However, due to the especially difficult annotation process to generate discourse trees, we initially develop such method to complement task-specific models in generating much larger and more diverse discourse treebanks.

Yasamin Medghalchi, Milad Yazdani, Amirhossein Dabiriaghdam et al.

Ultrasound is widely used in clinical practice due to its portability, cost-effectiveness, safety, and real-time imaging capabilities. However, image acquisition and interpretation remain highly operator dependent, motivating the development of robust AI-assisted analysis methods. Vision-language models (VLMs) have recently demonstrated strong multimodal reasoning capabilities and competitive performance in medical image analysis, including ultrasound. However, emerging evidence highlights significant concerns about their trustworthiness. In particular, adversarial robustness is critical because Med-VLMs operate via natural-language instructions, rendering prompt formulation a realistic and practically exploitable point of vulnerability. Small variations (typos, shorthand, underspecified requests, or ambiguous wording) can meaningfully shift model outputs. We propose a scalable adversarial evaluation framework that leverages a large language model (LLM) to generate clinically plausible adversarial prompt variants via "humanized" rewrites and minimal edits that mimic routine clinical communication. Using ultrasound multiple-choice question answering benchmarks, we systematically assess the vulnerability of SOTA Med-VLMs to these attacks, examine how attacker LLM capacity influences attack success, analyze the relationship between attack success and model confidence, and identify consistent failure patterns across models. Our results highlight realistic robustness gaps that must be addressed for safe clinical translation. Code will be released publicly following the review process.

Mahdi Mostajabdaveh, Timothy T. Yu, Samarendra Chandan Bindu Dash et al.

In this paper, we introduce and apply Operations Research Question Answering (ORQA), a new benchmark designed to assess the generalization capabilities of Large Language Models (LLMs) in the specialized technical domain of Operations Research (OR). This benchmark evaluates whether LLMs can emulate the knowledge and reasoning skills of OR experts when confronted with diverse and complex optimization problems. The dataset, developed by OR experts, features real-world optimization problems that demand multistep reasoning to construct their mathematical models. Our evaluations of various open source LLMs, such as LLaMA 3.1, DeepSeek, and Mixtral, reveal their modest performance, highlighting a gap in their ability to generalize to specialized technical domains. This work contributes to the ongoing discourse on LLMs generalization capabilities, offering valuable insights for future research in this area. The dataset and evaluation code are publicly available.

Chuyuan Li, Raymond Li, Thalia S. Field et al.

Alzheimer's Disease (AD) is a progressive neurodegenerative disorder that leads to dementia, and early intervention can greatly benefit from analyzing linguistic abnormalities. In this work, we explore the potential of Large Language Models (LLMs) as health assistants for AD diagnosis from patient-generated text using in-context learning (ICL), where tasks are defined through a few input-output examples. Empirical results reveal that conventional ICL methods, such as similarity-based selection, perform poorly for AD diagnosis, likely due to the inherent complexity of this task. To address this, we introduce Delta-KNN, a novel demonstration selection strategy that enhances ICL performance. Our method leverages a delta score to assess the relative gains of each training example, coupled with a KNN-based retriever that dynamically selects optimal "representatives" for a given input. Experiments on two AD detection datasets across three open-source LLMs demonstrate that Delta-KNN consistently outperforms existing ICL baselines. Notably, when using the Llama-3.1 model, our approach achieves new state-of-the-art results, surpassing even supervised classifiers.

Amirhossein Abaskohi, Yuhang He, Peter West et al.

Computer-use agents~(CUAs) rely on visual observations of graphical user interfaces, where each screenshot is encoded into a large number of visual tokens. As interaction trajectories grow, the token cost increases rapidly, limiting the amount of history that can be incorporated under fixed context and compute budgets. This has resulted in no or very limited improvement in the performance when using history unlike other domains. We address this inefficiency by introducing ReVision, which is used to train multimodal language models on trajectories where redundant visual patches are removed using a learned patch selector that compares patch representations across consecutive screenshots while preserving spatial structure required by the model. Across three benchmarks, OSWorld, WebTailBench, and AgentNetBench, when processing trajectories with 5 history screenshots using Qwen2.5-VL-7B, ReVision reduces token usage by approximately 46% on average while improving success rate by 3% over the no drop baseline. This establishes a clear efficiency gain, enabling agents to process longer trajectories with fewer tokens. With this improved efficiency, we revisit the role of history in CUAs and find that performance continues to improve as more past observations are incorporated when redundancy is removed. This suggests that the commonly observed saturation in visual history is not due to limited usefulness of past information, but rather a consequence of inefficient token representations.

Chuyuan Li, Giuseppe Carenini

We introduce BeDiscovER (Benchmark of Discourse Understanding in the Era of Reasoning Language Models), an up-to-date, comprehensive suite for evaluating the discourse-level knowledge of modern LLMs. BeDiscovER compiles 5 publicly available discourse tasks across discourse lexicon, (multi-)sentential, and documental levels, with in total 52 individual datasets. It covers both extensively studied tasks such as discourse parsing and temporal relation extraction, as well as some novel challenges such as discourse particle disambiguation (e.g., ``just''), and also aggregates a shared task on Discourse Relation Parsing and Treebanking for multilingual and multi-framework discourse relation classification. We evaluate open-source LLMs: Qwen3 series, DeepSeek-R1, and frontier model such as GPT-5-mini on BeDiscovER, and find that state-of-the-art models exhibit strong performance in arithmetic aspect of temporal reasoning, but they struggle with full document reasoning and some subtle semantic and discourse phenomena, such as rhetorical relation recognition.

Wen Xiao, Yujia Xie, Giuseppe Carenini et al.

Tailoring outputs from large language models, like ChatGPT, to implicit user preferences remains a challenge despite their impressive generative capabilities. In this paper, we propose a tri-agent generation pipeline comprising a generator, an instructor, and an editor to enhance output personalization. The generator produces an initial output, the instructor automatically generates editing instructions based on user preferences, and the editor refines the output to align with those preferences. The inference-only large language model (ChatGPT) serves as both the generator and editor, with a smaller model acting as the instructor to guide output generation. We train the instructor using editor-steered reinforcement learning, leveraging feedback from a large-scale editor model to optimize instruction generation. Experimental results on two abstractive summarization datasets demonstrate the effectiveness of our approach in generating outputs that better meet user expectations. Code is available at \url{https://github.com/Wendy-Xiao/chatgpt_editing_summ}

Wen Xiao, Iz Beltagy, Giuseppe Carenini et al.

We introduce PRIMERA, a pre-trained model for multi-document representation with a focus on summarization that reduces the need for dataset-specific architectures and large amounts of fine-tuning labeled data. PRIMERA uses our newly proposed pre-training objective designed to teach the model to connect and aggregate information across documents. It also uses efficient encoder-decoder transformers to simplify the processing of concatenated input documents. With extensive experiments on 6 multi-document summarization datasets from 3 different domains on zero-shot, few-shot and full-supervised settings, PRIMERA outperforms current state-of-the-art dataset-specific and pre-trained models on most of these settings with large margins. The code and pre-trained models can be found at \url{https://github.com/allenai/PRIMER}.

Yuwei Yin, Chuyuan Li, Giuseppe Carenini

Accurately understanding the intent behind speech, conversation, and writing is crucial to the development of helpful Large Language Model (LLM) assistants. This paper introduces IntentGrasp, a comprehensive benchmark for evaluating the intent understanding capability of LLMs. Derived from 49 high-quality, open-licensed corpora spanning 12 diverse domains, IntentGrasp is constructed through source datasets curation, intent label contextualization, and task format unification. IntentGrasp contains a large-scale training set of 262,759 instances and two evaluation sets: an All Set of 12,909 test cases and a more balanced and challenging Gem Set of 470 cases. Extensive evaluations on 20 LLMs across 7 families (including frontier models such as GPT-5.4, Gemini-3.1-Pro, and Claude-Opus-4.7) demonstrate unsatisfactory performance, with scores below 60% on All Set and below 25% on Gem set. Notably, 17 out of 20 tested models perform worse than a random-guess baseline (15.2%) on Gem Set, while the estimated human performance is ~81.1%, showing substantial room for improvement. To enhance such ability, this paper proposes Intentional Fine-Tuning (IFT), which fine-tunes the models on the training set in IntentGrasp, yielding significant gains of 30+ F1 points on All Set and 20+ points on Gem Set. Tellingly, the leave-one-domain-out (Lodo) experiments further demonstrate the strong cross-domain generalizability of IFT, verifying that it is a promising approach to substantially enhancing the intent understanding of LLMs. Overall, by benchmarking and boosting intent understanding ability, this study sheds light on a promising path towards more intentional, capable, and safe AI assistants for human benefits and social good.

Juntai Cao, Xiang Zhang, Raymond Li et al.

Recent advances in test-time scaling have shown promising results in improving Large Language Model (LLM) performance through strategic computation allocation during inference. While this approach has demonstrated strong improvements in logical and mathematical reasoning tasks, its application to natural language generation (NLG), particularly summarization, remains unexplored. Multi-Document Summarization (MDS), a fundamental task in NLG, presents unique challenges by requiring models to extract and synthesize essential information across multiple lengthy documents. Unlike reasoning tasks, MDS demands a more nuanced approach to prompt design and ensemble methods, as no single "best" prompt can satisfy diverse summarization requirements. We propose a novel framework leveraging test-time scaling for MDS. Our approach employs prompt ensemble techniques to generate multiple candidate summaries using various prompts, then combines them with an aggregator to produce a refined summary. To evaluate our method effectively, we also introduce two new LLM-based metrics: the Consistency-Aware Preference (CAP) score and LLM Atom-Content-Unit (LLM-ACU) score, which assess summary quality while addressing the positional bias inherent in traditional automatic evaluation. Our extensive experiments demonstrate that this framework significantly enhances summary quality while also revealing the practical scaling boundaries to MDS tasks.

Subhrangshu Nandi, Arghya Datta, Nikhil Vichare et al.

Large Language Models (LLMs) demonstrate impressive general-purpose reasoning and problem-solving abilities. However, they struggle with executing complex, long-horizon workflows that demand strict adherence to Standard Operating Procedures (SOPs), a critical requirement for real-world industrial automation. Despite this need, there is a lack of public benchmarks that reflect the complexity, structure, and domain-specific nuances of SOPs. To address this, we present three main contributions. First, we introduce a synthetic data generation framework to create realistic, industry-grade SOPs that rigorously test the planning, reasoning, and tool-use capabilities of LLM-based agents. Second, using this framework, we develop SOP-Bench, a benchmark of over 1,800 tasks across 10 industrial domains, each with APIs, tool interfaces, and human-validated test cases. Third, we evaluate two prominent agent architectures: Function-Calling and ReAct Agents, on SOP-Bench, observing average success rates of only 27% and 48%, respectively. Remarkably, when the tool registry is much larger than necessary, agents invoke incorrect tools nearly 100% of the time. These findings underscore a substantial gap between current agentic capabilities of LLMs and the demands of automating real-world SOPs. Performance varies significantly by task and domain, highlighting the need for domain-specific benchmarking and architectural choices before deployment. SOP-Bench is publicly available at http://sop-bench.s3-website-us-west-2.amazonaws.com/. We also release the prompts underpinning the data generation framework to support new domain-specific SOP benchmarks. We invite the community to extend SOP-Bench with SOPs from their industrial domains.

Felipe González-Pizarro, Giuseppe Carenini

Neural topic models can successfully find coherent and diverse topics in textual data. However, they are limited in dealing with multimodal datasets (e.g., images and text). This paper presents the first systematic and comprehensive evaluation of multimodal topic modeling of documents containing both text and images. In the process, we propose two novel topic modeling solutions and two novel evaluation metrics. Overall, our evaluation on an unprecedented rich and diverse collection of datasets indicates that both of our models generate coherent and diverse topics. Nevertheless, the extent to which one method outperforms the other depends on the metrics and dataset combinations, which suggests further exploration of hybrid solutions in the future. Notably, our succinct human evaluation aligns with the outcomes determined by our proposed metrics. This alignment not only reinforces the credibility of our metrics but also highlights the potential for their application in guiding future multimodal topic modeling endeavors.

Yuwei Yin, Giuseppe Carenini

Large language models (LLMs) have demonstrated impressive capabilities on complex evaluation benchmarks, many of which are formulated as question-answering (QA) tasks. Enhancing the performance of LLMs in QA contexts is becoming increasingly vital for advancing their development and applicability. This paper introduces ARR, an intuitive, effective, and general QA solving method that explicitly incorporates three key steps: analyzing the intent of the question, retrieving relevant information, and reasoning step by step. Notably, this paper is the first to introduce intent analysis in QA, which plays a vital role in ARR. Comprehensive evaluations across 10 diverse QA tasks demonstrate that ARR consistently outperforms the baseline methods. Ablation and case studies further validate the positive contributions of each ARR component. Furthermore, experiments involving variations in prompt design indicate that ARR maintains its effectiveness regardless of the specific prompt formulation. Additionally, extensive evaluations across various model sizes, LLM series, and generation settings solidify the effectiveness, robustness, and generalizability of ARR.

Amirhossein Abaskohi, Spandana Gella, Giuseppe Carenini et al.

Multimodal multihop question answering (MMQA) requires reasoning over images and text from multiple sources. Despite advances in visual question answering, this multihop setting remains underexplored due to a lack of quality datasets. Existing methods focus on single-hop, single-modality, or short texts, limiting real-world applications like interpreting educational documents with long, multimodal content. To fill this gap, we introduce FM2DS, the first framework for creating a high-quality dataset for MMQA. Our approach consists of a 5-stage pipeline that involves acquiring relevant multimodal documents from Wikipedia, synthetically generating high-level questions and answers, and validating them through rigorous criteria to ensure data quality. We evaluate our methodology by training models on our synthesized dataset and testing on two benchmarks: MultimodalQA and WebQA. Our results demonstrate that, with an equal sample size, models trained on our synthesized data outperform those trained on human-collected data by 1.9 in exact match (EM) score on average. Additionally, we introduce M2QA-Bench with 1k samples, the first benchmark for MMQA on long documents, generated using FM2DS and refined by human annotators. We believe our data synthesis method will serve as a strong foundation for training and evaluating MMQA models.

Yuwei Yin, EunJeong Hwang, Giuseppe Carenini

Intent, typically clearly formulated and planned, functions as a cognitive framework for communication and problem-solving. This paper introduces the concept of Speaking with Intent (SWI) in large language models (LLMs), where the explicitly generated intent encapsulates the model's underlying intention and provides high-level planning to guide subsequent analysis and action. By emulating deliberate and purposeful thoughts in the human mind, SWI is hypothesized to enhance the reasoning capabilities and generation quality of LLMs. Extensive experiments on text summarization, multi-task question answering, and mathematical reasoning benchmarks consistently demonstrate the effectiveness and generalizability of Speaking with Intent over direct generation without explicit intent. Further analysis corroborates the generalizability of SWI under different experimental settings. Moreover, human evaluations verify the coherence, effectiveness, and interpretability of the intent produced by SWI. The promising results in enhancing LLMs with explicit intents pave a new avenue for boosting LLMs' generation and reasoning abilities with cognitive notions.

Amirhossein Abaskohi, Amirhossein Dabiriaghdam, Lele Wang et al.

Memes, combining text and images, frequently use metaphors to convey persuasive messages, shaping public opinion. Motivated by this, our team engaged in SemEval-2024 Task 4, a hierarchical multi-label classification task designed to identify rhetorical and psychological persuasion techniques embedded within memes. To tackle this problem, we introduced a caption generation step to assess the modality gap and the impact of additional semantic information from images, which improved our result. Our best model utilizes GPT-4 generated captions alongside meme text to fine-tune RoBERTa as the text encoder and CLIP as the image encoder. It outperforms the baseline by a large margin in all 12 subtasks. In particular, it ranked in top-3 across all languages in Subtask 2a, and top-4 in Subtask 2b, demonstrating quantitatively strong performance. The improvement achieved by the introduced intermediate step is likely attributable to the metaphorical essence of images that challenges visual encoders. This highlights the potential for improving abstract visual semantics encoding.

Raymond Li, Amirhossein Abaskohi, Chuyuan Li et al.

Traditional neural topic models are typically optimized by reconstructing the document's Bag-of-Words (BoW) representations, overlooking contextual information and struggling with data sparsity. In this work, we propose a novel approach to construct semantically-grounded soft label targets using Language Models (LMs) by projecting the next token probabilities, conditioned on a specialized prompt, onto a pre-defined vocabulary to obtain contextually enriched supervision signals. By training the topic models to reconstruct the soft labels using the LM hidden states, our method produces higher-quality topics that are more closely aligned with the underlying thematic structure of the corpus. Experiments on three datasets show that our method achieves substantial improvements in topic coherence, purity over existing baselines. Additionally, we also introduce a retrieval-based metric, which shows that our approach significantly outperforms existing methods in identifying semantically similar documents, highlighting its effectiveness for retrieval-oriented applications.

EunJeong Hwang, Yuwei Yin, Giuseppe Carenini et al.

Theory of Mind (ToM)-an understanding of the mental states of others-is a key aspect of human social intelligence, yet, chatbots and LLM-based social agents do not typically integrate it. In this work, we demonstrate that LLMs that explicitly use ToM get better at dialogue, achieving goals more effectively. After showing that simply prompting models to generate mental states between dialogue turns already provides significant benefit, we further introduce ToMAgent (ToMA), a ToM-focused dialogue agent. ToMA is trained by pairing ToM with dialogue lookahead to produce mental states that are maximally useful for achieving dialogue goals. Experiments on the Sotopia interactive social evaluation benchmark demonstrate the effectiveness of our method over a range of baselines. Comprehensive analysis shows that ToMA exhibits more strategic, goal-oriented reasoning behaviors, which enable long-horizon adaptation, while maintaining better relationships with their partners. Our results suggest a step forward in integrating ToM for building socially intelligent LLM agents.

Ali Salamatian, Amirhossein Abaskohi, Wan-Cyuan Fan et al.

Charts are a crucial visual medium for communicating and representing information. While Large Vision-Language Models (LVLMs) have made progress on chart question answering (CQA), the task remains challenging, particularly when models attend to irrelevant regions of the chart. In this work, we present ChartGaze, a new eye-tracking dataset that captures human gaze patterns during chart reasoning tasks. Through a systematic comparison of human and model attention, we find that LVLMs often diverge from human gaze, leading to reduced interpretability and accuracy. To address this, we propose a gaze-guided attention refinement that aligns image-text attention with human fixations. Our approach improves both answer accuracy and attention alignment, yielding gains of up to 2.56 percentage points across multiple models. These results demonstrate the promise of incorporating human gaze to enhance both the reasoning quality and interpretability of chart-focused LVLMs.

Amirhossein Abaskohi, Raymond Li, Chuyuan Li et al.

We introduce CEMTM, a context-enhanced multimodal topic model designed to infer coherent and interpretable topic structures from both short and long documents containing text and images. CEMTM builds on fine-tuned large vision language models (LVLMs) to obtain contextualized embeddings, and employs a distributional attention mechanism to weight token-level contributions to topic inference. A reconstruction objective aligns topic-based representations with the document embedding, encouraging semantic consistency across modalities. Unlike existing approaches, CEMTM can process multiple images per document without repeated encoding and maintains interpretability through explicit word-topic and document-topic distributions. Extensive experiments on six multimodal benchmarks show that CEMTM consistently outperforms unimodal and multimodal baselines, achieving a remarkable average LLM score of 2.61. Further analysis shows its effectiveness in downstream few-shot retrieval and its ability to capture visually grounded semantics in complex domains such as scientific articles.

Chuyuan Li, Austin Xu, Shafiq Joty et al.

A key challenge in Multi-Document Summarization (MDS) is effectively integrating information from multiple sources while maintaining coherence and topical relevance. While Large Language Models have shown impressive results in single-document summarization, their performance on MDS still leaves room for improvement. In this paper, we propose a topic-guided reinforcement learning approach to improve content selection in MDS. We first show that explicitly prompting models with topic labels enhances the informativeness of the generated summaries. Building on this insight, we propose a novel topic reward within the Group Relative Policy Optimization (GRPO) framework to measure topic alignment between the generated summary and source documents. Experimental results on the Multi-News and Multi-XScience datasets demonstrate that our method consistently outperforms strong baselines, highlighting the effectiveness of leveraging topical cues in MDS.

Timothy Tin Long Yu, Mahdi Mostajabdaveh, Jabo Serge Byusa et al.

Large language models (LLMs) present intriguing opportunities to enhance user interaction with traditional algorithms and tools in real-world applications. An advanced planning system (APS) is a sophisticated software that leverages optimization to help operations planners create, interpret, and modify an operational plan. While highly beneficial, many customers are priced out of using an APS due to the ongoing costs of consultants responsible for customization and maintenance. To address the need for a more accessible APS expressed by supply chain planners, we present SmartAPS, a conversational system built on a tool-augmented LLM. Our system provides operations planners with an intuitive natural language chat interface, allowing them to query information, perform counterfactual reasoning, receive recommendations, and execute scenario analysis to better manage their operation. A short video demonstrating the system has been released: https://youtu.be/KtIrJjlDbyw

Amirhossein Abaskohi, Amrutha Varshini Ramesh, Shailesh Nanisetty et al.

We introduce AgentAda, the first LLM-powered analytics agent that can learn and use new analytics skills to extract more specialized insights. Unlike existing methods that require users to manually decide which data analytics method to apply, AgentAda automatically identifies the skill needed from a library of analytical skills to perform the analysis. This also allows AgentAda to use skills that existing LLMs cannot perform out of the box. The library covers a range of methods, including clustering, predictive modeling, and NLP techniques like BERT, which allow AgentAda to handle complex analytics tasks based on what the user needs. AgentAda's dataset-to-insight extraction strategy consists of three key steps: (I) a question generator to generate queries relevant to the user's goal and persona, (II) a hybrid Retrieval-Augmented Generation (RAG)-based skill matcher to choose the best data analytics skill from the skill library, and (III) a code generator that produces executable code based on the retrieved skill's documentation to extract key patterns. We also introduce KaggleBench, a benchmark of curated notebooks across diverse domains, to evaluate AgentAda's performance. We conducted a human evaluation demonstrating that AgentAda provides more insightful analytics than existing tools, with 48.78% of evaluators preferring its analyses, compared to 27.67% for the unskilled agent. We also propose a novel LLM-as-a-judge approach that we show is aligned with human evaluation as a way to automate insight quality evaluation at larger scale.

Giuseppe Carenini, Jordon Johnson, Ali Salamatian

Automatically captioning visualizations is not new, but recent advances in large language models(LLMs) open exciting new possibilities. In this tutorial, after providing a brief review of Information Visualization (InfoVis) principles and past work in captioning, we introduce neural models and the transformer architecture used in generic LLMs. We then discuss their recent applications in InfoVis, with a focus on captioning. Additionally, we explore promising future directions in this field.

Yuxi Feng, Raymond Li, Zhenan Fan et al.

While in-context Learning (ICL) has proven to be an effective technique to improve the performance of Large Language Models (LLMs) in a variety of complex tasks, notably in translating natural language questions into Structured Query Language (NL2SQL), the question of how to select the most beneficial demonstration examples remains an open research problem. While prior works often adapted off-the-shelf encoders to retrieve examples dynamically, an inherent discrepancy exists in the representational capacities between the external retrievers and the LLMs. Further, optimizing the selection of examples is a non-trivial task, since there are no straightforward methods to assess the relative benefits of examples without performing pairwise inference. To address these shortcomings, we propose DeTriever, a novel demonstration retrieval framework that learns a weighted combination of LLM hidden states, where rich semantic information is encoded. To train the model, we propose a proxy score that estimates the relative benefits of examples based on the similarities between output queries. Experiments on two popular NL2SQL benchmarks demonstrate that our method significantly outperforms the state-of-the-art baselines on one-shot NL2SQL tasks.

Raymond Li, Felipe González-Pizarro, Linzi Xing et al.

The standard approach for neural topic modeling uses a variational autoencoder (VAE) framework that jointly minimizes the KL divergence between the estimated posterior and prior, in addition to the reconstruction loss. Since neural topic models are trained by recreating individual input documents, they do not explicitly capture the coherence between topic words on the corpus level. In this work, we propose a novel diversity-aware coherence loss that encourages the model to learn corpus-level coherence scores while maintaining a high diversity between topics. Experimental results on multiple datasets show that our method significantly improves the performance of neural topic models without requiring any pretraining or additional parameters.

Patrick Huber, Linzi Xing, Giuseppe Carenini

RST-style discourse parsing plays a vital role in many NLP tasks, revealing the underlying semantic/pragmatic structure of potentially complex and diverse documents. Despite its importance, one of the most prevailing limitations in modern day discourse parsing is the lack of large-scale datasets. To overcome the data sparsity issue, distantly supervised approaches from tasks like sentiment analysis and summarization have been recently proposed. Here, we extend this line of research by exploiting distant supervision from topic segmentation, which can arguably provide a strong and oftentimes complementary signal for high-level discourse structures. Experiments on two human-annotated discourse treebanks confirm that our proposal generates accurate tree structures on sentence and paragraph level, consistently outperforming previous distantly supervised models on the sentence-to-document task and occasionally reaching even higher scores on the sentence-to-paragraph level.

Raymond Li, Wen Xiao, Linzi Xing et al.

The multi-head self-attention mechanism of the transformer model has been thoroughly investigated recently. In one vein of study, researchers are interested in understanding why and how transformers work. In another vein, researchers propose new attention augmentation methods to make transformers more accurate, efficient and interpretable. In this paper, we combine these two lines of research in a human-in-the-loop pipeline to first discover important task-specific attention patterns. Then those patterns are injected, not only to smaller models, but also to the original model. The benefits of our pipeline and discovered patterns are demonstrated in two case studies with extractive summarization and topic segmentation. After discovering interpretable patterns in BERT-based models fine-tuned for the two downstream tasks, experiments indicate that when we inject the patterns into attention heads, the models show considerable improvements in accuracy and efficiency.

Raymond Li, Wen Xiao, Lanjun Wang et al.

Transformers are the dominant architecture in NLP, but their training and fine-tuning is still very challenging. In this paper, we present the design and implementation of a visual analytic framework for assisting researchers in such process, by providing them with valuable insights about the model's intrinsic properties and behaviours. Our framework offers an intuitive overview that allows the user to explore different facets of the model (e.g., hidden states, attention) through interactive visualization, and allows a suite of built-in algorithms that compute the importance of model components and different parts of the input sequence. Case studies and feedback from a user focus group indicate that the framework is useful, and suggest several improvements.

Raymond Li, Enamul Hoque, Giuseppe Carenini et al.

The proliferation of text messaging for mobile health is generating a large amount of patient-doctor conversations that can be extremely valuable to health care professionals. We present ConVIScope, a visual text analytic system that tightly integrates interactive visualization with natural language processing in analyzing patient-doctor conversations. ConVIScope was developed in collaboration with healthcare professionals following a user-centered iterative design. Case studies with six domain experts suggest the potential utility of ConVIScope and reveal lessons for further developments.

Linzi Xing, Giuseppe Carenini

Dialogue topic segmentation is critical in several dialogue modeling problems. However, popular unsupervised approaches only exploit surface features in assessing topical coherence among utterances. In this work, we address this limitation by leveraging supervisory signals from the utterance-pair coherence scoring task. First, we present a simple yet effective strategy to generate a training corpus for utterance-pair coherence scoring. Then, we train a BERT-based neural utterance-pair coherence model with the obtained training corpus. Finally, such model is used to measure the topical relevance between utterances, acting as the basis of the segmentation inference. Experiments on three public datasets in English and Chinese demonstrate that our proposal outperforms the state-of-the-art baselines.

Patrick Huber, Wen Xiao, Giuseppe Carenini

Aiming for a better integration of data-driven and linguistically-inspired approaches, we explore whether RST Nuclearity, assigning a binary assessment of importance between text segments, can be replaced by automatically generated, real-valued scores, in what we call a Weighted-RST framework. In particular, we find that weighted discourse trees from auxiliary tasks can benefit key NLP downstream applications, compared to nuclearity-centered approaches. We further show that real-valued importance distributions partially and interestingly align with the assessment and uncertainty of human annotators.

Linzi Xing, Wen Xiao, Giuseppe Carenini

In news articles the lead bias is a common phenomenon that usually dominates the learning signals for neural extractive summarizers, severely limiting their performance on data with different or even no bias. In this paper, we introduce a novel technique to demote lead bias and make the summarizer focus more on the content semantics. Experiments on two news corpora with different degrees of lead bias show that our method can effectively demote the model's learned lead bias and improve its generality on out-of-distribution data, with little to no performance loss on in-distribution data.

Wen Xiao, Patrick Huber, Giuseppe Carenini

Previous work indicates that discourse information benefits summarization. In this paper, we explore whether this synergy between discourse and summarization is bidirectional, by inferring document-level discourse trees from pre-trained neural summarizers. In particular, we generate unlabeled RST-style discourse trees from the self-attention matrices of the transformer model. Experiments across models and datasets reveal that the summarizer learns both, dependency- and constituency-style discourse information, which is typically encoded in a single head, covering long- and short-distance discourse dependencies. Overall, the experimental results suggest that the learned discourse information is general and transferable inter-domain.

Patrick Huber, Giuseppe Carenini

Discourse information, as postulated by popular discourse theories, such as RST and PDTB, has been shown to improve an increasing number of downstream NLP tasks, showing positive effects and synergies of discourse with important real-world applications. While methods for incorporating discourse become more and more sophisticated, the growing need for robust and general discourse structures has not been sufficiently met by current discourse parsers, usually trained on small scale datasets in a strictly limited number of domains. This makes the prediction for arbitrary tasks noisy and unreliable. The overall resulting lack of high-quality, high-quantity discourse trees poses a severe limitation to further progress. In order the alleviate this shortcoming, we propose a new strategy to generate tree structures in a task-agnostic, unsupervised fashion by extending a latent tree induction framework with an auto-encoding objective. The proposed approach can be applied to any tree-structured objective, such as syntactic parsing, discourse parsing and others. However, due to the especially difficult annotation process to generate discourse trees, we initially develop a method to generate larger and more diverse discourse treebanks. In this paper we are inferring general tree structures of natural text in multiple domains, showing promising results on a diverse set of tasks.