Papers in nlp

Large Vision-Language Models (VLMs) have achieved remarkable progress in multimodal understanding, yet they struggle when reasoning over information-intensive images that densely interleave textual annotations with fine-grained graphical elements. The main challenges lie in precisely localizing critical cues in dense layouts and multi-hop reasoning to integrate dispersed evidence. We propose Speculative Verdict (SV), a training-free framework inspired by speculative decoding that combines multiple lightweight draft experts with a large verdict model. In the draft stage, small VLMs act as draft experts to generate reasoning paths that provide diverse localization candidates; in the verdict stage, a strong VLM synthesizes these paths to produce the final answer, minimizing computational cost while recovering correct answers. To further improve efficiency and accuracy, SV introduces a consensus expert selection mechanism that forwards only high-agreement reasoning paths to the verdict. Empirically, SV achieves consistent gains on challenging information-intensive and high-resolution visual question answering benchmarks, including InfographicVQA, ChartMuseum, ChartQAPro, and HR-Bench 4K. By synthesizing correct insights from multiple partially accurate reasoning paths, SV achieves both error correction and cost-efficiency compared to large proprietary models or training pipelines. Code is available at https://github.com/Tinaliu0123/speculative-verdict

With the widespread use of large language models (LLMs), many researchers have turned their attention to detecting text generated by them. However, there is no consistent or precise definition of their target, namely "LLM-generated text". Differences in usage scenarios and the diversity of LLMs further increase the difficulty of detection. What is commonly regarded as the detecting target usually represents only a subset of the text that LLMs can potentially produce. Human edits to LLM outputs, together with the subtle influences that LLMs exert on their users, are blurring the line between LLM-generated and human-written text. Existing benchmarks and evaluation approaches do not adequately address the various conditions in real-world detector applications. Hence, the numerical results of detectors are often misunderstood, and their significance is diminishing. Therefore, detectors remain useful under specific conditions, but their results should be interpreted only as references rather than decisive indicators.

With the rapid growth of research in AI and robotics now producing over 10,000 papers annually it has become increasingly difficult for researchers to stay up to date. Fast evolving trends, the rise of interdisciplinary work, and the need to explore domains beyond one's expertise all contribute to this challenge. To address these issues, we propose a generalizable pipeline capable of systematically analyzing any research area: identifying emerging trends, uncovering cross domain opportunities, and offering concrete starting points for new inquiry. In this work, we present Real Deep Research (RDR) a comprehensive framework applied to the domains of AI and robotics, with a particular focus on foundation models and robotics advancements. We also briefly extend our analysis to other areas of science. The main paper details the construction of the RDR pipeline, while the appendix provides extensive results across each analyzed topic. We hope this work sheds light for researchers working in the field of AI and beyond.

Recently, Sharma et al. suggested a method called Layer-SElective-Rank reduction (LASER) which demonstrated that pruning high-order components of carefully chosen LLM's weight matrices can boost downstream accuracy -- without any gradient-based fine-tuning. Yet LASER's exhaustive, per-matrix search (each requiring full-dataset forward passes) makes it impractical for rapid deployment. We demonstrate that this overhead can be removed and find that: (i) Only a small, carefully chosen subset of matrices needs to be inspected -- eliminating the layer-by-layer sweep, (ii) The gradient of each matrix's singular values pinpoints which matrices merit reduction, (iii) Increasing the factorization search space by allowing matrices rows to cluster around multiple subspaces and then decomposing each cluster separately further reduces overfitting on the original training data and further lifts accuracy by up to 24.6 percentage points, and finally, (iv) we discover that evaluating on just 100 samples rather than the full training data -- both for computing the indicative gradients and for measuring the final accuracy -- suffices to further reduce the search time; we explain that as adaptation to downstream tasks is dominated by prompting style, not dataset size. As a result, we show that combining these findings yields a fast and robust adaptation algorithm for downstream tasks. Overall, with a single gradient step on 100 examples and a quick scan of the top candidate layers and factorization techniques, we can adapt LLMs to new datasets -- entirely without fine-tuning.

A common strategy to reduce the computational costs of using long contexts in retrieval-augmented generation (RAG) with large language models (LLMs) is soft context compression, where the input sequence is transformed into a shorter continuous representation. We develop a lightweight and simple mean-pooling approach that consistently outperforms the widely used compression-tokens architecture, and study training the same compressor to output multiple compression ratios. We conduct extensive experiments across in-domain and out-of-domain QA datasets, as well as across model families, scales, and compression ratios. Overall, our simple mean-pooling approach achieves the strongest performance, with a relatively small drop when training for multiple compression ratios. More broadly though, across architectures and training regimes the trade-offs are more nuanced, illustrating the complex landscape of compression methods.

The rapid progress of graph generation has raised new security concerns, particularly regarding backdoor vulnerabilities. While prior work has explored backdoor attacks in image diffusion and unconditional graph generation, conditional, especially text-guided graph generation remains largely unexamined. This paper proposes BadGraph, a backdoor attack method targeting latent diffusion models for text-guided graph generation. BadGraph leverages textual triggers to poison training data, covertly implanting backdoors that induce attacker-specified subgraphs during inference when triggers appear, while preserving normal performance on clean inputs. Extensive experiments on four benchmark datasets (PubChem, ChEBI-20, PCDes, MoMu) demonstrate the effectiveness and stealth of the attack: less than 10% poisoning rate can achieves 50% attack success rate, while 24% suffices for over 80% success rate, with negligible performance degradation on benign samples. Ablation studies further reveal that the backdoor is implanted during VAE and diffusion training rather than pretraining. These findings reveal the security vulnerabilities in latent diffusion models of text-guided graph generation, highlight the serious risks in models' applications such as drug discovery and underscore the need for robust defenses against the backdoor attack in such diffusion models.

Linear-attention models that compress the entire input sequence into a fixed-size recurrent state offer an efficient alternative to Transformers, but their finite memory induces forgetfulness that harms retrieval-intensive tasks. To mitigate the issue, we explore a series of hybrid models that restore direct access to past tokens. We interleave token mixers with intermediate time and space complexity between linear and full attention, including sparse attention with token eviction, and the query-aware native sparse attention. Particularly, we propose a novel learnable token eviction approach. Combined with sliding-window attention, an end-to-end trainable lightweight CNN aggregates information from both past and future adjacent tokens to adaptively retain a limited set of critical KV-pairs per head, maintaining linear attention's constant time and space complexity. Efficient Triton kernels for the sparse attention mechanisms are provided. Empirical evaluations on retrieval-intensive benchmarks support the effectiveness of our approaches.

Current methods for evaluating large language models (LLMs) typically focus on high-level tasks such as text generation, without targeting a particular AI application. This approach is not sufficient for evaluating LLMs for Responsible AI dimensions like fairness, since protected attributes that are highly relevant in one application may be less relevant in another. In this work, we construct a dataset that is driven by a real-world application (generate a plain-text product description, given a list of product features), parameterized by fairness attributes intersected with gendered adjectives and product categories, yielding a rich set of labeled prompts. We show how to use the data to identify quality, veracity, safety, and fairness gaps in LLMs, contributing a proposal for LLM evaluation paired with a concrete resource for the research community.

Recent advancements in large reasoning models (LRMs) have introduced an intermediate "thinking" process prior to generating final answers, improving their reasoning capabilities on complex downstream tasks. However, the potential of LRMs as evaluators for machine translation (MT) quality remains underexplored. We provides the first systematic analysis of LRM-as-a-judge in MT evaluation. We identify key challenges, revealing LRMs require tailored evaluation materials, tend to "overthink" simpler instances and have issues with scoring mechanisms leading to overestimation. To address these, we propose to calibrate LRM thinking by training them on synthetic, human-like thinking trajectories. Our experiments on WMT24 Metrics benchmarks demonstrate that this approach largely reduces thinking budgets by ~35x while concurrently improving evaluation performance across different LRM scales from 7B to 32B (e.g., R1-Distill-Qwen-7B achieves a +8.7 correlation point improvement). These findings highlight the potential of efficiently calibrated LRMs to advance fine-grained automatic MT evaluation.

We present Empathic Prompting, a novel framework for multimodal human-AI interaction that enriches Large Language Model (LLM) conversations with implicit non-verbal context. The system integrates a commercial facial expression recognition service to capture users' emotional cues and embeds them as contextual signals during prompting. Unlike traditional multimodal interfaces, empathic prompting requires no explicit user control; instead, it unobtrusively augments textual input with affective information for conversational and smoothness alignment. The architecture is modular and scalable, allowing integration of additional non-verbal modules. We describe the system design, implemented through a locally deployed DeepSeek instance, and report a preliminary service and usability evaluation (N=5). Results show consistent integration of non-verbal input into coherent LLM outputs, with participants highlighting conversational fluidity. Beyond this proof of concept, empathic prompting points to applications in chatbot-mediated communication, particularly in domains like healthcare or education, where users' emotional signals are critical yet often opaque in verbal exchanges.

We present a multi-agent, human-in-the-loop workflow that co-designs quantum codes with prescribed transversal diagonal gates. It builds on the Subset-Sum Linear Programming (SSLP) framework (arXiv:2504.20847), which partitions basis strings by modular residues and enforces $Z$-marginal Knill-Laflamme (KL) equalities via small LPs. The workflow is powered by GPT-5 and implemented within TeXRA (https://texra.ai)-a multi-agent research assistant platform that supports an iterative tool-use loop agent and a derivation-then-edit workflow reasoning agent. We work in a LaTeX-Python environment where agents reason, edit documents, execute code, and synchronize their work to Git/Overleaf. Within this workspace, three roles collaborate: a Synthesis Agent formulates the problem; a Search Agent sweeps/screens candidates and exactifies numerics into rationals; and an Audit Agent independently checks all KL equalities and the induced logical action. As a first step we focus on distance $d=2$ with nondegenerate residues. For code dimension $K\in\{2,3,4\}$ and $n\le6$ qubits, systematic sweeps yield certificate-backed tables cataloging attainable cyclic logical groups-all realized by new codes-e.g., for $K=3$ we obtain order $16$ at $n=6$. From verified instances, Synthesis Agent abstracts recurring structures into closed-form families and proves they satisfy the KL equalities for all parameters. It further demonstrates that SSLP accommodates residue degeneracy by exhibiting a new $((6,4,2))$ code implementing the transversal controlled-phase $diag(1,1,1,i)$. Overall, the workflow recasts diagonal-transversal feasibility as an analytical pipeline executed at scale, combining systematic enumeration with exact analytical reconstruction. It yields reproducible code constructions, supports targeted extensions to larger $K$ and higher distances, and leads toward data-driven classification.

Objective: Fluoropyrimidines are widely prescribed for colorectal and breast cancers, but are associated with toxicities such as hand-foot syndrome and cardiotoxicity. Since toxicity documentation is often embedded in clinical notes, we aimed to develop and evaluate natural language processing (NLP) methods to extract treatment and toxicity information. Materials and Methods: We constructed a gold-standard dataset of 236 clinical notes from 204,165 adult oncology patients. Domain experts annotated categories related to treatment regimens and toxicities. We developed rule-based, machine learning-based (Random Forest, Support Vector Machine [SVM], Logistic Regression [LR]), deep learning-based (BERT, ClinicalBERT), and large language models (LLM)-based NLP approaches (zero-shot and error-analysis prompting). Models used an 80:20 train-test split. Results: Sufficient data existed to train and evaluate 5 annotated categories. Error-analysis prompting achieved optimal precision, recall, and F1 scores (F1=1.000) for treatment and toxicities extraction, whereas zero-shot prompting reached F1=1.000 for treatment and F1=0.876 for toxicities extraction.LR and SVM ranked second for toxicities (F1=0.937). Deep learning underperformed, with BERT (F1=0.873 treatment; F1= 0.839 toxicities) and ClinicalBERT (F1=0.873 treatment; F1 = 0.886 toxicities). Rule-based methods served as our baseline with F1 scores of 0.857 in treatment and 0.858 in toxicities. Discussion: LMM-based approaches outperformed all others, followed by machine learning methods. Machine and deep learning approaches were limited by small training data and showed limited generalizability, particularly for rare categories. Conclusion: LLM-based NLP most effectively extracted fluoropyrimidine treatment and toxicity information from clinical notes, and has strong potential to support oncology research and pharmacovigilance.

Large language models (LLMs) have seen rapid adoption for tasks such as drafting emails, summarizing meetings, and answering health questions. In such uses, users may need to share private information (e.g., health records, contact details). To evaluate LLMs' ability to identify and redact such private information, prior work developed benchmarks (e.g., ConfAIde, PrivacyLens) with real-life scenarios. Using these benchmarks, researchers have found that LLMs sometimes fail to keep secrets private when responding to complex tasks (e.g., leaking employee salaries in meeting summaries). However, these evaluations rely on LLMs (proxy LLMs) to gauge compliance with privacy norms, overlooking real users' perceptions. Moreover, prior work primarily focused on the privacy-preservation quality of responses, without investigating nuanced differences in helpfulness. To understand how users perceive the privacy-preservation quality and helpfulness of LLM responses to privacy-sensitive scenarios, we conducted a user study with 94 participants using 90 scenarios from PrivacyLens. We found that, when evaluating identical responses to the same scenario, users showed low agreement with each other on the privacy-preservation quality and helpfulness of the LLM response. Further, we found high agreement among five proxy LLMs, while each individual LLM had low correlation with users' evaluations. These results indicate that the privacy and helpfulness of LLM responses are often specific to individuals, and proxy LLMs are poor estimates of how real users would perceive these responses in privacy-sensitive scenarios. Our results suggest the need to conduct user-centered studies on measuring LLMs' ability to help users while preserving privacy. Additionally, future research could investigate ways to improve the alignment between proxy LLMs and users for better estimation of users' perceived privacy and utility.

Minimum Bayes Risk (MBR) decoding has seen renewed interest as an alternative to traditional generation strategies. While MBR has proven effective in machine translation, where the variability of a language model's outcome space is naturally constrained, it may face challenges in more open-ended tasks such as dialogue or instruction-following. We hypothesise that in such settings, applying MBR with standard similarity-based utility functions may result in selecting responses that are broadly representative of the model's distribution, yet sub-optimal with respect to any particular grouping of generations that share an underlying latent structure. In this work, we introduce three lightweight adaptations to the utility function, designed to make MBR more sensitive to structural variability in the outcome space. To test our hypothesis, we curate a dataset capturing three representative types of latent structure: dialogue act, emotion, and response structure (e.g., a sentence, a paragraph, or a list). We further propose two metrics to evaluate the structural optimality of MBR. Our analysis demonstrates that common similarity-based utility functions fall short by these metrics. In contrast, our proposed adaptations considerably improve structural optimality. Finally, we evaluate our approaches on real-world instruction-following benchmarks, AlpacaEval and MT-Bench, and show that increased structural sensitivity improves generation quality by up to 13.7 percentage points in win rate.

Language models continue to hallucinate despite increases in parameters, compute, and data. We propose neural diversity -- decorrelated parallel representations -- as a principled mechanism that reduces hallucination rates at fixed parameter and data budgets. Inspired by portfolio theory, where uncorrelated assets reduce risk by $\sqrt{P}$, we prove hallucination probability is bounded by representational correlation: $P(H) \leq f(\sigma^2((1-\rho(P))/P + \rho(P)), \mu^2)$, which predicts that language models need an optimal amount of neurodiversity. To validate this, we introduce ND-LoRA (Neural Diversity Low-Rank Adaptation), combining parallel LoRA adapters with Barlow Twins regularization, and demonstrate that ND-LoRA reduces hallucinations by up to 25.6% (and 14.6% on average) without degrading general accuracy. Ablations show LoRA adapters and regularization act synergistically, causal interventions prove neurodiversity as the mediating factor and correlational analyses indicate scale: a 0.1% neural correlation increase is associated with a 3.8% hallucination increase. Finally, task-dependent optimality emerges: different tasks require different amounts of optimal neurodiversity. Together, our results highlight neural diversity as a third axis of scaling -- orthogonal to parameters and data -- to improve the reliability of language models at fixed budgets.

This study presents the multilingual e-commerce search system developed by the Tredence_AICOE team. The competition features two multilingual relevance tasks: Query-Category (QC) Relevance, which evaluates how well a user's search query aligns with a product category, and Query-Item (QI) Relevance, which measures the match between a multilingual search query and an individual product listing. To ensure full language coverage, we performed data augmentation by translating existing datasets into languages missing from the development set, enabling training across all target languages. We fine-tuned Gemma-3 12B and Qwen-2.5 14B model for both tasks using multiple strategies. The Gemma-3 12B (4-bit) model achieved the best QC performance using original and translated data, and the best QI performance using original, translated, and minority class data creation. These approaches secured 4th place on the final leaderboard, with an average F1-score of 0.8857 on the private test set.

Cantonese, although spoken by millions, remains under-resourced due to policy and diglossia. To address this scarcity of evaluation frameworks for Cantonese, we introduce \textsc{\textbf{CantoNLU}}, a benchmark for Cantonese natural language understanding (NLU). This novel benchmark spans seven tasks covering syntax and semantics, including word sense disambiguation, linguistic acceptability judgment, language detection, natural language inference, sentiment analysis, part-of-speech tagging, and dependency parsing. In addition to the benchmark, we provide model baseline performance across a set of models: a Mandarin model without Cantonese training, two Cantonese-adapted models obtained by continual pre-training a Mandarin model on Cantonese text, and a monolingual Cantonese model trained from scratch. Results show that Cantonese-adapted models perform best overall, while monolingual models perform better on syntactic tasks. Mandarin models remain competitive in certain settings, indicating that direct transfer may be sufficient when Cantonese domain data is scarce. We release all datasets, code, and model weights to facilitate future research in Cantonese NLP.

Large Reasoning Models (LRMs) achieve strong performance on mathematical, scientific, and other question-answering tasks, but their multilingual reasoning abilities remain underexplored. When presented with non-English questions, LRMs often default to reasoning in English, raising concerns about interpretability and the handling of linguistic and cultural nuances. We systematically compare an LRM's reasoning in English versus the language of the question. Our evaluation spans two tasks: MGSM and GPQA Diamond. Beyond measuring answer accuracy, we also analyze cognitive attributes in the reasoning traces. We find that English reasoning traces exhibit a substantially higher presence of these cognitive behaviors, and that reasoning in English generally yields higher final-answer accuracy, with the performance gap increasing as tasks become more complex. However, this English-centric strategy is susceptible to a key failure mode - getting "Lost in Translation," where translation steps lead to errors that would have been avoided by question's language reasoning.

Curiosity serves as a pivotal conduit for human beings to discover and learn new knowledge. Recent advancements of large language models (LLMs) in natural language processing have sparked discussions regarding whether these models possess capability of curiosity-driven learning akin to humans. In this paper, starting from the human curiosity assessment questionnaire Five-Dimensional Curiosity scale Revised (5DCR), we design a comprehensive evaluation framework that covers dimensions such as Information Seeking, Thrill Seeking, and Social Curiosity to assess the extent of curiosity exhibited by LLMs. The results demonstrate that LLMs exhibit a stronger thirst for knowledge than humans but still tend to make conservative choices when faced with uncertain environments. We further investigated the relationship between curiosity and thinking of LLMs, confirming that curious behaviors can enhance the model's reasoning and active learning abilities. These findings suggest that LLMs have the potential to exhibit curiosity similar to that of humans, providing experimental support for the future development of learning capabilities and innovative research in LLMs.

This paper details our submission to the Ara- GenEval Shared Task on Arabic AI-generated text detection, where our team, BUSTED, se- cured 5th place. We investigated the effec- tiveness of three pre-trained transformer mod- els: AraELECTRA, CAMeLBERT, and XLM- RoBERTa. Our approach involved fine-tuning each model on the provided dataset for a binary classification task. Our findings revealed a sur- prising result: the multilingual XLM-RoBERTa model achieved the highest performance with an F1 score of 0.7701, outperforming the spe- cialized Arabic models. This work underscores the complexities of AI-generated text detection and highlights the strong generalization capa- bilities of multilingual models.

Evaluating large language models (LLMs) on final-answer correctness is the dominant paradigm. This approach, however, provides a coarse signal for model improvement and overlooks the quality of the underlying reasoning process. We argue that a more granular evaluation of reasoning offers a more effective path to building robust models. We decompose reasoning quality into two dimensions: relevance and coherence. Relevance measures if a step is grounded in the problem; coherence measures if it follows logically from prior steps. To measure these aspects reliably, we introduce causal stepwise evaluation (CaSE). This method assesses each reasoning step using only its preceding context, which avoids hindsight bias. We validate CaSE against human judgments on our new expert-annotated benchmarks, MRa-GSM8K and MRa-MATH. More importantly, we show that curating training data with CaSE-evaluated relevance and coherence directly improves final task performance. Our work provides a scalable framework for analyzing, debugging, and improving LLM reasoning, demonstrating the practical value of moving beyond validity checks.