Adaptive Backtracking for Privacy Protection in Large Language Models

The preservation of privacy has emerged as a critical topic in the era of artificial intelligence. However, current work focuses on user-oriented privacy, overlooking severe enterprise data leakage risks exacerbated by the Retrieval-Augmented Generation paradigm. To address this gap, our paper introduces a novel objective: enterprise-oriented privacy concerns. Achieving this objective requires overcoming two fundamental challenges: existing methods such as data sanitization severely degrade model performance, and the field lacks public datasets for evaluation. We address these challenges with several solutions. (1) To prevent performance degradation, we propose ABack, a training-free mechanism that leverages a Hidden State Model to pinpoint the origin of a leakage intention and rewrite the output safely. (2) To solve the lack of datasets, we construct PriGenQA, a new benchmark for enterprise privacy scenarios in healthcare and finance. To ensure a rigorous evaluation, we move beyond simple static attacks by developing a powerful adaptive attacker with Group Relative Policy Optimization. Experiments show that against this superior adversary, ABack improves the overall privacy utility score by up to 15\% over strong baselines, avoiding the performance trade-offs of prior methods.

Bimanual Robot-Assisted Dressing: A Spherical Coordinate-Based Strategy for Tight-Fitting Garments

Robot-assisted dressing is a popular but challenging topic in the field of robotic manipulation, offering significant potential to improve the quality of life for individuals with mobility limitations. Currently, the majority of research on robot-assisted dressing focuses on how to put on loose-fitting clothing, with little attention paid to tight garments. For the former, since the armscye is larger, a single robotic arm can usually complete the dressing task successfully. However, for the latter, dressing with a single robotic arm often fails due to the narrower armscye and the property of diminishing rigidity in the armscye, which eventually causes the armscye to get stuck. This paper proposes a bimanual dressing strategy suitable for dressing tight-fitting clothing. To facilitate the encoding of dressing trajectories that adapt to different human arm postures, a spherical coordinate system for dressing is established. We uses the azimuthal angle of the spherical coordinate system as a task-relevant feature for bimanual manipulation. Based on this new coordinate, we employ Gaussian Mixture Model (GMM) and Gaussian Mixture Regression (GMR) for imitation learning of bimanual dressing trajectories, generating dressing strategies that adapt to different human arm postures. The effectiveness of the proposed method is validated through various experiments.

Vec2Summ: Text Summarization via Probabilistic Sentence Embeddings

We propose Vec2Summ, a novel method for abstractive summarization that frames the task as semantic compression. Vec2Summ represents a document collection using a single mean vector in the semantic embedding space, capturing the central meaning of the corpus. To reconstruct fluent summaries, we perform embedding inversion -- decoding this mean vector into natural language using a generative language model. To improve reconstruction quality and capture some degree of topical variability, we introduce stochasticity by sampling from a Gaussian distribution centered on the mean. This approach is loosely analogous to bagging in ensemble learning, where controlled randomness encourages more robust and varied outputs. Vec2Summ addresses key limitations of LLM-based summarization methods. It avoids context-length constraints, enables interpretable and controllable generation via semantic parameters, and scales efficiently with corpus size -- requiring only $O(d + d^2)$ parameters. Empirical results show that Vec2Summ produces coherent summaries for topically focused, order-invariant corpora, with performance comparable to direct LLM summarization in terms of thematic coverage and efficiency, albeit with less fine-grained detail. These results underscore Vec2Summ's potential in settings where scalability, semantic control, and corpus-level abstraction are prioritized.

A Hybrid AI Methodology for Generating Ontologies of Research Topics from Scientific Paper Corpora

Taxonomies and ontologies of research topics (e.g., MeSH, UMLS, CSO, NLM) play a central role in providing the primary framework through which intelligent systems can explore and interpret the literature. However, these resources have traditionally been manually curated, a process that is time-consuming, prone to obsolescence, and limited in granularity. This paper presents Sci-OG, a semi-auto\-mated methodology for generating research topic ontologies, employing a multi-step approach: 1) Topic Discovery, extracting potential topics from research papers; 2) Relationship Classification, determining semantic relationships between topic pairs; and 3) Ontology Construction, refining and organizing topics into a structured ontology. The relationship classification component, which constitutes the core of the system, integrates an encoder-based language model with features describing topic occurrence in the scientific literature. We evaluate this approach against a range of alternative solutions using a dataset of 21,649 manually annotated semantic triples. Our method achieves the highest F1 score (0.951), surpassing various competing approaches, including a fine-tuned SciBERT model and several LLM baselines, such as the fine-tuned GPT4-mini. Our work is corroborated by a use case which illustrates the practical application of our system to extend the CSO ontology in the area of cybersecurity. The presented solution is designed to improve the accessibility, organization, and analysis of scientific knowledge, thereby supporting advancements in AI-enabled literature management and research exploration.

Learning the Topic, Not the Language: How LLMs Classify Online Immigration Discourse Across Languages

Large language models (LLMs) are transforming social-science research by enabling scalable, precise analysis. Their adaptability raises the question of whether knowledge acquired through fine-tuning in a few languages can transfer to unseen languages that only appeared during pre-training. To examine this, we fine-tune lightweight LLaMA 3.2-3B models on monolingual, bilingual, or multilingual data sets to classify immigration-related tweets from X/Twitter across 13 languages, a domain characterised by polarised, culturally specific discourse. We evaluate whether minimal language-specific fine-tuning enables cross-lingual topic detection and whether adding targeted languages corrects pre-training biases. Results show that LLMs fine-tuned in one or two languages can reliably classify immigration-related content in unseen languages. However, identifying whether a tweet expresses a pro- or anti-immigration stance benefits from multilingual fine-tuning. Pre-training bias favours dominant languages, but even minimal exposure to under-represented languages during fine-tuning (as little as $9.62\times10^{-11}$ of the original pre-training token volume) yields significant gains. These findings challenge the assumption that cross-lingual mastery requires extensive multilingual training: limited language coverage suffices for topic-level generalisation, and structural biases can be corrected with lightweight interventions. By releasing 4-bit-quantised, LoRA fine-tuned models, we provide an open-source, reproducible alternative to proprietary LLMs that delivers 35 times faster inference at just 0.00000989% of the dollar cost of the OpenAI GPT-4o model, enabling scalable, inclusive research.

Evaluating Podcast Recommendations with Profile-Aware LLM-as-a-Judge

Evaluating personalized recommendations remains a central challenge, especially in long-form audio domains like podcasts, where traditional offline metrics suffer from exposure bias and online methods such as A/B testing are costly and operationally constrained. In this paper, we propose a novel framework that leverages Large Language Models (LLMs) as offline judges to assess the quality of podcast recommendations in a scalable and interpretable manner. Our two-stage profile-aware approach first constructs natural-language user profiles distilled from 90 days of listening history. These profiles summarize both topical interests and behavioral patterns, serving as compact, interpretable representations of user preferences. Rather than prompting the LLM with raw data, we use these profiles to provide high-level, semantically rich context-enabling the LLM to reason more effectively about alignment between a user's interests and recommended episodes. This reduces input complexity and improves interpretability. The LLM is then prompted to deliver fine-grained pointwise and pairwise judgments based on the profile-episode match. In a controlled study with 47 participants, our profile-aware judge matched human judgments with high fidelity and outperformed or matched a variant using raw listening histories. The framework enables efficient, profile-aware evaluation for iterative testing and model selection in recommender systems.

HealthBranches: Synthesizing Clinically-Grounded Question Answering Datasets via Decision Pathways

HealthBranches is a novel benchmark dataset for medical Question-Answering (Q&A), specifically designed to evaluate complex reasoning in Large Language Models (LLMs). This dataset is generated through a semi-automated pipeline that transforms explicit decision pathways from medical source into realistic patient cases with associated questions and answers. Covering 4,063 case studies across 17 healthcare topics, each data point is based on clinically validated reasoning chains. HealthBranches supports both open-ended and multiple-choice question formats and uniquely includes the full reasoning path for each Q&A. Its structured design enables robust evaluation of LLMs' multi-step inference capabilities, including their performance in structured Retrieval-Augmented Generation (RAG) contexts. HealthBranches establishes a foundation for the development of more trustworthy, interpretable, and clinically reliable LLMs in high-stakes domains while also serving as a valuable resource for educational purposes.

SEADialogues: A Multilingual Culturally Grounded Multi-turn Dialogue Dataset on Southeast Asian Languages

Although numerous datasets have been developed to support dialogue systems, most existing chit-chat datasets overlook the cultural nuances inherent in natural human conversations. To address this gap, we introduce SEADialogues, a culturally grounded dialogue dataset centered on Southeast Asia, a region with over 700 million people and immense cultural diversity. Our dataset features dialogues in eight languages from six Southeast Asian countries, many of which are low-resource despite having sizable speaker populations. To enhance cultural relevance and personalization, each dialogue includes persona attributes and two culturally grounded topics that reflect everyday life in the respective communities. Furthermore, we release a multi-turn dialogue dataset to advance research on culturally aware and human-centric large language models, including conversational dialogue agents.

Combining Language and Topic Models for Hierarchical Text Classification

Hierarchical text classification (HTC) is a natural language processing task which has the objective of categorising text documents into a set of classes from a predefined structured class hierarchy. Recent HTC approaches use various techniques to incorporate the hierarchical class structure information with the natural language understanding capabilities of pre-trained language models (PLMs) to improve classification performance. Furthermore, using topic models along with PLMs to extract features from text documents has been shown to be an effective approach for multi-label text classification tasks. The rationale behind the combination of these feature extractor models is that the PLM captures the finer-grained contextual and semantic information while the topic model obtains high-level representations which consider the corpus of documents as a whole. In this paper, we use a HTC approach which uses a PLM and a topic model to extract features from text documents which are used to train a classification model. Our objective is to determine whether the combination of the features extracted from the two models is beneficial to HTC performance in general. In our approach, the extracted features are passed through separate convolutional layers whose outputs are combined and passed to a label-wise attention mechanisms which obtains label-specific document representations by weighing the most important features for each class separately. We perform comprehensive experiments on three HTC benchmark datasets and show that using the features extracted from the topic model generally decreases classification performance compared to only using the features obtained by the PLM. In contrast to previous work, this shows that the incorporation of features extracted from topic models for text classification tasks should not be assumed beneficial.

Prompt-Response Semantic Divergence Metrics for Faithfulness Hallucination and Misalignment Detection in Large Language Models

The proliferation of Large Language Models (LLMs) is challenged by hallucinations, critical failure modes where models generate non-factual, nonsensical or unfaithful text. This paper introduces Semantic Divergence Metrics (SDM), a novel lightweight framework for detecting Faithfulness Hallucinations -- events of severe deviations of LLMs responses from input contexts. We focus on a specific implementation of these LLM errors, {confabulations, defined as responses that are arbitrary and semantically misaligned with the user's query. Existing methods like Semantic Entropy test for arbitrariness by measuring the diversity of answers to a single, fixed prompt. Our SDM framework improves upon this by being more prompt-aware: we test for a deeper form of arbitrariness by measuring response consistency not only across multiple answers but also across multiple, semantically-equivalent paraphrases of the original prompt. Methodologically, our approach uses joint clustering on sentence embeddings to create a shared topic space for prompts and answers. A heatmap of topic co-occurances between prompts and responses can be viewed as a quantified two-dimensional visualization of the user-machine dialogue. We then compute a suite of information-theoretic metrics to measure the semantic divergence between prompts and responses. Our practical score, $\mathcal{S}_H$, combines the Jensen-Shannon divergence and Wasserstein distance to quantify this divergence, with a high score indicating a Faithfulness hallucination. Furthermore, we identify the KL divergence KL(Answer $||$ Prompt) as a powerful indicator of \textbf{Semantic Exploration}, a key signal for distinguishing different generative behaviors. These metrics are further combined into the Semantic Box, a diagnostic framework for classifying LLM response types, including the dangerous, confident confabulation.