Mounting evidence suggests that per- and polyfluoroalkyl substances (PFASs), are key drivers of glycometabolism disorders. However, the extent to which this association persists in the elderly population, as well as the underlying mechanisms involved, remains insufficiently characterized. This study aimed to comprehensively investigate associations between PFASs exposure and glycometabolic markers in the healthy elderly using integrated multiomics approach, elucidating underlying biological mechanisms. A panel study was conducted among 76 healthy elderly aged 60-69 years living in Jinan City, China, with five consecutive monthly surveys. Blood concentrations of nine PFASs and five glycometabolic markers were measured. Associations between both individual and combined exposures to PFASs and glycometabolic markers were investigated through a linear mixed-effects model (LMM) and Bayesian kernel machine regression (BKMR). Integrated multiomics analysis delineated mechanisms by which PFASs exposure disrupts glycometabolism. We observed that a single exposure to perfluorobutanoic acid (PFBA) was significantly associated with increases in all five glycometabolic markers. Exposure to perfluorodecanoic acid (PFDA) and perfluorohexanesulfonic acid (PFHxS) was associated with increases in fasting plasma glucose (FPG) and glycated serum protein (GSP), while PFHxS was additionally associated with the triglyceride-glucose index (TyG). Furthermore, the BKMR model further indicated a synergistic and positive association between PFAS mixtures and these markers. Multiomics analyses identified 48 canonical pathways potentially mediating these metabolic disruptions. These findings highlight the need to prioritize reduction of PFASs exposure as a public health strategy for preventing and managing type 2 diabetes mellitus (T2D), particularly in aging populations worldwide.

The safety alignment of Large Language Models (LLMs) remains vulnerable to Harmful Fine-tuning (HFT). While existing defenses impose constraints on parameters, gradients, or internal representations, we observe that they can be effectively circumvented under persistent HFT. Our analysis traces this failure to the inherent redundancy of the high-dimensional parameter space: attackers exploit optimization trajectories that are orthogonal to defense constraints to restore harmful capabilities while deceptively adhering to safety restrictions. To address this, we propose Safety Bottleneck Regularization (SBR). SBR shifts the defensive focus from the redundant parameter space to the unembedding layer, which serves as a geometric bottleneck. By anchoring the final hidden states of harmful queries to those of the safety-aligned model, SBR enables the model to maintain safe responses even under persistent HFT. Extensive experiments confirm SBR's effectiveness, demonstrating that utilizing just a single safety anchor is sufficient to reduce the Harmful Score to $<$10 while preserving competitive performance on benign downstream tasks.

Polybrominated diphenyl ethers (PBDEs) are implicated in dyslipidemia, but the molecular basis of individual susceptibility remains elusive. Here we report an analysis based on the China National Human Biomonitoring cohort, where we integrate exposome, genomic, and metabolomic data to identify 3,571 genetic variants that interact with PBDE exposure to influence dyslipidemia risk. Metabolomic analysis highlights glycine and glycerophosphate as key mediators. A polygenic risk score derived from these PBDE-interactive variants significantly enhances dyslipidemia prediction in highly exposed individuals. Among these, rs9869609 emerges as a candidate causal variant, showing the strongest association with hypercholesterolemia risk (β = 1.18, FDR = 0.0078). Further functional validation using single-base CRISPR/Cas9 editing reveals that the rs9869609-G allele downregulates SLC6A20 expression by strengthening BHLHE40 binding, which further impairs glycine transport and promotes cholesterol accumulation, particularly under 2,2',4,4'-Tetrabromodiphenyl ether exposure. Collectively, our study elucidates a gene-environment interaction mechanism through which genetic variants modulate lipid metabolism in response to PBDE exposure.

Recent advances in large language models (LLMs) promise more effective information extraction for review-based recommender systems, yet current methods still (i) mine free-form reviews without scope control, producing redundant and noisy representations, (ii) lack principled metrics that link LLM hallucination to downstream effectiveness, and (iii) leave the cost-quality trade-off across model scales largely unexplored. We address these gaps with the Hyper-Adaptive Dual-Stage Semantic Framework (HADSF), a two-stage approach that first induces a compact, corpus-level aspect vocabulary via adaptive selection and then performs vocabulary-guided, explicitly constrained extraction of structured aspect-opinion triples. To assess the fidelity of the resulting representations, we introduce Aspect Drift Rate (ADR) and Opinion Fidelity Rate (OFR) and empirically uncover a nonmonotonic relationship between hallucination severity and rating prediction error. Experiments on approximately 3 million reviews across LLMs spanning 1.5B-70B parameters show that, when integrated into standard rating predictors, HADSF yields consistent reductions in prediction error and enables smaller models to achieve competitive performance in representative deployment scenarios. We release code, data pipelines, and metric implementations to support reproducible research on hallucination-aware, LLM-enhanced explainable recommendation. Code is available at https://github.com/niez233/HADSF

Recent advancements in Text-to-Speech (TTS) technology have been remarkable, enabling current models to clone arbitrary unseen speakers and synthesize high-quality, natural-sounding speech. However, corresponding evaluation techniques appear to be lagging: existing Mean Opinion Score (MOS) estimation models typically perform regression-based scoring on entire speech segments, while a failed synthesized speech usually contains problematic elements in only a few isolated words rather than throughout the entire utterance. In this context, we present an intriguing finding: encoder-decoder ASR models, such as Whisper, leverage their extensive pre-training to precisely capture word-level mismatches between speech and text within their cross-attention mechanisms, thereby providing a fine-grained reward signal. Building upon this insight, we propose a novel TTS optimization method, which we term Word-level TTS Alignment by ASR-driven Attentive Reward (W3AR). Instead of relying on any explicit reward annotations, W3AR leverages the attention information within a pre-trained ASR model, enabling finer-grained alignment and optimization of the sequences predicted by the TTS model. Experimental results demonstrate that W3AR not only effectively improves the TTS generation quality of existing models but also further enhances zero-shot robustness based on both in-domain and out-of-domain prompt speakers. Additionally, our findings and proposed methodology offer a new insight for generative tasks: understanding models can potentially serve as evaluators, providing highly fine-grained and valuable feedback for generative optimization.

Multimodal sequential recommendation (MSR) leverages diverse item modalities to improve recommendation accuracy, while achieving effective and adaptive fusion remains challenging. Existing MSR models often overlook synergistic information that emerges only through modality combinations. Moreover, they typically assume a fixed importance for different modality interactions across users. To address these limitations, we propose \textbf{P}ersonalized \textbf{R}ecommend-ation via \textbf{I}nformation \textbf{S}ynergy \textbf{M}odule (PRISM), a plug-and-play framework for sequential recommendation (SR). PRISM explicitly decomposes multimodal information into unique, redundant, and synergistic components through an Interaction Expert Layer and dynamically weights them via an Adaptive Fusion Layer guided by user preferences. This information-theoretic design enables fine-grained disentanglement and personalized fusion of multimodal signals. Extensive experiments on four datasets and three SR backbones demonstrate its effectiveness and versatility. The code is available at https://github.com/YutongLi2024/PRISM.

Safety alignment instills in Large Language Models (LLMs) a critical capacity to refuse malicious requests. Prior works have modeled this refusal mechanism as a single linear direction in the activation space. We posit that this is an oversimplification that conflates two functionally distinct neural processes: the detection of harm and the execution of a refusal. In this work, we deconstruct this single representation into a Harm Detection Direction and a Refusal Execution Direction. Leveraging this fine-grained model, we introduce Differentiated Bi-Directional Intervention (DBDI), a new white-box framework that precisely neutralizes the safety alignment at critical layer. DBDI applies adaptive projection nullification to the refusal execution direction while suppressing the harm detection direction via direct steering. Extensive experiments demonstrate that DBDI outperforms prominent jailbreaking methods, achieving up to a 97.88% attack success rate on models such as Llama-2. By providing a more granular and mechanistic framework, our work offers a new direction for the in-depth understanding of LLM safety alignment.

Short videos have gained immense popularity, necessitating effective recommender systems that cater to individual preferences. The platforms use advanced algorithms to analyze user engagement and provide videos that satisfy users. A critical factor in user satisfaction is immersion, defined as the feeling of being deeply engaged in an activity. However, existing recommendation algorithms in the short video scenario have largely disregarded user immersion. Our study integrates user immersion into recommendation systems, aiming to predict immersion from user interactions and recommend items to enhance the overall viewing experience. Based on the user study of collecting and analyzing user immersion, we integrate immersion into the recommendations for both lab and large-scale scenarios. We adapt user-annotated immersion to large-scale real-world datasets without immersion labels. Specifically, we propose ImmersRec, an immersion-aware recommendation framework with immersion prediction fine-tuning, immersion knowledge alignment, and immersion-enhanced recommendation. Extensive experiments on two short video platforms indicate that our approach achieves significant enhancements among various context-aware recommender backbones. We investigate the predicted immersion and find it impacts not only short-term utility but also long-term user engagement. This research pioneers the incorporation of user immersion in short video recommendation algorithms, emphasizing its potential for improving recommendations with minimal data. The code can be available at https://github.com/hezy18/ImmersRec.

Personal care products (PCPs) are ubiquitously present in the environment, and the associated health risks have been increasingly concerned worldwide. However, knowledge regarding exposure assessments of older adults to these chemicals and their health risks remains largely limited. In the present study, five repeated surveys involving 76 healthy older adults in Jinan, Shandong Province, were performed to quantify urinary exposure levels of 14 chemicals in PCPs. Moreover, influencing factors and health risks associated with exposure to these chemicals were thoroughly analyzed. Our findings revealed that methyl paraben (MeP) was the predominant chemical in PCP in the urine of the elderly, with a median concentration of 16.17 μg/L. Dietary intake, particularly fish and milk products, along with exposure to ambient PM2.5, were identified as the primary sources of certain chemicals in PCPs. Additionally, an increased physical activity was associated with decreased concentrations of benzophenone-2 (BP-2) within the body. Risk assessment demonstrated that chemicals in PCPs currently posed minimal health risks to the elderly. Our findings provide substantial references for mitigating the health risks of these chemical exposures in healthy older adults, ultimately safeguarding their overall and physical well-being.

When users interact with Recommender Systems (RecSys), current situations, such as time, location, and environment, significantly influence their preferences. Situations serve as the background for interactions, where relationships between users and items evolve with situation changes. However, existing RecSys treat situations, users, and items on the same level. They can only model the relations between situations and users/items respectively, rather than the dynamic impact of situations on user-item associations (i.e., user preferences). In this paper, we provide a new perspective that takes situations as the preconditions for users' interactions. This perspective allows us to separate situations from user/item representations, and capture situations' influences over the user-item relationship, offering a more comprehensive understanding of situations. Based on it, we propose a novel Situation-Aware Recommender Enhancer (SARE), a pluggable module to integrate situations into various existing RecSys. Since users' perception of situations and situations' impact on preferences are both personalized, SARE includes a Personalized Situation Fusion (PSF) and a User-Conditioned Preference Encoder (UCPE) to model the perception and impact of situations, respectively. We conduct experiments of applying SARE on seven backbones in various settings on two real-world datasets. Experimental results indicate that SARE improves the recommendation performances significantly compared with backbones and SOTA situation-aware baselines.

Cross-domain recommendation (CDR) has emerged as a promising direction for expanding the applicability of recommendation systems. Recent advances in CDR have demonstrated the effectiveness of the unified distribution paradigm, which leverages shared distributions to transfer knowledge across domains and employs domain-specific adapters for targeted recommendations. While this well-designed paradigm shows promising performance, existing methods require extra supervision signals (e.g. contrastive learning on domain-masked embeddings) to maintain unified distributions across domains, leading to an inherent trade-off between unified objectives and domain-specific preference modeling. To address these limitations, we propose CD-CDR (Conditional Diffusion-CDR), a novel approach that leverages a shared conditional diffusion model to learn unified item distributions and facilitate knowledge transfer across domains. The key insight is to utilize the powerful generative capabilities of diffusion models to learn a shared distribution while naturally incorporating domain-specific characteristics through conditional generation. This design enables CD-CDR to replace traditional adapters with generation conditions as an integral part of the distribution model, thereby eliminating extra supervision signals and fundamentally resolving the trade-off between unified and domain-specific objectives. Extensive experiments on six domain pairs from two real-world datasets demonstrate that CD-CDR significantly outperforms existing methods for both normal and cold-start settings. To the best of our knowledge, this is the first work to explore the unified distribution paradigm in CDR using conditional diffusion models.

Emerging evidence suggests that exposure to polycyclic aromatic hydrocarbons (PAHs) may exacerbate cognitive deterioration and biological aging, but the biomolecular mechanisms, particularly those driving sex-specific vulnerability, remain unclear. This 5-month longitudinal panel study, conducted among 76 healthy participants aged 60-69 in Jinan, China. Urinary concentrations of six PAHs were measured, and biological age was estimated using the Klemera and Doubal method. The associations between individual and combined PAH exposures and cognitive function were examined using linear mixed-effects model, and the mediating effect of biological aging was assessed by mediation model. Additionally, lipidomic profiling including lipidome clustering and individual lipids helped identify key biomolecular mediators. The results showed that exposure to four PAHs was significantly associated with cognitive decline, with combined exposure analysis further indicated the detrimental effects of PAH mixtures. And biological aging partially mediated the association between 2-hydroxynaphthalene (2-NAP) and cognitive function. Key lipid levels, particularly phosphatidylethanolamines (PEs)-dominated clustering pattern, emerged as pivotal molecular mediator shaping this relationship. Additionally, more lipid species, notably phosphatidylcholines (PCs), mediated PAHs-induced cognitive decline in females. These findings elucidate lipid levels as a key mechanism linking biological aging and PAHs-induced cognitive decline, with specific lipid alterations partially explaining the increased vulnerability of females to cognitive decline.

Formed through the partial burning of carbon-based materials, polycyclic aromatic hydrocarbons (PAHs) represent a prevalent class of environmental contaminants associated with significant human health risks. However, comprehensive exposure and health risk assessment remain lacked, specifically among the elderly population. This research involved carrying out five sequential surveys across a five-month period, with 76 healthy elderly individuals participating from Jinan, Shandong Province, to quantify urinary concentrations of 16 monohydroxylated PAHs (OH-PAHs) and their associated health risks. To investigate determinants of PAHs exposure, we utilized Spearman correlation and linear mixed-effects modeling techniques. Among the 16 OH-PAHs, 1-hydroxynaphthalene (1-NAP) demonstrated the peak median urinary concentration at 1.78 μg/g creatinine (Cr). Urinary OH-PAHs concentrations were significantly influenced by demographic factors and dietary habits. The parameter showed a negative association with educational level and alcohol intake, yet aligned positively with the consumption of drinking water, rice, millets, corn products, and thermally processed fried foods. Hazard index (HI) analysis showed that no potential non-carcinogenic risks from PAH exposures among healthy older adults in this study. Furthermore, carcinogenic risk assessment demonstrated that all participants exceeded the upper safety threshold of 10-4, highlighting carcinogenicity as the predominant health concern. The findings of this study provide key scientific evidence for formulating targeted PAHs exposure mitigation policies, especially in the dietary exposure route. They provide novel insights to inform public health decision-making, enabling the optimization of dietary intervention strategies to reduce health risks associated with PAHs exposure among the elderly population, thereby contributing to the goal of healthy aging.

Providing reasonable explanations for a specific suggestion given by the recommender can help users trust the system more. As logic rule-based inference is concise, transparent, and aligned with human cognition, it can be adopted to improve the interpretability of recommendation models. Previous work that interprets user preference with logic rules merely focuses on the construction of rules while neglecting the usage of feature embeddings. This limits the model in capturing implicit relationships between features. In this paper, we aim to improve both the effectiveness and explainability of recommendation models by simultaneously representing logic rules and feature embeddings. We propose a novel model-intrinsic explainable recommendation method named Feature-Enhanced Neural Collaborative Reasoning ( FENCR ). The model automatically extracts representative logic rules from massive possibilities in a data-driven way. In addition, we utilize feature interaction-based neural modules to represent logic operators on embeddings. Experiments on two large public datasets show our model outperforms state-of-the-art neural logical recommendation models. Further case analyses demonstrate that FENCR can derive reasonable rules, indicating its high robustness and expandability 1 .

Cross-domain recommender (CDR) systems aim to enhance the performance of the target domain by utilizing data from other related domains. However, irrelevant information from the source domain may instead degrade target domain performance, which is known as the negative transfer problem. There have been some attempts to address this problem, mostly by designing adaptive representations for overlapped users. Whereas, representation adaptions solely rely on the expressive capacity of the CDR model, lacking explicit constraint to filter the irrelevant source-domain collaborative information for the target domain. In this paper, we propose a novel Collaborative information regularized User Transformation (CUT) framework to tackle the negative transfer problem by directly filtering users' collaborative information. In CUT, user similarity in the target domain is adopted as a constraint for user transformation learning to filter the user collaborative information from the source domain. CUT first learns user similarity relationships from the target domain. Then, source-target information transfer is guided by the user similarity, where we design a user transformation layer to learn target-domain user representations and a contrastive loss to supervise the user collaborative information transferred. The results show significant performance improvement of CUT compared with SOTA single and cross-domain methods. Further analysis of the target-domain results illustrates that CUT can effectively alleviate the negative transfer problem.

Sequential recommender systems predict items that may interest users by modeling their preferences based on historical interactions. Traditional sequential recommendation methods rely on capturing implicit collaborative filtering signals among items. Recent relation-aware sequential recommendation models have achieved promising performance by explicitly incorporating item relations into the modeling of user historical sequences, where most relations are extracted from knowledge graphs. However, existing methods rely on manually predefined relations and suffer the sparsity issue, limiting the generalization ability in diverse scenarios with varied item relations. In this paper, we propose a novel relation-aware sequential recommendation framework with Latent Relation Discovery (LRD). Different from previous relation-aware models that rely on predefined rules, we propose to leverage the Large Language Model (LLM) to provide new types of relations and connections between items. The motivation is that LLM contains abundant world knowledge, which can be adopted to mine latent relations of items for recommendation. Specifically, inspired by that humans can describe relations between items using natural language, LRD harnesses the LLM that has demonstrated human-like knowledge to obtain language knowledge representations of items. These representations are fed into a latent relation discovery module based on the discrete state variational autoencoder (DVAE). Then the self-supervised relation discovery tasks and recommendation tasks are jointly optimized. Experimental results on multiple public datasets demonstrate our proposed latent relations discovery method can be incorporated with existing relation-aware sequential recommendation models and significantly improve the performance. Further analysis experiments indicate the effectiveness and reliability of the discovered latent relations.

BACKGROUND:Evidence suggested that abiotic airborne exposures may be associated with changes in body composition. However, more evidence is needed to identify key pollutants linked to adverse health effects and their underlying biomolecular mechanisms, particularly in sensitive older adults. OBJECTIVES:Our research aimed to systematically assess the relationship between abiotic airborne exposures and changes in body composition among healthy older adults, as well as the potential mediating mechanisms through the serum lipidome. METHODS:From September 2018 to January 2019, we conducted a monthly survey among 76 healthy adults (60-69 years old) in the China Biomarkers of Air Pollutant Exposure (BAPE) study, measuring their personal exposures to 632 abiotic airborne pollutions using MicroPEM and the Fresh Air wristband, 18 body composition indicators from the InBody 770 device, and lipidomics from venous blood samples. We used an exposome-wide association study (ExWAS) and deletion/substitution/addition (DSA) model to unravel complex associations between exposure to contaminant mixtures and body composition, a Bayesian kernel machine regression (BKMR) model to assess the overall effect of key exposures on body composition, and mediation analysis to identify lipid intermediators. RESULTS:The ExWAS and DSA model identified that 2,4,5-T methyl ester (2,4,5-TME), 9,10-Anthracenedione (ATQ), 4b,8-dimethyl-2-isopropylphenanthrene, and 4b,5,6,7,8,8a,9,10-octahydro-(DMIP) were associated with increased body fat mass (BFM), fat mass indicators (FMI), percent body fat (PBF), and visceral fat area (VFA) in healthy older adults [Bonferroni-Hochberg false discovery rate (FDRBH)<0.05]. The BKMR model demonstrated a positive correlation between contaminants (anthracene, ATQ, copaene, di-epi-α-cedrene, and DMIP) with VFA. Mediation analysis revealed that phosphatidylcholine [PC, PC(16:1e/18:1), PC(16:2e/18:0)] and sphingolipid [SM, SM(d18:2/24:1)] mediated a significant portion, ranging from 12.27% to 26.03% (p-value <0.05), of the observed increase in VFA. DISCUSSION:Based on the evidence from multiple model results, ATQ and DMIP were statistically significantly associated with the increased VFA levels of healthy older adults, potentially regulated through lipid intermediators. These findings may have important implications for identifying potentially harmful environmental chemicals and developing targeted strategies for the control and prevention of chronic diseases in the future, particularly as the global population is rapidly aging. https://doi.org/10.1289/EHP13865.

Large language models (LLMs) are essential tools that users employ across various scenarios, so evaluating their performance and guiding users in selecting the suitable service is important. Although many benchmarks exist, they mainly focus on specific predefined model abilities, such as world knowledge, reasoning, etc. Based on these ability scores, it is hard for users to determine which LLM best suits their particular needs. To address these issues, we propose to evaluate LLMs from a user-centric perspective and design this benchmark to measure their efficacy in satisfying user needs under distinct intents. Firstly, we collect 1,846 real-world use cases from a user study with 712 participants from 23 countries. This first-hand data helps us understand actual user intents and needs in LLM interactions, forming the User Reported Scenarios (URS) dataset, which is categorized with six types of user intents. Secondly, based on this authentic dataset, we benchmark 10 LLM services with GPT-4-as-Judge. Thirdly, we show that benchmark scores align well with human preference in both real-world experience and pair-wise annotations, achieving Pearson correlations of 0.95 and 0.94, respectively. This alignment confirms that the URS dataset and our evaluation method establish an effective user-centric benchmark. The dataset, code, and process data are publicly available at https://github.com/Alice1998/URS.

Fairness of recommender systems (RS) has attracted increasing attention recently. Based on the involved stakeholders, the fairness of RS can be divided into user fairness, item fairness, and two-sided fairness which considers both user and item fairness simultaneously. However, we argue that the intersectional two-sided unfairness may still exist even if the RS is two-sided fair, which is observed and shown by empirical studies on real-world data in this paper, and has not been well-studied previously. To mitigate this problem, we propose a novel approach called Intersectional Two-sided Fairness Recommendation (ITFR). Our method utilizes a sharpness-aware loss to perceive disadvantaged groups, and then uses collaborative loss balance to develop consistent distinguishing abilities for different intersectional groups. Additionally, predicted score normalization is leveraged to align positive predicted scores to fairly treat positives in different intersectional groups. Extensive experiments and analyses on three public datasets show that our proposed approach effectively alleviates the intersectional two-sided unfairness and consistently outperforms previous state-of-the-art methods.