Beyond Latent Personalization: Integrating HRV-Conditioned Personalization via LLM for Interpretable Stress Detection

BEYOND LATENT PERSONALIZATION: INTEGRATING HRV-CONDITIONED PERSONALIZATION VIA LLM FOR INTERPRETABLE STRESS DETECTION Wei-Hsiang Ma1 , Ya-Tse Wu2 , Hsing-Hang Chou2 , Chi-Chun Lee2

arXiv:2610.23555v1 [eess.SP] 16 Oct 2026

1

Institute of Communication Engineering, National Tsing Hua University, Taiwan 2 Department of Electrical Engineering, National Tsing Hua University, Taiwan

ABSTRACT Stress detection based on heart rate variability (HRV) has received increasing attention. However, existing personalized approaches integrate personal attributes with HRV through latent representations, limiting the interpretability of their interactions. In this work, we propose an LLM-enhanced personalized stress detection framework that integrates HRV features and personal attributes through LLM-based semantic interpretation. The proposed framework achieves a 29.9% relative MCC improvement over the strongest baseline on TILES2018 while providing human-readable descriptions of individual physiological responses. The results highlight the potential of LLM-based semantic interpretation for achieving both effective and interpretable personalized stress detection. Index Terms— Personalized Stress Detection, Heart Rate Variability (HRV), large language models (LLMs), semantic interpretation 1. INTRODUCTION With increasing awareness of mental health, physiological signal-based stress detection has become an important research topic. Among various physiological indicators, heart rate variability (HRV) has received considerable attention, as numerous studies have demonstrated its association with psychological stress [1, 2]. However, HRV responses vary across individuals due to personal attributes, such as age and gender [3, 4], as well as work-related factors, such as shift schedules [5], suggesting that personal context should be considered when interpreting HRV patterns for stress detection. Recent studies have incorporated personal attributes into HRV-based stress detection by aligning HRV with survey information through contrastive learning [6], regularizing HRV embeddings using pairwise attribute similarities [7], and modeling higher-order HRV–attribute interactions through hypergraph learning [8]. Despite their effectiveness, these feature-based approaches implicitly encode HRV–attribute interactions in latent representations, providing limited insight into how relatively static personal attributes contextualize dynamically changing HRV responses for stress detection [9]. Large language models (LLMs) offer a promising alternative, having been applied across medical domains to integrate clinically informative text into downstream tasks [10]. Re-

cent efforts have translated and aligned non-linguistic, highdimensional physiological signals with LLMs for various downstream tasks [11, 12, 13]. Their ability to express contextual information provides a natural means of describing how personal attributes relate to physiological responses. Motivated by this capability and the limitations of conventional feature-based integration, we propose the BeyondLatent Personalization (BLP) framework, an LLM-enhanced stress classification framework that takes HRV features and LLM-generated domain-specific descriptions as inputs, with each description integrating personal attributes with HRV information from a specific physiological domain. As each HRV domain reflects distinct physiological characteristics, these descriptions provide complementary perspectives on an individual’s stress-related physiological responses. To reduce redundancy across their semantic representations, we impose an orthogonality constraint and adaptively aggregate them through learnable source weights, while an intra-subject contrastive objective is employed to enhance stress-discriminative representation learning. This design provides an interpretable alternative to conventional blackbox integration: the generated descriptions explicitly reveal how the LLM integrates personal context with physiological information, while the learned source weights quantify the relative contribution of each HRV domain to stress detection. On TILES-2018, BLP achieves a 29.9% relative MCC improvement over the strongest baseline, while its performance on WESAD supports its generalizability across datasets. These results demonstrate improved stress classification and transparency into how personal attributes complement physiological information, revealing higher-level relationships less accessible through conventional feature-based integration. 2. METHODOLOGY As illustrated in Fig. 1, BLP consists of four blocks: HRV representation learning, LLM-based semantic interpretation, semantic representation learning, and multimodal fusion. The learned representations from multiple daily samples of the same subject are jointly processed through the bidirectional cross-attention fusion block to capture complementary physiological and semantic information, then concatenated and classified by an MLP for stress detection.

HRV Representation Learning

HRV Embedding 𝒉𝒉

Physiological Interpretation Prompt

Projection Head

HRV Feature 𝝌𝒉

Multimodal Fusion & Classifier

HRV Representation HRV Projector

Heart Rate

Age: 28 Gender: Female Shift: Night Shift … HRV Integration

𝑝#

𝑍#

Bidirectional Cross-Attention

LLM Interpretation

Domain-Specific Description

LLM-based semantic interpretation

Projection Head

Text Projector

Source Projector

Semantic Embedding 𝓛𝒐𝒓𝒕𝒉𝒐 𝒉𝐭

𝑝$

C

𝑍$ 𝑝$

𝑓$

Semantic Representation

𝑦" !,#

Fused Multimodal Representations

𝑓#

𝑝#

𝓛𝒊𝒏𝒕𝒓𝒂 Semantic Embedding Extractor

Attribute Integration

Attribute

Transformer Encoder

Inputs Template [Participant ID] is a [Gender] participant in the…

f

Day 1 Day 2 Day 3 Day 4

Classifier (MLP)

𝑦!,#

. . . Same subject (Multiple days)

𝓛𝑪𝑬

Semantic Representation Learning

Fig. 1. Overview of the proposed BLP framework. Table 1. Physiological Interpretation Prompt Template Attribute Integration [Neutral Baseline Profile] This individual is a {gender} participant in the {age} age group, and {student desc}. Professionally, {subject} works as a(n) {position}, with {nurseyears} of professional experience and {duration} of tenure at the current organization. Working the {shift}, {subject} averages {hours} per week with approximately {overtime} of monthly overtime, and supervises {quantsup}. Regarding personal background, {subject} identifies as {race} and has {englyrs} of English language experience. {possessive} current relationship status is {relationship}, with an annual household income of {income}. {Subject} has {children} child(ren) under 18 in the household and {housing desc}. {possessive} typical daily commute is {commute time}. HRV Integration (Take time domain as example) [Data Interpretation Context] Note: The following HRV metrics are not instantaneous readings. They are strictly calculated based on rigorous daily aggregation: Feature Group: ’Time’: ”Time-domain HRV features (e.g., meanNN, sdNN, RMSSD, pNN50, cvNN)”; Base calculation: Extracted from 5-minute rolling windows of Normal-to-Normal (NN) intervals. Time-domain differences and Frequency domains were normalized; Suffixes meaning: Metric ending in ‘ mean’ is the daily average; ‘ std’ is the daily standard deviation; ‘ cv’ (Coefficient of Variation) shows daily volatility; ‘ skew’/‘ kurt’ show distribution shape; ‘ 1quartile’/‘ 3quartile’ are the 25th/75th percentiles. Please interpret these metrics as the overall autonomic stability and allostatic load for the entire day. [Current Physiological Markers] {feature 1}: {value 1} {feature 2}: {value 2}... {feature N}: {value N} General [Task] Based on the individual’s baseline profile and today’s HRV metrics, write a daily analytical log assessing their current stress state. [Output Constraints] Combine your final analysis into a SINGLE, cohesive paragraph; Focus strictly on physiological and psychological interpretation; DO NOT directly state, conclude, or predict whether the individual is “stressed” or “not stressed”; DO NOT list, quote, or repeat the raw HRV numbers provided above. Use qualitative descriptions (e.g., “highly volatile,” “elevated parasympathetic activity”) instead; If there’s ‘Unknown’ in the description, it means the data is missing and you can ignore it.

2.1. HRV Representation Learning For each subject, HRV feature vectors xh are temporally ordered to form a sequence that captures the evolution of physiological responses over time. Positional encodings are added to xh , which are then processed by a Transformer encoder to model temporal dependencies, yielding encoded HRV embeddings hh . An HRV projector then maps hh to ph for bidirectional cross-attention. 2.2. LLM-based Semantic Interpretation The semantic interpretation module transforms structured physiological and personal information into a textual representation. Following prior studies that contextualize physiological measurements with domain knowledge and user information for LLM-based health analysis [11, 12], we convert each sample’s HRV features and personal attributes into textual components and incorporate them into predefined

Physiological Interpretation Prompts. Specifically, the HRV features are organized into four physiological domains: time, frequency, multi-scale RR, and multi-scale dRR, with a separate prompt constructed for each domain. As shown in Table 1, the subject’s personal attributes are converted into the [Neutral Baseline Profile], which is shared across the four domain-specific prompts, while the HRV features and their interpretation context for each domain are organized into the [Data Interpretation Context] and [Current Physiological Markers]. Together with the predefined [Task] and [Output Constraints], these sections form the complete prompt. During LLM Interpretation, the prompt is processed by an LLM to generate a Domain-Specific Description that describes the physiological state in the context of the individual’s baseline characteristics without directly predicting the stress label. 2.3. Semantic Representation Learning The Domain-Specific Descriptions generated by Qwen3.59B [14] are encoded using Qwen3-Embedding-0.6B [15], yielding four 1024-dimensional semantic embeddings ht for each sample. Each embedding is then transformed by a domain-specific Source Projector, producing four learnable representations on which the orthogonal loss is applied. These representations are adaptively aggregated using softmax-normalized learnable weights to capture the relative contribution of each HRV domain. Finally, the aggregated representation is projected into a 100-dimensional shared embedding space, yielding the semantic representation pt for subsequent cross-modal fusion. 2.4. Training Objective To capture subject-specific stress patterns while discriminating between stress states within each subject, we employ an intra-subject Multi-Similarity Loss Lintra on the HRV and semantic representations:
Lintra = LM S [Zh ; Zt ], c , cs,d = (s, ys,d ), (1) h t where Z and Z denote the HRV and semantic representations obtained through their projection heads, respectively, and cs,d is defined by the subject identity s and stress label ys,d . To reduce redundancy among the domain-specific semantic representations, we further introduce an orthogonal loss: 2 X 1  e (p) ⊤ e (q) E , (2) Lortho = E N2 F 1≤p<q≤4

e (p) and E e (q) denote the representations from two where E HRV domains, and N denotes the number of valid samples. This constraint encourages different domain-specific representations to preserve complementary semantic information. Finally, the complete framework is optimized using Ltotal = LCE + λintra Lintra + λortho Lortho ,

Table 2. Overview of the 61 HRV Features

(3)

where LCE denotes the weighted cross-entropy loss for stress classification, while λintra and λortho control the contributions of the two auxiliary objectives. We provide our source code in a GitHub repository1 . 3. EXPERIMENT SETUP 3.1. Dataset and Feature Extraction TILES-2018 [16] is a relatively large-scale longitudinal dataset collected from 212 hospital workers over 10 weeks under naturalistic conditions, whereas WESAD [17] is a smaller controlled dataset of 15 participants under Baseline, Stress, and Amusement conditions. Together, they provide complementary settings for evaluating personalized stress detection and framework generalizability. Both datasets use the same 61 HRV features across four domains (Table 2). For TILES-2018, following the data quality screening and label construction in [7], 185 participants are retained. HRV features are extracted from 5-min RR segments derived from heart rate measurements, with a 4-min overlap, and aggregated daily using 13 statistical functionals: mean, standard deviation (SD), coefficient of variation (CV), median, minimum, maximum, 1st and 3rd quartiles, skewness, kurtosis, weighted mean, weighted SD, and weighted CV, resulting in 793 features that are reduced to 100 by Recursive Feature Elimination (RFE). WESAD uses 700-Hz ECG signals to extract HRV features from RR intervals using the recommended 60-s windows with a 0.25-s shift; windows with less than 80% RR coverage are excluded, and each valid window retains its 61-dimensional HRV vector without daily aggregation or RFE. Following [17], Baseline and Amusement are labeled non-stress and Stress as stress. The 17 TILES-2018 attributes from self-reported surveys are gender, age group, student status, current position, professional experience, organizational tenure, work shift, weekly working hours, monthly overtime, number of supervisees, race, English-language experience, relationship status, annual household income, number of children under 18, housing status, and daily commute time. The 11 WESAD attributes include age, height, weight, gender, and dominant hand; coffee consumption, sports participation, smoking status, and illness status on the day of data collection; and coffee consumption and smoking within the last hour. The HRV encoder uses one Transformer layer with two attention heads and a hidden dimension of 100, while bidirectional cross-attention uses four heads. Training uses a batch size of 16 subjects for up to 400 epochs, with λintra and λortho 1 https://github.com/Justinma0201/BLP

selected by grid search over [0.1, 1]. Evaluation uses subjectindependent five-fold cross-validation with unweighted average recall (UAR), accuracy (Acc), and MCC, with MCC as the primary metric due to class imbalance. Additional implementation details are provided in the released code. 3.2. Comparison Models We compare BLP with three baselines and six ablation models. Trans.+MTL+GRL [18] incorporates personal attributes through multi-task learning with a gradient reversal layer; Trans.+MGCN+MS [19] integrates attributes through graphbased modeling with Multi-Similarity loss; and DESHM [8] captures higher-order relationships between HRV representations and personal attributes through dynamic hypergraph learning. For the ablations, LLM-Only uses only the semantic representation with a zeroed HRV branch, whereas HRV-Only uses only the HRV representation with a zeroed LLM branch. w/o Attribute Integration and w/o HRV Integration retain both branches but remove personal attributes and HRV information from LLM interpretation, respectively. w/o Embedding Extractor retains the LLM interpretation but removes the subsequent embedding extractor, directly using the normalized last-token representation from the LLM as the semantic embedding. w/o LLM Interpretation removes LLM-based semantic interpretation and directly encodes the original prompt with the embedding extractor. 4. RESULTS ANALYSIS AND ABLATION 4.1. Performance Comparison Baselines: Table 4 compares the proposed method with three baselines under the same subject-independent setting. The baseline results are reproduced using the original implementation provided by the authors of [8]. Statistical significance is evaluated using within-fold stratified BCa bootstrap with subject-level paired resampling, with two-sided p-values derived from the 95% CIs of pooled MCC differences and corrected using the Holm–Bonferroni method. The proposed model achieves the highest UAR/Acc/MCC of 0.630/0.631/0.261, with MCC improvements of 51.7%, 33.2%, and 29.9% over Trans.+MTL+GRL, Trans.+MGCN+ MS, and DESHM, respectively, all statistically significant (p < 0.05). These results demonstrate the effectiveness of BLP for personalized stress detection compared with conventional attribute-integration approaches.

Table 3. Ablation Results on the TILES-2018 and WESAD Datasets. Interp. = Interpretation; Integ. = Integration; Attr. = Attribute; Extractor = Semantic Embedding Extractor. Methods HRV LLM-Only HRV-Only w/o Attribute Integration w/o HRV Integration w/o Embedding Extractor w/o LLM Interpretation Proposed

× √ √ √ √ √

TILES-2018 Semantic Representation

LLM Interp. √

Extractor √

HRV Integ. √

Attr. Integ. √

× √ √ √

× √ √

× √

× × √

× √

√

× √

× √ √

√ √

√

√

√

Table 4. Baseline Comparison on TILES-2018 (∗ p < 0.05).

WESAD

UAR

Acc

MCC

UAR

Acc

MCC

0.516 0.600 0.594 0.560 0.607 0.619

0.504 0.588 0.600 0.557 0.598 0.610

0.040 0.201 0.189 0.119 0.215 0.240

0.616 0.799 0.798 0.787 0.801 0.799

0.683 0.821 0.820 0.809 0.818 0.815

0.260 0.599 0.595 0.576 0.600 0.595

0.630

0.631

0.261

0.808

0.823

0.614

Table 5. Case Study of HRV-Conditioned LLM Interpretation

Method

UAR

Acc

MCC

MCC p-value

w/o HRV Integration (Pred: 0, GT: 1): ...the cumulative effect of

Trans.+MTL+GRL [18]

0.584

0.603

0.172

0.0019*

Trans.+MGCN+MS [19]

0.597

0.613

0.196

0.0085*

DESHM [8]

0.598

0.619

0.201

0.0496*

Proposed

0.630

0.631

0.261

–

prolonged overtime, leadership duties, and a lengthy commute creates a state of sustained activation where the body and mind are likely operating under continuous demand, potentially leading to physiological weariness and psychological strain... Proposed (Pred: 1, GT: 1): ...This pattern is further underscored by the significant variability in very-low-frequency power, which reflects a disrupted circadian rhythm or allostatic load, likely exacerbated by the cumulative demands of his supervisory role, extended overtime hours, and long commute...

Ablation Study: Table 3 shows that, on TILES-2018, removing HRV or personal attributes from the LLM prompt reduces MCC by 54.4% and 27.6%, respectively. The former shows that HRV provides day-to-day physiological variations and distinct domain information, whose complementarity is preserved by the orthogonal constraint; the latter shows that personal attributes provide subject-specific baselines to contextualize such dynamic information. Together, they enable more complete personalized interpretations. LLMOnly substantially underperforms HRV-Only, indicating that semantics complement rather than replace physiological representations. Removing LLM interpretation reduces MCC by 8.0%/3.1% (TILES-2018/WESAD), showing that, although the structured prompt preserves HRV and attribute information, LLM interpretation further establishes semantic relationships between them. Removing the semantic embedding extractor causes larger reductions of 17.6%/2.3%, demonstrating its role in extracting task-relevant representations from rich semantics. Its greater impact on TILES-2018 may reflect the more complex daily physiological and personal information in its longitudinal naturalistic setting. 4.2. Case Study Table 5 presents cases where BLP corrects prediction errors from the w/o HRV Integration setting. Incorporating samplespecific HRV information allows physiological patterns to be interpreted alongside personal attributes. In the stress case, VLF variability is related to occupational demands, whereas in the non-stress case, elevated pNN50 and RMSSD are associated with parasympathetic activity and recovery capacity, with key differences highlighted in red. These cases illustrate how BLP explicitly relates physiological patterns to personal attributes to form individualized, physiologically meaningful

w/o HRV Integration (Pred: 1, GT: 0): ...With over 15 years of experience in clinical support and a tenure of 6 to 9 years at the current organization, she possesses significant institutional knowledge and likely established coping mechanisms ... Furthermore, her stable marital status and the presence of a child under 18 indicate strong social anchoring... Proposed (Pred: 0, GT: 0): ...significantly elevated parasympathetic activity, as evidenced by the exceptionally high pNN50 and RMSSD values, which suggest a strong capacity for recovery and stress buffering... Given her extensive professional tenure and established home environment, this pattern aligns with a resilient allostatic load profile where the body effectively manages daily demands without exhibiting signs of chronic depletion or dysregulation...

representations. This explicit relationship further supports the interpretability of personalized stress detection. 5. CONCLUSION AND FUTURE WORK In this work, we proposed BLP, which integrates HRV and personal attributes through LLM-based semantic interpretation to generate personalized, human-readable descriptions fused with temporal HRV representations for stress classification. BLP outperforms all compared baselines on TILES-2018, while its performance on WESAD supports its generalizability across datasets. Ablation results verify the complementary contributions of HRV features, personal attributes, semantic interpretation, and dedicated embedding extraction. However, the framework remains reliant on HRV representations, as LLM-generated semantic representations alone are insufficiently discriminative. Future work will enhance the task-relevant information in LLM-generated descriptions while preserving semantic interpretability, aiming to better balance interpretability and predictive performance.

6. ACKNOWLEDGMENT We use OpenAI ChatGPT only to polish manuscript grammar and refine the final delivered code. All experimental results and analyses are conducted without generative AI. 7. REFERENCES [1] Sarah Immanuel, Meseret N Teferra, Mathias Baumert, and Niranjan Bidargaddi, “Heart rate variability for evaluating psychological stress changes in healthy adults: a scoping review,” Neuropsychobiology, vol. 82, no. 4, pp. 187–202, 2023. [2] Beatrice Thielmann, Jonas Hartung, and Irina Böckelmann, “Objective assessment of mental stress in individuals with different levels of effort reward imbalance or overcommitment using heart rate variability: A systematic review,” Systematic reviews, vol. 11, no. 1, pp. 48, 2022. [3] Mariaconsuelo Valentini and Gianfranco Parati, “Variables influencing heart rate,” Progress in cardiovascular diseases, vol. 52, no. 1, pp. 11–19, 2009. [4] Julian Koenig and Julian F Thayer, “Sex differences in healthy human heart rate variability: A meta-analysis,” Neuroscience & Biobehavioral Reviews, vol. 64, pp. 288–310, 2016. [5] Taihe Zhan, Ziying Zhang, Zhimin Shi, Hongyan Xie, Daogang Zha, and Xiumei Wei, “Factors influencing heart rate variability in nurses following night shifts: a prospective observational clinical study,” BMC nursing, vol. 24, no. 1, pp. 1318, 2025. [6] Zeyu Yang, Han Yu, and Akane Sano, “Contrastive pretraining for stress detection with multimodal wearable sensor data and surveys,” Proceedings of machine learning research, vol. 287, pp. 166, 2025. [7] Jing-Chun Wang, Woan-Shiuan Chien, Huan-Yu Chen, and Chi-Chun Lee, “In-the-wild hrv-based stress detection using individual-aware metric learning,” in 2024 IEEE 20th International Conference on Body Sensor Networks (BSN). IEEE, 2024, pp. 1–4. [8] Jing-Chun Wang, Woan-Shiuan Chien, and Chi-Chun Lee, “A dynamic edge-selection mechanism in hrv hypergraph learning for improved stress detection,” in ICASSP 2025-2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2025, pp. 1–5. [9] Marco Bolpagni, Susanna Pardini, Marco Dianti, and Silvia Gabrielli, “Personalized stress detection using biosignals from wearables: A scoping review,” Sensors, vol. 24, no. 10, pp. 3221, 2024. [10] Huan-Yu Chen, Woan-Shiuan Chien, Ching-Heng Lin, and Chi-Chun Lee, “Leveraging foundation models for

clinically instructed tumor image synthesis in renal cell carcinoma,” in 2025 IEEE 22nd International Symposium on Biomedical Imaging (ISBI). IEEE, 2025, pp. 1– 5. [11] Zengding Liu, Chen Chen, Jiannong Cao, Minglei Pan, Jikui Liu, Nan Li, Fen Miao, and Ye Li, “Large language models for cuffless blood pressure measurement from wearable biosignals,” in Proceedings of the 15th ACM international conference on bioinformatics, computational biology and health informatics, 2024, pp. 1– 11. [12] Yubin Kim, Xuhai Xu, Daniel McDuff, Cynthia Breazeal, and Hae Won Park, “Health-llm: Large language models for health prediction via wearable sensor data,” arXiv preprint arXiv:2401.06866, 2024. [13] Kai Yang, Massimo Hong, Jiahuan Zhang, Yizhen Luo, Suyuan Zhao, Ou Zhang, Xiaomao Yu, Jiawen Zhou, Liuqing Yang, Ping Zhang, et al., “Ecg-lm: understanding electrocardiogram with a large language model,” Health Data Science, vol. 5, pp. 0221, 2025. [14] Qwen Team, “Qwen3.5: Towards native multimodal agents,” February 2026. [15] Yanzhao Zhang, Mingxin Li, Dingkun Long, Xin Zhang, Huan Lin, Baosong Yang, Pengjun Xie, An Yang, Dayiheng Liu, Junyang Lin, Fei Huang, and Jingren Zhou, “Qwen3 embedding: Advancing text embedding and reranking through foundation models,” arXiv preprint arXiv:2506.05176, 2025. [16] Karel Mundnich, Brandon M Booth, Michelle l’Hommedieu, Tiantian Feng, Benjamin Girault, Justin L’hommedieu, Mackenzie Wildman, Sophia Skaaden, Amrutha Nadarajan, Jennifer L Villatte, et al., “Tiles-2018, a longitudinal physiologic and behavioral data set of hospital workers,” Scientific Data, vol. 7, no. 1, pp. 354, 2020. [17] Philip Schmidt, Attila Reiss, Robert Duerichen, Claus Marberger, and Kristof Van Laerhoven, “Introducing wesad, a multimodal dataset for wearable stress and affect detection,” in Proceedings of the 20th ACM international conference on multimodal interaction, 2018, pp. 400–408. [18] Shakeel A Sheikh, Md Sahidullah, Fabrice Hirsch, and Slim Ouni, “Robust stuttering detection via multi-task and adversarial learning,” in 2022 30th European signal processing conference (EUSIPCO). IEEE, 2022, pp. 190–194. [19] Zihao Li, Yakun Chen, Xianzhi Wang, Lina Yao, and Guandong Xu, “Multi-view gcn for loan default risk prediction,” Neural Computing and Applications, vol. 36, no. 20, pp. 12149–12162, July 2024.

Related documents

• Local LLM-based sentiment analysis of emotional patterns in complex-to-manage and treatment-resistant psoriatic arthritis.#988370
• RIACT: A Responsible AI System for Personalized Study Habit Tracking and Early Burnout Signal Detection in University Students#361080
• Electrodermal Activity as a Unimodal Signal for Aerobic Exercise Detection in Wearable Sensors#1110202
• Leveraging AI for Automatic Item Generation for Psychological Scales#215932
• Leveraging AI for Automatic Item Generation for Psychological Scales#16321