arXiv subjects

Ciro Beneduce

Publications and source records attributed to Ciro Beneduce.

7 recordsLinked to original sources

arXiv2026

Mirror, Mirror on the Wall: Can VLM Agents Tell Who They Are at All?

In the animal kingdom, mirror self-recognition is a canonical probe of higher-order cognition, emerging only in some species. We ask whether an analogous functional capability emerges in embodied vision-language model (VLM) agents: can they recognize themselves in a mirror? We introduce a controlled 3D benchmark where a first-person VLM agent must infer a hidden body attribute from its reflection and select the matching target, while avoiding self-other misattribution. To separate mirror-grounded self-identification from shortcuts, we test mirror removal, misleading cues, and occluded reflections. We also evaluate the decision process through mirror seeking, temporal ordering, self-attribution, and reasoning-action consistency. Our experiments show that mirror-based self-identification emerges mainly in stronger VLMs. These models can use reflected evidence for action, whereas weaker models often inspect the mirror but fail to extract self-relevant information or misattribute their reflection. Language-vision conflict further shows that self-referential language alone is not evidence of grounded self-identification. Overall, mirror-based evaluation provides a diagnostic for whether embodied self-grounding is causally rooted in perception and action rather than priors, prompt compliance, or confabulation.

cs.AI↗

arXiv2025

TS-Mob: Social and Geographical-Aware Time Series Foundation-Model Framework for Human Mobility Prediction

Short-term forecasting of aggregated human mobility flows supports urban planning, intelligent transportation systems, and emergency response, yet existing models often require substantial mobility history and learn spatial structure implicitly through grids or graphs. Time series foundation models provide strong temporal priors but typically lack explicit geographic and social conditioning for origin-destination interactions. We introduce TS-Mob, a framework that conditions a fine-tuned time series foundation model (TimesFM) forecaster on a gravity-inspired destination-attractiveness index that encodes geographic and social signals computed from open data (living population, centroid distances, and Overture POI counts), together with weather covariates. Evaluated on commonly used benchmarks like Bike New York City, Taxi Beijing, and a nation-scale Spain origin-destination matrix estimated through mobile phone data, TS-Mob outperforms classical, deep spatio-temporal, and a set of foundation model-based baselines across RMSE, MAE, and CPC, with gains up to 78.71% lower MAE and 137.93% higher CPC over the best classical baseline, and up to about 4.27% lower RMSE over the strongest foundation baseline. Stratified analyses further show robustness across different temporal regimes, like weekdays/weekends.

cs.LG↗

arXiv2025

AI's Blind Spots: Geographic Knowledge and Diversity Deficit in Generated Urban Scenario

Diffusion-based text-to-image models are increasingly used for urban analysis and scenario generation, but their geographic knowledge and representational biases remain poorly understood. We evaluate FLUX 1-schnell and Stable Diffusion 3.5-Large in the United States by generating 150 street-view images for each state, each state capital, and a generic "USA" prompt. Images are embedded with DINO-v2 ViT-S/14 and compared with Fr\'echet Inception Distance (FID). Pairwise FID clustering shows that geographically proximate states and capitals often group together, indicating implicit geographic structure. However, the generic ``USA'' prompt collapses this diversity into a metropolitan stereotype: frontier, desert, tropical, rural, and small-city environments are underrepresented or distant in FID space. These results show that diffusion models can encode fine-grained geography while still reproducing narrow national-scale visual stereotypes.

cs.AI↗

arXiv2025

pySpainMobility: Unlocking Spanish Open Mobility Data for Spatial Inequality Research

Human mobility shapes access to resources, opportunities, and services, making movement data a powerful lens for studying spatial and social inequality. Yet despite the growing availability of official open mobility datasets, their research potential is rarely realized because the technical overhead of retrieving, harmonizing, and processing them often crowds out substantive analysis. To address this, we introduce pySpainMobility, a Python package that automates the retrieval and harmonization of Spain's open mobility data across spatial resolutions and demographic strata, streamlining national-scale, reproducible analysis. Using the package, we study income-stratified mobility inequality across Spain's inter-province network, drawing on district-level origin-destination flows for four representative weeks spanning the seasons of 2023. We construct income-specific mobility layers and show that socioeconomic stratification is deeply embedded in the structure of the national mobility system: low-income mobility is disproportionately concentrated in a narrow set of destinations and shorter in spatial reach, while high-income groups access a broader and more distant hierarchy of destinations. Low- and high-income layers consistently follow weakly aligned destination hierarchies across seasons, indicating that income groups navigate distinct mobility geographies rather than a shared one at different volumes. We further show that destination provinces themselves differ systematically in the income composition of the travelers they receive, with several provinces attracting arrivals disproportionately skewed toward one income group relative to the national seasonal baseline. These results demonstrate how official open mobility data, combined with accessible tooling, can be operationalized to reveal spatial inequality as a structural property of national mobility networks.

cs.CY↗

arXiv2025

The LLM Wears Prada: Analysing Gender Bias and Stereotypes through Online Shopping Data

With the wide and cross-domain adoption of Large Language Models, it becomes crucial to assess to which extent the statistical correlations in training data, which underlie their impressive performance, hide subtle and potentially troubling biases. Gender bias in LLMs has been widely investigated from the perspectives of works, hobbies, and emotions typically associated with a specific gender. In this study, we introduce a novel perspective. We investigate whether LLMs can predict an individual's gender based solely on online shopping histories and whether these predictions are influenced by gender biases and stereotypes. Using a dataset of historical online purchases from users in the United States, we evaluate the ability of six LLMs to classify gender and we then analyze their reasoning and products-gender co-occurrences. Results indicate that while models can infer gender with moderate accuracy, their decisions are often rooted in stereotypical associations between product categories and gender. Furthermore, explicit instructions to avoid bias reduce the certainty of model predictions, but do not eliminate stereotypical patterns. Our findings highlight the persistent nature of gender biases in LLMs and emphasize the need for robust bias-mitigation strategies.

cs.AI↗

arXiv2025

Multimodal Large Language Models Predict Urban Safety Perception but Encode Non-Neutral Demographic Priors

Understanding how people perceive urban environments is essential for inclusive planning, yet conventional surveys are costly and difficult to scale. We investigate whether Multimodal Large Language Models (MLLMs) can assess perceived urban safety from street-view imagery while accounting for the observer-dependent nature of perception. Using Place Pulse 2.0, we evaluate four open and proprietary MLLMs across 56 cities under a Neutral prompt and socio-demographic personas defined by gender, age, and race or ethnicity. We also analyse the keywords generated to justify each classification. All four models display comparable zero-shot capability, with city-macro F1 scores of 65--69%, and preserve meaningful cross-city variation. However, they systematically favour the Safe class, underpredict unsafety, and compress differences between cities. Their explanations converge on a shared visual lexicon: maintenance, greenery, order, and residential character support Safe judgements, whereas deterioration, isolation, poor lighting, and limited pedestrian activity support Unsafe judgements. Persona prompting produces substantial and structured shifts while holding the image fixed. Female personas yield more Unsafe classifications than Male personas across all models; age effects are model-dependent, although Middle-aged personas generally remain closest to Neutral. Black/African American and Native American personas frequently show the largest departures, while the closest race or ethnicity match varies by model. These findings show that MLLMs can provide scalable signals of perceived urban safety, but not from a demographically neutral standpoint.

cs.CY↗

arXiv2024

Large Language Models are Zero-Shot Next Location Predictors

Predicting the locations an individual will visit in the future is crucial for solving many societal issues like disease diffusion and reduction of pollution. However, next-location predictors require a significant amount of individual-level information that may be scarce or unavailable in some scenarios (e.g., cold-start). Large Language Models (LLMs) have shown good generalization and reasoning capabilities and are rich in geographical knowledge, allowing us to believe that these models can act as zero-shot next-location predictors. We tested more than 15 LLMs on three real-world mobility datasets and we found that LLMs can obtain accuracies up to 36.2%, a significant relative improvement of almost 640% when compared to other models specifically designed for human mobility. We also test for data contamination and explored the possibility of using LLMs as text-based explainers for next-location prediction, showing that, regardless of the model size, LLMs can explain their decision.

cs.CY↗