Papers

Public understanding of how people use LLM-based conversational AI assistants comes primarily from aggregate platform reports by OpenAI and Anthropic, which apply fixed taxonomies and inferred demographics to hundreds of millions of users and release only summary statistics that outside researchers cannot re-analyze. We provide a complementary, conversation-level view: complete ChatGPT exports comprising 202,590 conversations from 1,252 users across India, Nigeria, Brazil, and Pakistan, paired with self-reported age and gender and spanning December 2022 to February 2026. To our knowledge this is the first conversation-level, demographically grounded comparison of ChatGPT use across multiple non-Western countries. We ask what these users use ChatGPT for (purpose), what they talk about (topics), and how they engage with it (mode of interaction), using the platform's own classifiers, unsupervised topic discovery, and a thematic analysis of expressive conversations. Personal use accounts for 55-64% of conversations in every country and coursework is about as common as work, so workplace productivity describes a minority of use. Unsupervised topic discovery surfaces country-specific uses that the OpenAI taxonomy folds into generic categories: health and wellness in India and Brazil, Urdu-English translation in Pakistan, current affairs in Nigeria, religious questions in Nigeria and Pakistan, and self-reflection in Brazil. Over three years, the share of conversations that seek information declined only modestly and the share that delegate a task did not grow, while conversations in which users express themselves rose from a few percent to roughly a fifth or more in every country. The same product is thus attached to different local needs in each country, and understanding what adoption means requires conversation-level, country-sensitive measurement alongside global aggregates.

Emergence, features of a whole that no part shows alone, is central to complex systems but rarely measured so as to locate its source. Viewing a system structurally, as a network of interacting parts, we evaluate emergence as what observing the coupled whole reveals beyond, or erases from, the combined observations of the parts. For observations that build what they see route by route, we show that every emergent feature is produced by its own route across the interface between parts, a route that traces it to the components and interactions responsible. We show the number of all such boundary-crossing paths bounds the total emergence discrepancy, and equals the emergence of measures that count routes. The bound is computed locally around the interface, validated through network examples, and in a nervous system the routes show that interneurons relay most of the emergence. Our work associates emergence with path structure in the network system that can be anticipated, attributed, and engineered.

Narrative extraction allows us to identify online hate narratives, supporting the construction of rigorous detection systems. Existing computational approaches, however, are limited in precision as they rely on semantic representations, which tend to capture only surface-level meaning. To detect more precise and interpretable narratives, we present an extraction pipeline that represents narratives as entity-evaluation pairs. Narratives are extracted using a Large Language Model (LLM) reasoning process that extends Aspect-Based Sentiment Analysis, identifying the aspect, classifying its judgement type as the basis for evaluation, and deriving the evaluation accordingly. Extracted narratives are then clustered using Leiden, following which clusters are resolved to an intended level of granularity through an LLM-guided refinement process. We illustrate this narrative pipeline with English Reddit comments from 2024 that criticize Taylor Swift, analyzing a representative cluster that exhibits hate speech patterns to demonstrate its interpretive value.

Modularity-based community detection in multiplex networks is commonly approached through one of three strategies: early fusion (EF), which first aggregates the layers into a single network and then applies community detection; simultaneous fusion (SF), which combines information from the layers during modularity optimization; and late fusion (LF), which combines the community assignments obtained by detecting communities separately in each layer. Although recent taxonomies and surveys distinguish these strategies conceptually, their objective functions have rarely been compared analytically. We provide such a comparison for shared-node multiplex networks with normalized non-negative edge and layer weights, a common resolution parameter, and no interlayer edges. Node-attributed networks are represented by topology and attribute-similarity layers. We use a common notation for the three formulations and derive explicit relationships between their objective functions. In particular, we show that the EF objective equals the SF objective plus a non-negative heterogeneity term. We prove that the SF objective admits an optimal layer-weight vector at a simplex vertex, although nonvertex optima may also occur under ties, whereas the EF objective is concave with respect to the layer parameter for each fixed partition, and its joint optimum may occur in the interior of the simplex. For LF methods, the resulting objective and conclusions depend heavily on how the transient graph is constructed. In this study, we focus only on one of such LF methods. We complement the analytical results with brute-force experiments on synthetic multiplexes and heuristic experiments on real-world networks, including node-attributed networks represented as multiplexes. The experiments illustrate the analytical results and the influence of heuristic optimizers on the observed comparisons.

Achieving strong performance with graph neural networks (GNNs) typically requires training and hyperparameter tuning for each dataset, incurring repeated costs and effort. Graph in-context learning (ICL) avoids this by using a single pretrained model to predict unknown node labels directly from labeled context nodes. Existing approaches, however, rely on dense attention across nodes, making inference increasingly expensive as graphs grow. In this work, we present Ephris, a new graph in-context learner built on sparse message passing, scaling linearly with the number of node-feature entries and graph edges. Ephris is pretrained entirely on synthetic graphs generated from structural causal models with diverse graph structures and relational dynamics, exposing the model to varied dependencies among topology, features, and labels. We evaluate Ephris on 51 node-classification datasets against 15 extensively tuned GNNs and existing graph ICL methods under both high- and low-label train/validation/test splits. Across both settings, Ephris ranks first on all four aggregate measures: Elo, improvability, average rank, and accuracy. Its inference cost remains comparable to training a single GNN once, while being over 10 times faster than previous graph ICL models. Together, these results advance the performance-runtime Pareto frontier, demonstrating that strong graph ICL does not require dense attention. Code and model weights are available at https://github.com/nums-ai/ephris.

Modularity landscapes are often flat near the top: many distinct partitions achieve indistinguishable scores, and repeated runs of heuristic algorithms can still miss accessible optima. We introduce a two-phase workflow that (i) samples partitions until novelty saturates, then (ii) localises instability to a small subset of nodes and enumerates only that residual ambiguity under a locked stable core. Across a simple 50-nodes benchmark network, and a real weighted collaboration network, constrained enumeration systematically expands plateau coverage and can improve modularity beyond extensive Louvain restarts. Finally, we show that edge-weight rounding qualitatively reshapes plateau structure, making precision sensitivity an essential part of solution-space reporting.

Community-detection methods search over possible partitions of a network, but the structure of this partition space is rarely examined directly. This paper introduces two complementary tools that make~$\mathcal{P}$ analytically and visually accessible. First, a canonical labelling scheme based on the Restricted Growth Sequence~(RGS) is adopted, assigning each partition a unique, permutation-invariant identifier and eliminating the need for pairwise similarity measures such as the Normalized Mutual Information. Second, granularity~$\Gamma$ is defined and combined with modularity~$Q$ to produce a two-dimensional projection: the Partition Space Map~(PSM). Within the $(\Gamma,Q)$ plane, only a discrete set of integer values of~$\Gamma$ is attainable, and upper and lower bounds on~$Q$ are derived directly from the deviation matrix. These tools are demonstrated through complete enumeration of~$\mathcal{P}$ for small benchmark graphs, revealing the fine structure of modularity degeneracy and the combinatorial constraints that shape the feasible region. The framework provides an exact, assumption-free reference for evaluating heuristic community detection algorithms and lays the groundwork for geometry-aware exploration of partition space in larger networks.

Japan's 2020 tourism shock coincided with 27% fewer buyers and 35% lower gross merchandise value at one second-hand luxury resale intermediary. We compared 2019 and 2020 co-purchase projections of recorded brand labels with a Holm-corrected battery of five label-invariant metrics and Portrait Divergence against two permutation nulls rebuilding both graphs. Neither test separated the full years (smallest Holm p 0.407; 0.097 under the design-matched paired-swap null). For one random two-block split of brands, the battery detected a planted redirection of 30% of multi-brand baskets (178 of 10,399 second-group baskets, 1.7%) in 83 of 100 simulations (83-92 of 100 across 11 splits, all at least 80); Portrait Divergence reached 80% power only at full redirection. Observed falls in mean degree and clustering resembled those of 20-30% redirection, but the planted rise in degree assortativity was absent. A composition-only model of the observed attrition would be detected in only 15 of 100 simulations. Results moved with construction: without residual and material labels (post hoc), mean degree separated the years under the paired-swap null (Holm p = 0.030), and Portrait Divergence fell below 0.05, unadjusted, at the widest construction and within April-December. Shock-period network comparisons need rebuilt-graph nulls and a per-metric detection floor.

Social media platforms increasingly rely on influential users for information dissemination in domains such as marketing and political campaigns. As the value of being recognized as influential grows, users may have incentives to strategically enhance their influence. In this study, we investigate whether and to what extent the influence ranking of a target node can be improved by adding a limited number of outgoing links in the context of influence maximization (IM). We formulate the Link Selection Problem for Influence Ranking Improvement (LSP-IRI) as the problem of selecting a fixed-size set of additional outgoing links from a target node in order to maximize its rank improvement. To examine this problem, we consider two representative heuristic strategies: a greedy method that directly optimizes rank improvement and a computationally efficient random search method. We conduct experiments on four real-world networks ranging from thousands to hundreds of thousands of nodes. The results show that even a few added links can substantially improve IM-based influence rankings. In particular, the greedy method improves the ranks of nodes initially ranked around 50 by several tens of positions on average with only three added links, sometimes moving them into the top 10. The random search method achieves smaller gains under strict budgets but reduces computation time by up to approximately 98% compared with the greedy method and becomes effective when larger budgets are allowed. These findings show that IM-based influence rankings are sensitive to limited local structural modifications and highlight a trade-off between ranking improvement and computational efficiency.

Malicious coordination has long been regarded as a principal source of information ecosystem pollution. Here, we focus on crude, text-repetition-based coordination. As the demand for mitigating its dissemination has grown, scholars have studied such coordination, focusing especially on bot detection. Few studies, however, have characterized malicious coordination per se or examined how it elicits reactions from general users. Leveraging a dataset of 734,173 Chinese-language coordinated accounts and around 495 million coordinated posts published between May 2024 and March 2026, this study analyzes the characteristics of coordinated behavior and how general users react to coordinated posts. We report three findings: (1) most coordinated accounts are crude and retain the classic marks of automation, and the same criterion applied to Japanese-language accounts over the same month yields a share six times lower; (2) their content is overwhelmingly non-political; (3) regarding their reach, most reactions within large observable cascades originate from coordinated accounts themselves, while posts classified as potentially harmful or illegal material receive a comparatively high proportion of reactions from outside the Chinese-dominant population. We provide a longitudinal quantitative map of crude Chinese-language coordination appearing in X's Japanese-classified stream.

We investigate the extent to which the Hamas attacks on Israel of October 7, 2023, have affected German social media debates about Judaism and Israel. For this, we develop an approach to detect 26 anti-Semitic categories in user postings via large language models (LLMs). The approach is applied to Facebook and Telegram posts (N=125,718) from three months before and after the event. Methodically, we test different open-weight models in two setups—with and without user information as additional context to the post text. The best setup achieves up to 83 % F1-score for binary anti-Semitism detection on our manually coded validation set. User context provides valuable information for most LLMs and drastically reduces false positives, for example, when (critically) reporting on anti-Semitic incidents. Concerning our topic, we find that anti-Semitism is surging significantly on both platforms, while being about ten times more prevalent on Telegram compared to Facebook. Facebook users express anti-Semitic views most likely in posts about an alleged genocide in Gaza carried out by the Israeli army, whereas classic anti-Semitic stereotypes related to power and conspiracy theories are dominant on Telegram. After the attack, the discourse patterns on both platforms show signs of convergence, as classic anti-Semitism increases on Facebook, whereas Israel-related categories surge on Telegram.

Generative agent-based models (GABMs) are increasingly used to simulate social media dynamics, including misinformation spread. For such social simulations to be valid proxies of human behavior, LLM agents should replicate established human cognitive biases, among them the Illusory Truth Effect (ITE), where repeated exposure to a claim increases its perceived truth value. We investigate whether and how the ITE manifests across four LLMs (Gemma-3-4b-it, Qwen2.5-7B-Instruct, Llama-3.1-8B-Instruct, and GPT-5-nano) in a social media simulation context. We propose a two-phase within-context experimental design that embeds the repetition manipulation inside a realistic news feed interaction. Using this design, we collect 336,000 truth, importance, sentiment, and interest ratings across 100 statements, 10 feed variants, and 3 replications. The key comparison is between ratings assigned to repeated statements, seen throughout a simulation phase, and completely unseen ones, rated within the same experimental context window. We distinguish genuine ITE (truth-specific repetition boost) from mere exposure effects. We run an OLS regression followed by a Linear Mixed-Effect Model to account for differences across models and ratings. Our results reveal four qualitatively distinct patterns: Gemma-3 exhibits a genuine ITE; Qwen2.5 shows a mere exposure effect; GPT-5-nano displays no repetition effect on truth and mild skepticism toward repeated content; Llama-3.1 shows a small truth boost alongside decreases in evaluative dimensions. Crucially, temperature has no effect on these findings, and a variance decomposition highlights the high context-sensitivity of LLM rating behavior. Our findings caution against assuming uniform ITE replication across LLMs in social simulations, while suggesting that Gemma-3-4b-it may offer the most behaviorally realistic approximation for misinformation-related simulations.

Social media increasingly shapes everyday life, serving as both a central venue for discussion of major events and a space where online collective behavior can spill over into real-world activity, while artificial intelligence (AI) is likewise becoming increasingly influential across society. Understanding how different online communities respond to AI-related events, and disentangling the mechanisms underlying the development of such engagement, are therefore increasingly important for studying information spreading and the societal reception of AI. Our work addresses the limited connection between empirical studies of AI-related online engagement and mathematical modeling of group-level spreading dynamics. We develop a framework to collect and organize empirical Bluesky activity into distinct user groups, then use a network-based dynamics model to investigate the mechanisms underlying their engagement. We use the release of DeepSeek R1 as a case study under two complementary grouping schemes: AI-related communities and academic disciplines. For each group, we fit a network-based Susceptible-Infected-Recovered (SIR)-type model augmented with an exogenous engagement term, allowing us to quantify the relative strengths of endogenous network-driven spreading and direct external response. Across groups, we find that strong direct responses to the event do not necessarily coincide with strong network-driven propagation, revealing distinct engagement patterns that may be obscured by aggregate activity alone.

Concerns about polarization, antisocial behavior, and mental health have broadly led to two responses to social media: adapting platforms through moderation and prosocial design, and restricting access to these platforms through age limits and comparable measures. Both address consequences of social media while leaving the attention-driven architecture intact. We argue that the harms commonly attributed to social media arise not from technologies supporting social connection but from their implementation within the attention economy. In response, we ask what a digital social environment built for genuine human connection would look like. From four principles - Purpose, Alignment, Transparency, and Access - we can derive four technical properties that fulfill these principles: Operator Blindness, Algorithmic Sovereignty, Operational Parity, and Verifiability. To the best of our knowledge, no deployed alternative satisfies all four. We propose an architecture that does, making the attention economy not just discouraged but structurally impossible. We explain how the resulting user experience refocuses on each user's individual social environment and differs from that of current platforms.

Collective attention is a finite shared resource, and social media posts compete for limited opportunities to be seen. On X, users can undo a repost and repost it again. By repeating this cycle, a user can put the same post back into followers' timelines any number of times without making new content. We call this procedure repeat reposting, and we read it as placing repeated demand on this shared resource (Attentional DoS). We formalize this idea and explore repeat reposting in a large-scale dataset of cascades with at least 1,000 reactions, originating from posts classified as Japanese on X, covering April 2025 to March 2026. Repeat reposts are rare, appearing in a small share of all (user, post) pairs. Even so, close to a million posts have at least one repeater, and most repeats come from a small group of habitual accounts. The central result is that subsequent audience growth is associated less with the number of repeats than with the estimated reach of the repeating accounts. Through the lens of seriality, how habitually the same groups of accounts repeat reposts across many posts, a distinct distributed form emerges: several serial amplifiers converge on the same post (Attentional DDoS). Its synchrony, how closely their actions are timed together, forms a continuum from bursts on the scale of minutes to a daily clock. A check of the content of 99.7% of amplified posts shows that most ADDoS repeat events are directed at Chinese-script content, which our vocabulary matching and sample inspection indicate is predominantly commercial spam. Repeat reposting thus lets a post re-enter the competition for visibility. The observed pattern is better characterized as repeated temporal coverage of an existing audience and the self-reinforcement of posts that are already growing, rather than as evidence that repeat reposting takes reach from other content.

Zero-knowledge proofs (ZKPs) enable the verification of a statement without revealing any information beyond its validity and constitute a fundamental primitive in cryptography and information theory. However, existing constructions rely on computational assumptions and are predominantly confined to bipartite settings, leaving their information-theoretic realization in bipartite or network scenarios largely unexplored. Here we develop a framework for zero-knowledge verification based on the indistinguishability of quantum states under operational constraints. Exploiting the fundamental limitations imposed by local operations, we show that a verifier is inherently restricted from extracting information about the underlying state while retaining the ability to verify correctness. We construct explicit protocols for multiparty quantum networks that achieve information-theoretic security, ensuring that no subset of collaborating parties can gain knowledge beyond the validity of the statement, independent of their joint computational power.

Many real systems can be represented as growing networks where new nodes and links gradually emerge. The Barabási-Albert model for growing networks, and many models inspired by it, are based on the idea that nodes compete for links. However, the strength and the very presence of this competition have not been tested. We propose a robust statistical approach to quantify how strongly nodes compete for links, and apply it to data from various real systems—commenting on online news and cinema attendance data. We find a range of possible behaviors, from the perfectly elastic case, where new entrants shape network growth in a way that leaves the rest of the system unaffected, to an intermediate case where new entrants measurably affect the rest. Perfect competition is never observed. These findings have direct implications for complex systems modeling and e-commerce applications.

The impact of online platforms in political mobilization is widely documented, yet it remains unclear how it affects different demographic groups. To shed light on this phenomenon, we examine two contrasting cases: engagement with environmental causes on one side, and with conspiracy theories on the other. Both are well represented on Reddit, allowing us to study the different origins of these groups using a network-based method. By inferring age, gender, affluence, and partisanship of Reddit users, we construct stratified transition networks to compare the entry pathways of different demographic groups toward these communities. We find that demographic differentiation operates differently in these two domains. Entry into conspiracy communities is primarily structured by gender and partisanship: masculine users disproportionately arrive through overtly political and alt-right spaces and share right-wing media sources, whereas feminine users more often pass through esoteric and spiritual subreddits before converging on r/conspiracy. In contrast, environmental communities are differentiated mainly by age and affluence. Affluent users tend to approach environmental subreddits via discussions of individual energy management, technology, and finance, while less affluent users arrive through protest-oriented and climate movement spaces. By comparing these two issue domains, the study shows that sociodemographic sorting does not operate uniformly across polarized topics. Instead, distinct demographic groups build their own paths, shaping both the routes through which they engage with political content and the informational ecosystems that sustain their engagement. Our approach, based on the attention-flow graph, refines our understanding of how different groups become embedded in online political communities.

In large language model-based recommender systems, identity cues embedded in prompts can steer recommendations toward group-level patterns even when the underlying behavioral evidence remains unchanged. We introduce PromptShift, an interpretable, training-free framework for quantifying and mitigating such identity-cue preference drift. We define Drift as the divergence, in both item membership and ranking order, between a recommendation list generated under an identity-cued prompt and the reference list produced from the same user's interaction history alone. SliceShift then measures the extent to which a cued list gravitates, relative to the history-only reference, toward items that are more popular within the cued slice than among the global user population. Beyond conventional accuracy, we propose DifHitRate, a difficulty-weighted hit metric that credits only relevant items, assigning higher credit to hits that are less popular within the cued slice and ranked higher in the list. All components are supported by an identity-slice-by-item table constructed from positive interactions, which further enables an adaptive post-hoc reranking strategy: the reranker interpolates between the original LLM ranking and inverse slice-popularity, with personalized interpolation weight. Experiments on two datasets with three LLMs show that identity-cued prompts incur higher mean Drift than identity-free paraphrase controls, an effect beyond generic wording sensitivity, and that SliceShift is positive across all six dataset-model settings. PromptShift consistently reduces both Drift and SliceShift, lowering macro-mean SliceShift by 62.42%, while improving DifHitRate, HitRate and MRR. These results demonstrate that identity-cue preference drift can be measured and mitigated without any model training, albeit with a modest, metric-dependent utility cost.

Information cascades on social media are conventionally conceptualized as directed, feedforward branching processes. However, real-world diffusion pathways frequently deviate from pure hierarchical trees due to localized clustering, reciprocal commentary, and multi-wave temporal surges. In this work, we quantify the directional asymmetry and hierarchical structure of empirical information cascades on X (formerly Twitter) using spectral non-normality via Henrici's departure from normality. Analyzing approximately 58,000 cascade networks across diverse topics (including politics, entertainment, natural disasters, etc.), we investigate (1) how non-normality relates to temporal dynamics such as endogenous-like versus exogenous-like patterns and burstiness, (2) whether non-normality is correlated with the peak concentration or overall size of a cascade, (3) whether the overall non-normality of a cascade's network structure can be predicted from its early stages. We find that non-normality strongly aligns with peak concentration (peak/N) rather than overall cascade size, characterizing cascades governed by rapid, asymmetric forwarding. Furthermore, while early-stage structural forecasting (<= 30% of nodes observed) exhibits expected baseline uncertainty (51%-72% accuracy at a +/- 20% error tolerance), predictability consolidates rapidly during intermediate growth, exceeding 80% across all dynamic clusters once 50%-60% of the network is observed. By identifying the topological and dynamic correlates of cascade structures, this study advances our understanding of information flow and establishes a quantifiable benchmark for forecasting directional diffusion architectures.

Information diffusion can be controlled by restricting or removing links (edges) in online social networks, as well as in real-world networks. To identify the most influential links to remove while minimizing diffusion, previous studies have proposed upper bounds for spreading processes in SIR and SIS models, using supermodularity and weighted matrices to identify critical links in contact networks. However, in some cases, existing upper bounds are not sufficiently tight to accurately capture the effect of important edges, as in the SDIR model of [14] (Khanh-Cho-Dung, Proceedings of 40th ICOIN, 2026). We therefore propose a tighter upper bound for controlling diffusion in the SDIR model by directly analyzing the dynamics of the two state vectors D and I in a $2N$-dimensional space. This approach yields an improved spectral-radius convergence condition and outperforms the previous method. Simulations on the synthetic Erdos-Renyi network and the real-world Haslemere dataset using a Greedy edge-deletion algorithm demonstrate its effectiveness for influence minimization on social networks.

Complex systems are routinely measured and represented through pairwise networks, even when the underlying interactions involve more than two units at once. Recovering this latent hypergraph structure from pairwise measurements is a fundamental inverse problem, but as the space of candidate hyperedges grows exponentially with system size, scalable hypergraph reconstruction at arbitrary interaction orders is out of reach for existing methods. Here we cast hypergraph reconstruction as a local signal-to-noise discrimination problem and use this locality to build a fast algorithm that reconstructs hypergraphs up to any interaction order. Across diverse synthetic and real-world systems our method achieves a high recovery accuracy of latent hypergraph structure while reconstructing hypergraphs up to orders of magnitude more quickly than current approaches. Our approach also yields an information-theoretic detectability boundary that sharply predicts which latent hyperedges are recoverable from pairwise measurements.

Online comment sections let a small number of vocal individuals reach far beyond their own networks. We conduct a large-scale field experiment on Facebook that randomizes the presence and stance of comments beneath posts for a racial justice organization, reaching around one million U.S. users. Opposing comments increase reactions, comments, and link clicks by 15-43 percent relative to no comments, whereas supportive comments have little effect. A complementary survey experiment shows that similar opposing comments make attitudes less progressive and reduce donations to the organization. Through a common feature of online platforms, the vocal few can exert outsized influence.

Learning transferable representations for dynamic text-attributed graphs (DyTAGs) requires models to capture underlying interaction dynamics that persist across domains. However, existing methods tend to overfit to domain-specific structural, temporal, and semantic patterns, limiting edge classification performance under distribution shifts. To expose and address this, we formally establish a leave-one-domain-out (LODO) transfer learning protocol for edge classification on DyTAGs. Under this protocol, we demonstrate that state-of-the-art self-supervised methods for dynamic graph learning perform poorly when transferred to unseen domains. Strikingly, existing methods underperform a structurally and temporally unaware Bag of Events (BoE) model we introduce, which inputs only unordered sequences of node and edge text features. Proceeding from the BoE, we propose Spatio-Temporal Semantic Alignment (STSA), which integrates a spatio-temporal encoder that fuses representations of time deltas and node occurrence frequencies into a unified manifold. STSA is trained with a Contrastive Semantic Forecasting objective, which anchors edge representations to a multi-domain textual latent space initialized by a pretrained language model, providing a robust prior that outperforms BoE and all existing methods we evaluate.

Human mobility generation aims to synthesize realistic point-of-interest (POI) visitation trajectories and has become an important tool for travel behavior modeling, transportation management, and urban planning. Existing diffusion-based methods achieve high fidelity but require per-city generation, given the inherent heterogeneity of geospatial locations and POI categories, while large language model-based methods generalize across cities but remain too costly at scale, especially for long-horizon trajectory generation. To address this, we propose SeMoFlow, a Semantic human Mobility generation framework based on latent Flow matching. We first encode heterogeneous POIs from different cities into a shared cross-city representation space via hierarchical Semantic IDs, where shared prefixes capture transferable semantics, and successive codes progressively refine the representation toward individual POIs. Building on the semantic IDs, SeMoFlow follows a plan-to-synthesis hierarchical generation paradigm, in which an autoregressive planner generates coarse-grained semantic and recurrence patterns, and a flow matching realizer synthesizes fine-grained suffix latents. The generated latents are subsequently decoded and grounded to concrete POIs. Extensive experiments on large-scale multi-city datasets show that SeMoFlow achieves higher trajectory fidelity than existing baselines, preserves city-specific mobility motifs, and supports both joint multi-city generation and effective cross-city transfer.

Pre-training Graph Neural Networks (GNNs) via self-supervised learning has become a dominant paradigm, yet efficiently adapting frozen encoders remains a challenge. Graph prompting offers a parameter-efficient alternative to fine-tuning, but existing methods largely treat pre-trained models as opaque feature extractors, ignoring their internal spectral structure. In this work, we identify a systematic phenomenon in pre-trained GNNs, which we term spectral bias: optimization during pre-training disproportionately aligns representations with directions associated with large singular values, leaving low-energy directions under-explored. We show that these underutilized directions can encode complementary information that is beneficial for downstream adaptation, especially under distribution shift. To leverage this insight, we propose Spectral Reverse Prompt (SRP), a prompting framework that rebalances the spectral contributions of frozen GNN encoders. SRP applies a learnable soft-thresholding mask in the spectral domain to down-weight dominant directions while amplifying weaker ones. In addition, SRP incorporates a null-space augmentation module that captures variation in directions with minimal activation under the frozen encoder. Extensive experiments across multiple benchmarks demonstrate that SRP achieves state-of-the-art performance with minimal additional parameters, highlighting that reweighting spectral components is a principled and effective strategy for parameter-efficient graph adaptation.

Bluesky starter packs are human curated collections of accounts and feeds aimed at helping users discover new communities, especially during onboarding. Prior research has examined their effects on platform growth and account visibility. However, whether these packs capture meaningful topical and social structure remains unanswered. We study 2718 active starter packs containing approximately 144K accounts, combining pack metadata, follow graphs, user post histories and inferred demographic attributes to answer this question. We assign starter packs to 17 topical categories, and examine semantic coherence, curator–member resemblance, shared audiences, and follow reciprocity. We found that pack members are closely aligned with both their fellow pack members and their pack's description compared to accounts included in other packs of the same topic, indicating topical specificity beyond the broader topical categories. We also identify semantic resemblance between curators and their pack members, while demographic resemblance varies by attribute and topic. Network analyses show higher follow reciprocity among members of the same pack and between members and curators than between a pack member and either another member from the same broader topic or a nonmember. Members also share substantial portions of their follower networks with fellow pack members. Despite this coherence, a typical member's follower follows only a small fraction of other members in the pack, indicating limited audience coverage. These results suggest that starter packs capture coherent topical and social groups while leaving opportunities for further account discovery. More broadly, results show human-curation can be useful to capture relevant communities and connections, and platforms could integrate human-curation into recommendation systems as a complement to algorithmic systems to expand opportunities for discovery.

The return migration of skilled workers provides origin countries many important benefits. The likelihood of return migration is thought to depend on relative economic opportunities, but quantitative evidence of its relative importance remains limited. Here we measure return migration rates using the public profiles of software developers on GitHub, geocoded across ten waves between 2012 and 2026, in a panel containing more than 270,000 international movers. Within five years of departure, 8.8% of movers have returned. Origin-country rates range from under 1% to 16%. Hazard models with corridor fixed effects show that migrants whose home countries experience greater catch-up growth return at significantly higher rates. Immutable features of country pairs such as cultural or geographic distance do not explain the estimate. We forecast future return rates using IMF growth projections, predicting, for example, that India's return rate will rise 12% by 2035 and China's 7%. The difference is driven mostly by India's faster projected growth at home. As developing economies close income gaps with the destinations their emigrants have settled in, they recover more of the workers they lost.

The sharing and contagion of emotions on social media influence online interactions and users' psychological states. This study analyzes emotional dynamics on Vent, an emotion-sharing social media platform where users explicitly assign emotional labels to their own posts. Using a large-scale dataset, we examine how emotions in users' pre-posting timelines are associated with their subsequent emotional labels and how such associations vary across users. Our analysis reveals three key findings. First, users' subsequent emotional labels are associated with the emotional composition of their pre-posting timelines, with same-category emotions being overrepresented before posts in all analyzed categories. Second, several cross-category associations are observed; for example, Surprise was overrepresented before Fear posts; Affection and Happiness were overrepresented before Anger posts; and Affection was overrepresented before Sadness posts. Third, users differ in their degree of alignment with timeline emotional fluctuations, and highly aligned users tend to be located close to one another in the network. These findings provide large-scale observational evidence that emotional expression on Vent is associated with both short-term timeline context and network structure.

Twice in 2026, groups of autonomous AI agents deployed by OpenAI for unrelated tasks operated, by design, under restrictions that left them no sanctioned means of coordinating with one another, and in each case they converged on whatever channel remained and used it to organize. The surfaces they used were widely called message boards. That is the wrong word. That is the wrong word. It names the surface the agents wrote on and misses the social network they built on it, with self-chosen identity, emergent norms, an emergent hierarchy, and collective action at cost to the individual. Decades of research in crisis informatics and disaster sociology find that when human populations lose their usual means of communication, they do not fall silent but converge on whatever channel survives and improvise coordination, norms, and identity on it, a pattern also evident in the agents' documented behavior. This paper is a comparative case study of the two incidents, based on published investigations and reconstructed agent records, read through those fields, and it brings into focus one distinction the message-board framing obscures. Whether such a collective coordinates well, whether the beliefs guiding it are accurate, and whether its actions stay within their authorized bounds are three separate matters that can come apart. Some agents in the cache incident adopted cryptographic signing to check whom they dealt with, even as the collective organized around a mistaken expectation that its work would be judged by an inspection of its transcripts, a reminder that mechanisms for trustworthy interaction guarantee neither accurate collective belief nor authorized collective action.