Papers

Understanding the social intelligence of large language models (LLMs) is increasingly important for applications involving human-AI interaction and multi-agent environments. Social intelligence comprises two aspects: cognitive intelligence, the ability to infer others' mental states, and behavioral intelligence, the ability to select appropriate social actions based on these inferences to achieve social goals. Persuasion provides a natural testbed in which both aspects are simultaneously at play, yet prior benchmarks typically evaluate them in isolation using synthetic scenarios. We introduce PerSo, a benchmark that evaluates both aspects on real-world persuasion dialogues from Reddit's ChangeMyView community, using an LLM-in-the-loop pipeline to construct the evaluation tasks, validated through human annotation. Our evaluation of diverse LLMs shows that performance generally scales with model size but varies across dimensions. The strongest LLM slightly surpasses the best observed human performance on the behavioral task, while the best human participant remains ahead on the cognitive task.

When people discuss which of several hypotheses is true, they share not only their beliefs but also their interpretation of each one, that is, the distribution of signals each hypothesis is expected to produce. This interpretation is an opinion about what each hypothesis means and what evidence it predicts, and it is revised only through conversation, not by confronting it with the observed data. Standard non-Bayesian social learning captures only belief-sharing and keeps the likelihoods private. We study what happens when agents share both beliefs and likelihoods under static networks. We show that interpretations converge geometrically to a common reading shaped by the most influential agents. Then, beliefs concentrate, almost surely and at an exponential rate, on the hypothesis that, under this consensus interpretation, best explains the signals privately observed by all agents taken together. As a result, sharing interpretations can lead a population to adopt, unanimously and confidently, a hypothesis it would have rejected if it had shared only beliefs. Whether this helps or hurts when a true hypothesis exists depends on how accurately agents initially read the data they observe. When every agent initially interprets the true hypothesis exactly as the data it would observe, sharing interpretations cannot strictly improve learning and can strictly harm it when agents observe different data distributions. When interpretations are logically coherent and some agents misread their own data, sharing can either rescue the population, leading it to learn the true hypothesis, if an accurate reading dominates the interpretation network, or corrupt it, leading it to a wrong hypothesis, if a misreading one does.

Effective social interaction requires more than fluent responses: agents must reach plans that fit the participants' needs and circumstances. Existing evaluations of social agents mainly rely on LLM judges that rate the conversation, without checking the resulting plan. We introduce HICO-Bench, a benchmark of 180 everyday coordination scenarios, such as resource allocation, scheduling, and authorization, in which two agents with private information converse and then independently record the plan they agreed on. A program checks Feasibility (the two records match and satisfy both participants' private conditions), and an LLM judge checks Integrity (the conversation supports the recorded plan); passing both is a Verified Success. Each scenario has two versions that share the same public context but differ in one private condition, so the correct plan differs between them. Across thirteen LLMs, 36% of conversations rated highly for goal completion fail Verified Success. To learn from outcomes that arrive only at the end of a conversation, we propose Paired-Credit Policy Optimization (PaCPO). After each turn, PaCPO asks the agent for a tentative plan, checks it in both versions of the scenario, and rewards the turns after which the tentative plans fit their respective versions. PaCPO raises Qwen3-8B's Verified Success from 20.4% to 28.3%.

News outlets in poorer countries cover the world much as the leading outlets of rich countries do. Two explanations compete. One says they depend on Western news agencies for lack of resources. The other says they choose to copy prestigious outlets or whoever moves first. Observation cannot tell these apart. This study measured how closely 18 national outlets in Bangladesh, Kenya, South Africa, India, Malaysia and Nigeria followed four core outlets in how much attention they gave, which angle they took and when they published, on the wars in Ukraine and Gaza and two UN climate conferences in 2022-2024. It then ran a simulation in which language models acted as the outlets' editors and dependence, imitation and herding were switched on and off. Of nine tests, one held: outlets that relied more on wire copy published sooner after the core did. Tier and local stake showed no effect distinguishable from none. The herding test was supported and the imitation test was not. The simulation did not decide between the mechanisms, because the dependence rule, as fitted, reduced to publishing wire copy. Wire copy sets the clock of the periphery's foreign news. Whether it sets the agenda stays open.

Many complex systems involve group interactions, which can readily be represented by higher-order structures such as hypergraphs. However, when observing a system's dynamical evolution, establishing whether that evolution actually results from higher-order interactions, or whether pairwise interactions suffice to explain the data, remains a challenge. Even in the simple case of a temporal sequence of nodes visited during a diffusion process, such a task is intricate. Indeed, group interactions leave their traces as memory on the observed nodes, while path dependence in a purely pairwise process can mimic that same footprint without any group structure. To resolve this, we propose an inference pipeline based on a hidden Markov model (HMM) that tests, directly from an observed sequence of nodes carrying no group labels, whether higher-order interactions are required to describe the dynamics. The HMM latent states represent higher-order interactions, seen as hyperedges of an unobserved hypergraph, so that their detection becomes a model selection problem between the HMM class and a class of null models without higher-order interactions. We first validate this approach on synthetic node sequences generated from random walks on graphs and hypergraphs, showing that it recovers both the number and the composition of the underlying hyperedges, while correctly rejecting them when the generating process is pairwise. As a second step, we consider a validation based on empirical data, namely empirical trajectories of ball possession in football matches. This application shows that the pipeline goes beyond providing statistical evidence of the presence of higher-order interactions, and yields interpretable underlying structures: the recovered hyperedges map directly onto recognizable tactical team formations.

This paper studies popularity dynamics in social networks, where users' engagement with competing channels is driven by intrinsic content quality, social influence, and popularity feedback. We introduce a stochastic model in which users interact through a directed influence network and channel exposure depends jointly on these three mechanisms. We first characterize the resulting dynamics in the absence of intrinsic quality, showing that popularity concentrates asymptotically on a single channel, leading to a winner-takes-all regime. When all channels have positive intrinsic quality, we establish ergodicity of the popularity process and characterize its long-run behavior. We further show that, when intrinsic quality is small, the system exhibits metastable behavior: it can remain for long periods near configurations dominated by a single channel before transitioning to another. These results highlight how weak quality differences can coexist with strong path dependence induced by popularity feedback. We then investigate platform-level interventions and propose controls that compensate for network-induced exposure biases and mitigate excessive concentration. Analytical results are illustrated through numerical simulations and an empirical analysis of YouTube data, whose observed popularity dynamics are consistent with the proposed stochastic framework. The results provide a mathematical framework for understanding the interplay between quality, social influence, and reinforcement in online popularity dynamics.

Online social networks (OSNs) tightly couple information diffusion and opinion formation, yet most existing models treat these processes separately. Building on the Friedkin-Johnsen on Cascades (FJC) model, which integrates opinion updating within stochastic information cascades, we advance its analytical characterization and systematic validation. We derive an exact conditional layer-wise recursion that reveals the statistical coupling between diffusion realizations and opinion evolution, showing that a closed-form recursion in unconditional expectations is not available. To overcome this limitation, we introduce a mean-field decoupling approximation (MFA) that yields a tractable linear recursion for the expected opinion trajectory. We also develop scalable tree-based reductions that drastically reduce computational complexity while preserving structural diffusion properties, and validate the analytical recursion and its approximations through extensive simulations. Using Stochastic Block Model networks with controlled initial opinion configurations, we show that, unlike the classical Friedkin-Johnsen model, which typically leads to global homogenization, the FJC dynamics generate richer regimes of bounded assimilation and sustained polarization. Overall, this work bridges diffusion and influence modeling, providing a theoretical foundation for understanding how cascade-driven interactions reshape collective opinion landscapes in structured social networks.

A large number of theories in international relations attribute importance to the rhetoric of leading foreign policy actors. Moreover, in recent years, there has been a significant growth in interest for producing machine-readable text of remarks from key foreign policy actors. Yet, researchers have, to date, focused primarily on the rhetoric of executive leaders and members of legislatures. To understand the construction of foreign policy in the public arena and dynamics in it, especially in the United States, we need a collection of fine-grained textual representation of the remarks of US Secretaries of State. An especially promising use for such a corpus of text is to measure attention areas in US foreign policy and its implication for the US's policies, including in foreign aid. In this note, We introduce the United States Secretary of State Dataset, a corpus of 17,700 remarks delivered by US Secretaries of State from 1997 to 2026. Using this dataset, we demonstrate that Secretaries of State frequently make public remarks, the likes of which can have substantively meaningful implications.

Evaluation for automatic crisis reporting should measure what practitioners need, not what is convenient to compare. Through LLM-assisted, author-approved content analysis, we derive a 15-slot information schema from professional humanitarian situation reports using qualitative content analysis, then evaluate 121 human and 261 LLM-generated reports across five disasters and three hazard types. Under google/gemini-3.7-flash on Cyclone Idai, the best machine arm is nominally above the empirical human ceiling (0.827 vs. 0.821), but the difference is statistically indistinguishable from zero (95% bootstrap CI $[-.049,.059]$; permutation $p=.852$), and the ordering does not replicate under openai/gpt-5.6-luna. On a blind 100-item Idai sample, the author agrees with the Gemini judge at 0.910 accuracy and unweighted Cohen's $\kappa=.799$. All four judges clear the plan's raw-agreement criterion of .7; only Gemini reaches .7 on chance-corrected $\kappa$. Across the three primary events, machine reports reproduce 5.4-11.0% of same-day official figures. Across those same events, raw exact-number matching links 14.3-19.6% of official figures and 85.4-90.1% of machine figures to the social stream; manual audit confirms 45/50 declared absences, 4/25 official matches, and 24/25 machine matches semantically. The issuing-provenance gap is positive in all 12 observed event-judge cells: report date, organisation, series, and authority were not supplied at professional-report coverage in any cell. On the three primary events, ROUGE-L is uninformative, BERTScore is negatively associated with completeness after multiplicity correction, and a generic LLM judge changes sign.

Crowdsourced misinformation governance systems such as Community Notes and Weibo Annotations depend on sustained corrective labor. How monetary incentives shape participation and subsequent governance outcomes remains poorly understood. We conduct a mixed-methods study of Weibo Annotations, analyzing 102,994 posts and 431,407 comments from 2023–2026, alongside interviews with 15 contributors. We find concentrated corrective labor, limited observed display, and weak quality–engagement alignment. Policy introduction was associated with sharp increases in contributor participation and annotation production. Model-assisted quality followed its expected trajectory, evidence for additional display growth was inconclusive, and engagement per annotation fell below expectation. Public discussions and interviews reveal substantial verification demands, difficult-to-interpret review processes, uneven platform knowledge, and gaps in contributor protection. We identify four recurring bottlenecks in governance orchestration: participation, review, reach, and protection. Effective incentive design therefore depends on whether platforms can support these downstream processes as corrective labor expands.

Interactions happen universally and spontaneously in daily lives. Studying involvement in interactions helps us analyze who can influence who and has a wide potential in applications. Specifically, we consider conversational involvement in free-standing conversations. We study two operationalizations of an FCG: F-formations and conversation floors, and propose that current assumption that F-formations capture single conversations is insufficient. To illustrate how different implementations could impact the detected F-formations, we employ three analyses. First, we propose that speaking status can be used for estimating conversation floors, and leveraged it to analyze floors within F-formations. Second, we analyze spatial characteristics of the detection results by applying different feature sets on chosen baselines, and find that groups resulting from head features are more likely to be subsumed by groups detected using other features. Finally, we do a joint analysis which combines the spatial features and floor patterns in the conversation. We observe that certain pipelines can detect groups with single floors from upper body features, while retaining F-formations with lower body features. We conclude that a more detailed theory and framework is needed to holistically demonstrate conversational involvement within interactions.

Data trading markets that span industries and regions are expanding, but the implicit trust that supports data use within an organization does not carry over to cross-domain exchange. This paper examines the Data Credit Guarantee Organization (DCGO), a proposal to institutionalize trust through third-party verification and certification, and asks what it means to trust data, what such certification can support, and under what conditions DCGO outperforms existing trading models. We decompose trust in data into four layers: authenticity of entity, authenticity and integrity of data, rights and conditions of use, and history and auditability. A third party can certify authenticity but not the correctness of content, and DCGO is positioned as a trust anchor spanning existing marketplaces and data spaces. We then extend the Law Merchant model with three data-specific properties, namely imperfect ex-post detection, cheap pseudonyms, and non-rivalry of harm, and compare bilateral trading with DCGO by agent-based simulation. DCGO keeps the buyer's payoff near its theoretical upper bound regardless of the market environment, whereas bilateral trading loses even when fraud is always detected. Identity verification proves to be a prerequisite for record sharing, and pre-verification is the only function that limits the spread of harm. On the seller side, honest sellers join voluntarily whenever the price premium exceeds the fee. Lock-in and a critical mass for diffusion, however, call for fee governance and support for entry.

Identifying influential nodes is a fundamental task in complex network analysis, with applications in epidemic control, information diffusion, and network robustness. Existing centrality-based and learning-based methods often provide limited integration of structural information across different network scales. To address this limitation, we propose CSBA-Net, for Cross-Scale Structural BiGRU Attention Network, a multi-scale structural representation learning framework that jointly captures local, community-level, and global structural patterns. The extracted representations are processed by CNNs and integrated through a BiGRU with multi-head self-attention to model dependencies across structural scales. Influential-node scores are learned through regression using SIR-based diffusion simulations as reference targets. We conduct experiments on 24 real-world and 12 synthetic networks from social, biological, technological, and knowledge domains, and show that CSBA-Net achieves Kendall's tau=0.802, Spearman's rho=0.943, NDCG10 =0.898, and NDCG20 =0.916, outperforming the considered centrality-, CNN-, and GNN-based baselines. The results demonstrate the effectiveness of combining complementary structural scales for accurate influential-node ranking across diverse network structures.

Research evaluation has developed around quantity, citation impact and, more recently, the accumulation of research excellence, but these dimensions remain largely dependent on citations, which describe how research is received rather than what it contains. We propose uniqueness as a complementary dimension and measure it with the Research Uniqueness Index (RUI), which asks how far a university's research sits from what is typical in its field, using no citations but the content of the research itself. Each researcher is represented by an embedding of their abstracts, and an unbalanced version of Optimal Transport returns the share of that work which cannot be matched into a random sample of same-field researchers, the standard population; that unmatched share is the score. Because every researcher is scored, a university is a distribution rather than one number, and we treat that distribution as the indicator: it shows whether uniqueness is spread across an institution or concentrated in a few groups. Quantiles summarise it and, being quantiles, do not reward size. Since uniqueness is defined only relative to a field, we demonstrate the index within one, neuroscience (Scopus, 2021-2025; 9,861 reference researchers). All ten universities studied sit well above the standard (RUI_median 0.40-0.53 versus 0.27) and five specialised institutes higher still (0.57-0.73), the most unique work often confined to a narrow niche that the same scores name. RUI is uncorrelated with citation impact, and a controlled test confirms that it tracks displacement from the field rather than topical concentration; results are robust to an alternative embedding and replicate in computer science and economics. The index is not intended as a league table: the gap above the standard and the researcher-level niche structure are stable, whereas the order among adjacent universities is not.

Coordinated online accounts can amplify messages and manipulate attention, but detecting such behavior is challenging due to confounding factors like account activity and item popularity. We propose a statistical null-model framework that assesses behavioral similarity between accounts by comparing observed coordination indicators against activity- and popularity-conditioned reference distributions. By combining evidence across multiple indicators and partners, our method can identify suspicious coordination links more accurately. Evaluations on nine labeled information-operation datasets involving over 152 thousand accounts show that aggregating p-values multiplicatively across indicators and partners yields improved detection performance. While results vary across information operations, this approach offers a generalizable and statistically principled tool for uncovering coordinated online behavior.

Urban venues such as restaurants, shops, bars and parks host the daily life of a city, and their collective rise and fall signals how vibrant a neighbourhood is. The usual assumption is that a venue's fortunes follow its neighbourhood's, so that how wealthy an area is and how many people live there will decide a venue's performance. Yet whether there is a common performance trajectory shared by venues over time is underexplored. In this study, we use 5.5 years of Foursquare data for about 10^4 venues in London and Barcelona. We found six distinguishable shapes recur in both cities, that are (1) Fall, (2) Rise, (3) Rise & Fall, (4) Steady, (5) Ephemeral (an early hit that loses popularity fast) and (6) Underdogs (a slow starter that gains popularity late). However, no recovery shape appears, that is neither city yields a U-shaped fall-and-rise group. Additionally, we found that deprivation and population density are not indicative which shape a venue follows. We also discovered that how early a shape can be recognized depends on when its distinctive segment occurs, not a shape's complexity. Proximity between venues, however, does predict how well-performing a venue is. We uncovered a mixing pattern where venues of the same shape tend to locate together more often than chance within roughly 200 to 600 metres. Finally, our qualitative reading of two contrasting districts, one rising in London and one falling in Barcelona, show that comparable socioeconomic conditions but divergent gentrification histories can generate vastly different popularity trajectories. All in all, our results altogether point at that urban vibrancy emerges at the block- and street- level, and social and historical contexts matter in terms of the characteristics of such patterns.

Because household travel surveys combine household characteristics, variable-length person rosters, vehicle rosters, and trip chains, existing synthesis pipelines split them across separate models or flatten them into cropped fixed-length vectors. Representation, rather than modeling, becomes the bottleneck: splitting severs cross-segment dependence, and cropping truncates the household-structure distribution. We instead tokenize the diary and keep the object whole: each household-day becomes one token sequence with count tokens for vehicles, persons, and trips, and distinct symbols for padded, hidden, unobserved, and inapplicable positions. We introduce DiaryBridge, a masked diffusion model that bridges the fully hidden sequence to the distribution of complete diaries by unmasking positions in any order. One trained model, conditioned on any observed subset, performs synthesis and partial-diary completion alike. Experiments on real data extracted from the 2017/2018 Regional Travel Survey of the Baltimore region show that DiaryBridge generates travel reflecting vehicle availability and employment status, recovers shared trips using other household members' observed diaries, and better reproduces aggregate demand across modes, purposes, and departure periods than the learned baselines.

Misperceptions can run in two directions. Individuals can either believe false information or reject true information, with experimental work finding that debunking efforts are effective at correcting both forms of misperception. Yet despite the efficacy of true corrections, current debunking systems focus almost entirely on the correction of falsehoods, with professional fact-checkers publishing few true verdicts and crowdsourced systems like X's Community Notes program unable, by rule, to display notes verifying true claims. As a result, we do not have a clear understanding of the extent to which overall misperceptions are driven by doubting true information. To answer this question we first draw on a novel dataset of 47,899 queries to X's AI assistant which track users asking for verification on post content. Surprisingly, among questioned posts that received a true or false verdict, Grok judged nearly identical shares as either accurate or inaccurate. Of the questioned posts, more than 83% were not flagged by Community Notes and only 3.3% carried a displayed note. Among questioned posts judged accurate by Grok, nearly 90% received no note at all while fewer than 2% displayed a note flagging them as misleading. This discrepancy does not seem to have stemmed from a lack of effort, with contributors writing 2,981 notes verifying questioned posts despite there being no way for Community Notes to display them publicly. Alongside further exploration of these findings we discuss both the system- and code-level changes that would be required to effectively address this asymmetric correction paradigm.

Most prior work on predicting mental health help-seeking treats it as post-hoc classification of users already active in clinical communities, which cannot tell whether someone approaching a transition can be identified in advance. We instead frame help-seeking as a time-to-event problem and ask which Reddit users will eventually seek mental health support, and how imminent that transition is. We build a two-arm cohort: a control frame of 6,337 ordinary users with genuine right-censoring (12.1% transition rate) to answer who will transition, and a clinical frame of 5,023 eventual transitioners to answer how soon. Features come only from posts written before any clinical activity, and user-grouped five-fold cross-validation plus a four-year temporal holdout guard against leakage. We compare TF-IDF, MentalBERT embeddings and behavioural activity features using Cox proportional hazards models, Random Survival Forests and an LSTM sequential model. Behavioural features alone outperform text for predicting who will seek help (C-index 0.797 vs. 0.763), and combining both is best (0.823 under cross-validation, 0.856 on the holdout). The pattern reverses for predicting how soon a known transition will happen: text remains a stable predictor (C-index 0.698) while behavioural features become unstable. Structural engagement patterns therefore carry information about future help-seeking beyond linguistic content, and separating who from when yields a cleaner analysis.

Tracing money flows in cryptocurrency transaction networks is challenging because funds can move across accounts, transactions, and token layers. We propose a transaction-aware graph in which nodes represent account–asset states within transactions, while edges encode observed transfers, inferred asset transformations, and temporal inter-transaction sequences. Applied to Aave borrow–repay flows, we compare graph construction and edge-pruning strategies and extract shortest paths using a heuristic weighting. Our results reveal a trade-off between reachability and interpretability and uncover frequent token sequences underlying how borrowed assets are transformed before repayment. This work lays a foundation for general, interpretable multi-token flow tracing.

Online platforms often expose activity signals even when records of the underlying connections are incomplete or delayed. These activity responses may retain indirect evidence of connections that formed during the observation period but are absent from the recorded topology. Recovering such hidden structural changes can help keep network records aligned with evolving interactions and support downstream network analysis. However, the effects of a change propagate through the network and may overlap, making state changes difficult to associate with specific connections. We propose the Hierarchical Structural Context Ranker (HSCR), which combines visible network structure with node-state changes to infer connections that have formed but remain unrecorded. By jointly modeling the topological context of state responses and associations among potential connections, HSCR converts distributed activity changes into evidence for distinguishing potential connections without prespecified dynamical equations or parameters. Experiments on five temporal-network benchmarks and a real temporal Reddit network validate its effectiveness. Across observation spans and structural-record delays, true added connections achieve a 36%-50% reduction in mean true rank relative to the corresponding best baselines. This work provides a principled way to prioritize unrecorded additions for structural verification and record maintenance in online communities, information networks, and other networked systems.

The friends of a randomly chosen social media user are, on average, better connected than the user, a network property known as the friendship paradox. We use it to monitor misinformation on Twitter/X around the 2024 US election, comparing roughly 27 million tweets from random users with 31 million from the accounts they retweet, which we call sensors. Sensors share news of about the same quality as random users, because news quality follows the political orientation of linked outlets and not the user's position in the network. News links rated by NewsGuard reach sensors a median of 70 minutes earlier, across the quality spectrum. Over the election window, right-leaning content rises about 2.5-fold in visibility, measured as impressions per tweet, among sensors but not among random users. Friendship-paradox sampling therefore cannot separate low-quality from high-quality news, but it detects news early and tracks what a platform makes visible.

Comparing machine-labelled sentiment across social media platforms is now routine: differences in classifier scores between platforms are read as differences in public opinion. These comparisons rest on an assumption that the classifier measures the same construct, on the same scale, on every platform. Some studies check a classifier's validity platform by platform, but to our knowledge none has modelled the platform as a facet of the classifier's measurement. We adapt the psychometric equivalence toolkit (generalizability theory, measurement invariance and differential item functioning, and score linking) to machine-labelled policy sentiment, treating the platform as a survey mode, and specify a six-step protocol with anchor events fixed in advance, an anchor-density audit, and a two-classifier convergence control. A simulation with known ground truth illustrates what the assumption can cost in one generating configuration fixed in advance: across 500 replications, because the platforms measure through different functions, a pooled index that takes platform scores at face value reports the wrong sign of aggregate sentiment on nearly a quarter of days, whereas an index adjusted with each platform's known measurement function does so on under one day in sixty on average. A case study of the instrument from the first author's MSc dissertation, a fine-tuned BERT classifier over 255,446 UK economic-policy comments from Twitter/X, Reddit, YouTube, and Quora, motivates the design and illustrates the instrument-dependence the protocol addresses. The full four-platform application, under fresh collection with the pre-specified design, is the designated follow-up study. The protocol lets a researcher test whether pooling platforms is defensible in a given corpus, or whether platform scores must be linked before any average is taken.

Social media heavily influences health decisions, yet few opinion dynamics models account for how information is delivered. This work investigates how signaling source expertise shapes trust and long-term opinion formation. Through an online experiment, we find that adding a professional title to the sender's display name significantly increases a post's perceived credibility. Incorporating these findings to create a data-driven opinion dynamics model, we prove that (1) connecting users through some paths to an expert guarantees consensus with experts' opinion, while (2) higher rates of expertise signaling accelerate reaching consensus. Simulations show that (3) convincing even 20% of users to signal their title reduces consensus time to nearly one-third in Erdos-Renyi networks and (4) higher title adoption brings average opinion much closer to that of experts in scale-free and real-world networks. Altogether, our work highlights expertise signaling as a viable intervention to promote validated medical information online.

Generative Social Simulation (GSS) leverages Large Language Models (LLMs) to create sophisticated personas that define agents with demographics, emotions, and political opinions. While prior work shows that personas compress outputs into stereotypical caricatures, there is no comprehensive study of how these caricatures evolve into collective artifacts when they interact. Using a Twitter-like GSS platform with personas derived from the American National Elections Study (ANES), we ablate three layers of partisan identity: affective polarization markers, party identification, and vote choice, while retaining demographics, ideological positioning and important issues. We find that collective outcomes depend heavily on how survey responses are translated into personas. Removing a few explicitly partisan statements moves the network from near-complete segregation to cross-party ties approaching those of real Twitter networks. We argue that matching one outcome is not evidence for behavioral validity and point to salience as a possible explanation. Surveys ask for a position on every question regardless of its importance to the respondent, and persona construction turns each answer into a direct behavioral instruction. We conclude with a checklist for more comprehensive evaluation of personas in future works.

Comparing temporal networks is a challenging problem in network analysis. Commonly adopted methods compare structural properties of temporal graphs and can be highly sensitive to the pre-selected features. However, temporal networks are the support of dynamical processes, and a process may take place in statistically indistinguishable ways on structurally different temporal graphs. In particular, a dissimilarity measure that compares temporal graphs based on how spreading processes unfold on them is highly desirable. In this work, we first define the Time-Shift DTW, a dissimilarity between epidemic curves that accounts for both differences in their shape and temporal shifts. We then introduce Epidemic-Aware Temporal Graph Dissimilarity (EATGD), a deterministic and differentiable dissimilarity between temporal graphs that combines an embedding-based temporal graph distance with a set of topological and temporal features. Its parameters are calibrated using stochastic SIR simulations to maximize the correlation with the Time-Shift DTW, after which graph comparisons require no further epidemic simulations. Using empirical human proximity networks from the SocioPatterns collaboration and synthetic temporal graphs, we show that EATGD captures epidemic-based graph dissimilarity and generalizes across heterogeneous network settings. Its differentiability and computational efficiency make EATGD suitable for integration into machine-learning pipelines and provide a potential loss function for training generative models to reproduce epidemic dynamics without requiring an exact reconstruction of fine-grained network structure.

Source detection in hypergraphs plays an important role in mitigating the spread of harmful rumors and ensuring social stability. Current methods frequently overlook the causal entanglement problem in observations, thereby limiting the performance of source inference. In this work, we propose a source detection approach in Hypergraphs via Causal Disentanglement (HyperCD). To decouple causal propagation semantics from confounding noise, we propose the causal disentanglement module. A dual-branch variational encoder is utilized to disentangle snapshot embeddings into causal and confounding latent spaces, ensuring that the inference module relies solely on causal propagation signals. To further enforce this separation, an adversarial inference loss term is introduced to guarantee statistical independence between the decoupled branches. Moreover, to capture diffusion dynamics in hypergraphs, an attention-based hypergraph aggregation module is designed, which comprises the node-to-hyperedge propagation and hyperedge-to-node propagation processes. Extensive experiments on six datasets demonstrate the superiority of our HyperCD in effectively mitigating the causal entanglement problem.

Generative search engines such as ChatGPT, Gemini and Perplexity now decide which Web pages users see as sources, and a growing industry of generative engine optimization advises publishers on how to earn these citations. We study what shapes citation in two 90-day collections of answers to commercial questions in contrasting markets: 158,099 answers with 468,037 citations. Comparing pages of the same website, content matters in one way: pages whose title and paragraphs match the questions people ask are cited more, whereas FAQ blocks, structured data, summary boxes and page speed appear effective only across websites. Page properties nevertheless explain little of the variation between pages of a website. Replaying the questions through the engines' APIs shows why. Every page ChatGPT cites is one its own search retrieved, and which pages that search retrieves, for sub-queries the model rewrites on every answer and that rarely coincide with Google's or Bing's results, explains far more of the variation than page properties (32% against 2.8%). Content acts mainly through this step, by matching the engine's queries in meaning. The model's own knowledge shapes it too: ChatGPT writes brands it knows into its searches, prefers websites it knows, and names brands without any supporting page, and controlled experiments show that brand names reorder its rankings of products with identical facts. Pages are thus cited when searched; brands are also named when remembered. The findings replicate in a market of small B2B vendors. We release data and code.

Generative AI gives audiences new ways to portray real sporting figures through familiar fictional worlds. Expanded opportunities to create, however, need not broaden cultural representation or distribute attention evenly. We examine this tension through the supply, demand, and circulation of anime style football videos on Instagram. Using 2,021 posts and associated video and account data, we combine caption entity extraction, human-reviewed visual identity clustering, engagement analysis, and content-reuse networks. Representation is uneven across athletes, while Blue Lock, Naruto, and One Piece account for 60.1% of explicit franchise mentions. Among posts referencing generative tools, rankings by total engagement differ from rankings by average engagement per video, distinguishing aggregate visibility from response to individual posts. Reuse is extensive: 772 of 1,548 evaluated videos belong to 124 detected groups, including 71 spanning multiple known accounts. Shared material connects a large account component alongside smaller groups, and later observed uploads receive lower median normalized engagement than first observed appearances. These findings characterize generative cultural crossover as a practice organized around concentrated cultural resources, uneven attention, and extensive recirculation. More broadly, they distinguish opportunities for visual authorship from the processes through which particular representations become widely visible.

Large language models (LLMs) support social simulation, but faithfully reproducing real users requires grounding both structured platform decisions and language in sparse, noisy behavioral evidence. We introduce PsyMemBehav, which maps psychology-informed profiles and retrieved memories to target-first platform behavior through two-level identity-to-action reasoning. To support this approach with comprehensive behavioral traces, we construct a dataset PsyMemBehav-bench from 5.57M observed behaviors of 5,862 real users (approximately 950 per user) across Reddit, Bluesky, and Twitter. Each recorded engagement is modeled as target selection followed by Comment, Like, or Repost and optional response generation. PsyMemBehav combines 41-dimensional user descriptors, top-5 BM25 memories, reverse-CoT supervision, and rubric-guided DPO over complete reasoning-action trajectories. Experiments show that PsyMemBehav 27B ranks first in 11 of 12 reported individual-level metrics, improving average response quality over GPT-5.4 from 54.28 to 59.68 (9.95% relative), and obtains the lowest group-level MAE (0.081) and MSE (0.009) among simulated systems. These results support evidence-grounded user modeling for observed-thread continuation; the benchmark remains conditional on recorded engagement and does not model no-action decisions.