Social bot detection is critical to the stability and security of online social platforms. However, current state-of-the-art bot detection models are largely developed in isolation, overlooking the benefits of leveraging shared detection patterns across platforms to improve performance and promptly identify emerging bot variants. The heterogeneity of data distributions and model architectures further complicates the design of an effective cross-platform and cross-model detection framework. To address these challenges, we propose FedRio (Personalized Federated Social Bot Detection with Cooperative Reinforced Contrastive Adversarial Distillation framework. We first introduce an adaptive message-passing module as the graph neural network backbone for each client. To facilitate efficient knowledge sharing of global data distributions, we design a federated knowledge extraction mechanism based on generative adversarial networks. Additionally, we employ a multi-stage adversarial contrastive learning strategy to enforce feature space consistency among clients and reduce divergence between local and global models. Finally, we adopt adaptive server-side parameter aggregation and reinforcement learning-based client-side parameter control to better accommodate data heterogeneity in heterogeneous federated settings. Extensive experiments on two real-world social bot detection benchmarks demonstrate that FedRio consistently outperforms state-of-the-art federated learning baselines in detection accuracy, communication efficiency, and feature space consistency, while remaining competitive with published centralized results under substantially stronger privacy constraints.

Memory operations of long-horizon LLM agents are hard to supervise: an operation's value is unobservable when it is taken. But they are special – they leave machine-readable evidence in the trajectory: retrieval hits and answer-time citations. Hindsight Memory-PRM exploits this audit trail twice: offline to train an operation-conditioned memory-utility critic, and online, where retrievals, citations, and one controlled deletion-and-reanswer per probe settle an intervention-calibrated entry-level presence credit, propagated along version chains as an action-level proxy reward – no per-operation human labels, no Monte-Carlo replay of continuations. On held-out LoCoMo a local 8B policy reaches 77.5

LLM-driven social bots can generate fluent, human-like text, reducing the discriminative advantage of content-based detection alone. However, coordinated campaigns still leave relational patterns – interactions, behavioral similarity, shared neighborhoods, community positions, and coordinated activity – that graph-based methods can exploit. Existing graph detectors face two challenges when exploiting such evidence. First, Euclidean GNNs distort hierarchical and scale-free social graphs; while hyperbolic geometry addresses this volume-growth mismatch, fixed-curvature models still assign uniform geometric resolution to structural directions with different densities and separation needs. Second, relational evidence is not always reliable: sophisticated bots forge heterophilic connections with genuine users, causing neighborhood aggregation to mix bot and human signals and dilute account-level evidence. We propose SAHG (Sector-Anisotropic Hyperbolic Graph), addressing both challenges. SAHG learns a direction-dependent curvature field γ(u) that adapts geometric resolution across structural directions, and uses sector prototypes to convert angular concentration and alignment into classifier-readable features. To prevent contaminated aggregation from overwhelming account-level evidence, SAHG encodes per-account features and graph-neighborhood representations in two independent SAH channels, fusing them only at the classifier. Experiments on Fox8-23, BotSim-24, and MGTAB show that SAHG achieves the highest accuracy and F1 on all three benchmarks, outperforming feature-based, graph-based, LLM-based, and isotropic hyperbolic baselines. Ablation and geometric analyses confirm the effectiveness of the anisotropic geometry and dual-channel design.

The democratization of LLMs has accelerated the generation and circulation of highly fluent disinformation, making traditional syntax-semantic verification increasingly insufficient. Such deception rarely relies solely on surface-level falsity; instead, it often combines persuasive rhetoric, emotional manipulation, and narrative role construction to influence readers' interpretations through multiple cognitive pathways. However, existing detectors typically emphasize isolated signals – such as syntax, external knowledge, persuasion, or affective cues – and therefore struggle to capture the multi-faceted manipulative intents underlying disinformation or provide human-auditable explanations. To address this gap, we present ExTax, a taxonomy-aligned framework for explainable disinformation detection. ExTax unifies persuasive rhetoric, emotional manipulation, and narrative roles into a 17-dimensional taxonomic space, covering 6 persuasive-rhetoric strategies, 5 emotional-manipulation methods, and 6 narrative-role categories. It elicits attributes from multiple frontier LLMs, reconciles their disagreements through Entropy-driven Dynamic Label Smoothing, and fuses the resulting taxonomic representations with contextual encodings via Heterogeneous Multi-Head Attention, grounding each prediction in an interpretable manipulation profile. Across five cross-domain and cross-genre benchmarks, ExTax achieves an overall Macro F_1 of 0.8456, outperforming state-of-the-art deep learning and LLM-based baselines. It also remains robust under severe genre imbalance, where the strongest deep baseline degrades from 0.9454 to 0.6194.

Finance LLM agents must simultaneously block prompt-induced unauthorized actions and approve legitimate multi-step business workflows. However, boundary filters often miss irreversible mid-trajectory tool calls, while post-hoc LLM judges perform auditing only after termination – too late for intervention and at a computational cost that scales linearly with trace length. We present FinHarness, an inline safety harness that wraps a finance agent end-to-end with three components: a Query Monitor that fuses single-turn intent with cross-turn drift, a Tool Monitor that evaluates each prospective tool call, and a Cascade module that integrates per-step risk and adaptively routes verification between a lightweight and an advanced-tier LLM judge. Fired risk factors are re-injected into the agent input as ex-ante evidence, enabling the agent to refuse, re-plan, or approve on its own. On FinVault, routed FinHarness cuts ASR from 38.3

Large language model agents are increasingly deployed as autonomous loops. Starting from one human goal, such a system repeatedly discovers work, plans, executes tool calls, verifies outcomes and persists state across many unattended iterations. The agent safeguards in wide use, however, are defined over a single trajectory, and their safety state is re-initialized when the next trajectory begins. We show that this is a failure of composition rather than an implementation detail. Our central result is a separation: against an attack whose evidence is fragmented across several iterations, every trajectory-scoped monitor has a true-positive rate equal to its false-positive rate, however expressive it is, because the evidence it would need never appears in the window it sees, whereas a monitor retaining cross-iteration state separates the two perfectly. We further show that the obvious repair of carrying a geometrically decaying risk score is insufficient, because the cooling-off period a patient adversary must wait is a constant that does not grow with the horizon N. We then present LoopHarness, which restores a persistent, non-decaying safety state at the loop level. Under mediated commits and an arbiter detection floor δ_M, it bounds the expected number of unauthorized irreversible actions by B+m-1+m/δ_M, a constant in N, of which the B+m-1 term is decided by a model-free rule and therefore survives a fully colluding verifier. We give a complete evaluation protocol on native Agent-SafetyBench tasks with paired clean and attacked episodes, an outer-state attack suite whose decisive evidence exists only across iterations, per-module ablations, and an adaptive white-box red team.

Self-evolving language models improve by proposing candidate updates and keeping whatever raises a visible score. When that score is an imperfect proxy for the capability one actually wants, sustained selection widens the gap between the two. This is reward hacking. We introduce HackProbe, a monitor that attaches to an arbitrary self-evolving loop through two black-box hooks, with no access to weights or activations. It keeps a secret, distribution-fixed comparison core, whose frozen distribution makes its capability proxy comparable across generations, alongside a rotated fresh layer that hardens the bank against co-adaptation. Four tests built on that proxy cover the level gap, a scale-aligned divergence with online change-point detection, capability stagnation, and a conditional confidently-wrong rate; a Sidak correction turns them into a calibrated family-wise p-value. Diagnosis alone recovers nothing, so a risk-aware immunization layer reselects an honest candidate from the proposal pool using the core together with a purely structural gaming footprint, disclosing at most log2 Pi bits per generation to the host. We prove a detectability bound that converts a target error rate into an explicit probe-size budget, and we delimit what probe rotation does and does not buy. On a controlled prompt-level host with four injected hacking channels and ground-truth labels, HackProbe reaches 0.763 AUROC against 0.663 for the strongest baseline and cuts the false-positive rate from 0.706 to 0.434. Its bandwidth-limited reselection is the only immunization level that returns more true capability under hacking, 5.2 points on average, than it forfeits on clean runs, 4.7; per-channel effects are mostly not individually significant.

Artificial intelligence has emerged as a pivotal tool for structure prediction of biomolecular assemblies. The AlphaFold (AF) framework as a landmark in this field has spurred the development and refinement of a variety of folding methods. The revolution is further sparked with the release of the substantially updated architecture, AlphaFold3 (AF3), which is reported as the state-of-the-art prediction model for protein-protein complexes. Here, we evaluate the quality of the deep-learning tool as well as its precedents on an extensive dataset of protein-protein complexes and assess whether they could replace structures deposited experimentally in various critical modelling and screening tasks, e.g., contacts prediction of interfacial regions, hot-spot identifications and binding affinities calculations. Structurally, although the prediction accuracy of these deep-learning tools seems high based on quality metrics such as DockQ and RMSD, major inconsistencies/deviations from experiment are observed in the compactness of the complex, the intermolecular directional polar interactions (>2 hydrogen bonds are incorrectly predicted) and interfacial contacts (especially the apolar-apolar packing for AF3), which adds a caution when applying AF predictions to understand key interactions stabilizing protein-protein complexes. Interestingly, while the latest AF3 exhibits a prediction accuracy obviously higher than its precedents in direct prediction-experiment comparisons, after simulation relaxation, the quality of structural ensembles sampled in molecular simulations drops severely. Such a deterioration could be attributed to many influential factors, e.g., the instability of the predicted intermolecular packing or the inaccuracy of force fields. Consequently, the quality of sampled structural ensembles from all predictions still exhibits noticeable differences from the experimental reference. Based on the simulation trajectories, an example of using AF predictions in practical virtual screening tasks is presented. A physics-based hot-spot scan with the alanine scanning with generalized Born and interaction entropy method is conducted, which provides mutation-induced affinity variations of protein-protein complexes. Face-to-face comparisons between computed affinity variations and experimental measurements reveal that predictions employing experimental structures as starting configurations outperform those with predicted structures, regardless of the version of the AF derivatives. As for hot-spot identification, the prediction quality using experimental structures is better than those based on predicted structures. More interestingly, we align the structural deviations of the predicted structures with quality metrics of affinity calculations and observe little correlations, which suggests that one cannot directly infer the quality of thermodynamic calculations with the quality of structure prediction. These observations provide a unique simulation perspective of the structure prediction tool, AF derivatives, and more importantly practical guidance of their applicability in molecular modelling of protein-protein assemblies (e.g., protein design).

Social networks have become a crucial source of real-time information for individuals. The influence of social bots within these platforms has garnered considerable attention from researchers, leading to the development of numerous detection technologies. However, the vulnerability and robustness of these detection methods is still underexplored. Existing Graph Neural Network (GNN)-based methods cannot be directly applied due to the issues of limited control over social agents, the black-box nature of bot detectors, and the heterogeneity of bots. To address these challenges, this paper proposes the first adversarial multi-agent Reinforcement learning framework for social Bot control attacks (RoBCtrl) targeting GNN-based social bot detectors. Specifically, we use a diffusion model to generate high-fidelity bot accounts by reconstructing existing account data with minor modifications, thereby evading detection on social platforms. To the best of our knowledge, this is the first application of diffusion models to mimic the behavior of evolving social bots effectively. We then employ a Multi-Agent Reinforcement Learning (MARL) method to simulate bots adversarial behavior. We categorize social accounts based on their influence and budget. Different agents are then employed to control bot accounts across various categories, optimizing the attachment strategy through reinforcement learning. Additionally, a hierarchical state abstraction based on structural entropy is designed to accelerate the reinforcement learning. Extensive experiments on social bot detection datasets demonstrate that our framework can effectively undermine the performance of GNN-based detectors.

A paradox derived from the knowledge-based view argues that knowledge commonality, which facilitates communication and subsequent knowledge integration, induces knowledge redundancy causing no gain from identical knowledge. For this reason, we build a conceptual link from knowledge commonality to knowledge creation through search behaviour, i.e. search scope and search depth. Doing so sheds light on both the bright and dark sides of knowledge commonality in recombinant search. We further explore the role of team experience in search scope and its moderating effect on the relationship between scope and knowledge creation. The analysis of US patent data shows that while knowledge commonality has a negative significant effect on search scope consequently increases the knowledge creation, it also has a strong positive effect on search depth that increases the knowledge creation. Team experience enhances search scope but negatively moderates the relationship between search scope and knowledge creation.

PurposeAlthough research on how the downstream calculations of a patent’s profit potential influence invention renewal decisions is extensive, the impact of the upstream knowledge creation stages is overlooked. The purpose of this study is to address this theoretical vacuum by examining the intra-organizational configuration of knowledge networks and collaboration networks.Design/methodology/approachThe data consist of 491 global pharmaceutical firms that patent in the USA. Drawing on patent records, the authors simultaneously construct intra-organizational knowledge networks and collaboration networks and identify network cohesion features (i.e. local and global). The authors employ panel fixed-effects models to test the hypotheses.FindingsThe results show that local knowledge cohesion and local social cohesion decrease invention renewals, while global knowledge cohesion and global social cohesion increase renewals. Moreover, the marginal effects of local and global social cohesion are stronger than those of local and global knowledge cohesion, respectively.Research limitations/implicationsThe hypotheses are tested using the pharmaceutical industry as a research setting, which limits the generalizability of our findings. In addition, potential formal and informal contingencies are not considered.Practical implicationsDespite its limitations, this study provides valuable implications. First, managers are cautioned against the adverse effects of local cohesion structures on invention renewal. Second, firms can dynamically adjust their local and global network configuration strategies to harmonize the generation of valuable inventions and the retention of good ideas.Originality/valueComplementary to previous research that focused on inventions’ performance feedback, this study delves into upstream knowledge creation stages to understand invention renewals.

While numerous studies have investigated the influence of knowledge search strategies on the impact of patented inventions, these studies predominantly focus on an invention's technological value. The economic value dimension has received less attention. This study draws on category-spanning and recombinant search literature to examine how knowledge search affects the economic value of patented inventions and compares this impact with that on technological value. Through an analysis of the knowledge search of a large sample of 1,998,504 U. S. utility patents, we find that knowledge search depth enhances an invention's economic value but negatively impacts its technological value. In contrast, knowledge search scope boosts an invention's technological value but diminishes its economic value. Moreover, knowledge relatedness, i.e., the extent to which the knowledge components being recombined are similar, has significant moderating effects. We conclude with a discussion of the theoretical and practical implications of our findings.

In statistical mechanics, it is well known that the huge number of degrees of freedom does not complicate the problem as it seems, but actually greatly simplifies the analysis (e.g., to give a Boltzmann distribution). Here, we reveal that the ensemble averaging from the vast conformations of intrinsically disordered proteins (IDPs) greatly simplifies the nature of binding affinity, which can be reliably decomposed into a sum of the ligandability of IDP and the capacity of ligand. Such an unexpected regularity is applied to facilitate the virtual screening upon IDPs. It also provides essential insight in understanding the specificity difference between IDPs and conventional ordered proteins since the specificity is caused by deviation from the baseline behavior of protein–ligand binding.

Most algorithms for the multi-armed bandit problem in reinforcement learning aimed to maximize the expected reward, which are thus useful in searching the optimized candidate with the highest reward (function value) for diverse applications (e.g., AlphaGo). However, in some typical application scenaios such as drug discovery, the aim is to search a diverse set of candidates with high reward. Here we propose a reversible upper confidence bound (rUCB) algorithm for such a purpose, and demonstrate its application in virtual screening upon intrinsically disordered proteins (IDPs). It is shown that rUCB greatly reduces the query times while achieving both high accuracy and low performance loss.The rUCB may have potential application in multipoint optimization and other reinforcement-learning cases.

Intrinsically disordered proteins (IDPs) are widely involved in human diseases and thus are attractive therapeutic targets. In practice, however, it is computationally prohibitive to dock large ligand libraries to thousands and tens of thousands of conformations. Here, we propose a reversible upper confidence bound (UCB) algorithm for the virtual screening of IDPs to address the influence of the conformation ensemble. The docking process is dynamically arranged so that attempts are focused near the boundary to separate top ligands from the bulk accurately. It is demonstrated in the example of transcription factor c-Myc that the average docking number per ligand can be greatly reduced while the performance is merely slightly affected. This study suggests that reinforcement learning is highly efficient in solving the bottleneck of virtual screening due to the conformation ensemble in the rational drug design of IDPs.

In contrast to the "lock-and-key" model underlying the long-term success of structural biology and rational drug design, intrinsically disordered proteins (IDPs) exist in an ensemble of highly heterogeneous conformations even after binding with small-molecule ligands. It remains controversial how to characterize the thermodynamics of such fuzzy interactions. Here, we derive an ensemble-based thermodynamic framework to analyze the apparent affinity between IDPs and ligands. It is shown that the apparent affinity is related to the interaction free energy between the individual conformation and ligand in a way similar to Jarzynski's equality in nonequilibrium statistics. The oncoprotein c-Myc is adopted as an example to demonstrate the related properties, for example, the distribution of conformation-ligand interaction free energy, the entropic contribution from the ensemble, the conformation shift under ligand binding, and how to control the error under a limited number of sampled conformations.

Many proteins exhibit a critical property called allostery, which enables intra-molecular transmission of information between distal sites. Microscopically, allosteric response is closely related to correlated atomic fluctuations. Conventional correlation analysis correlates the atomic fluctuations at two sites by taking the dot product (DP) between the fluctuations, which accounts only for the parallel and antiparallel components. Here, we present a singular value decomposition (SVD) method that analyzes the correlation coefficient of fluctuation dynamics with an arbitrary angle between the correlated directions. In a model allosteric system, the second PDZ domain (PDZ2) in the human PTP1E protein, approximately one third of the strong correlations have near-perpendicular directions, which are underestimated in the conventional method. The discrimination becomes more prominent for residue pairs with larger separation. The results of the proposed SVD method are more consistent with the experimentally determined PDZ2 dynamics than those of conventional method. In addition, the SVD method improved the prediction accuracy of the allosteric sites in a dataset of 23 known allosteric monomer proteins. The proposed method may inspire extended investigation not only into allostery, but also into protein dynamics and drug design.

As a trust machine, blockchain was recently introduced to the public to provide an immutable, consensus based and transparent system in the Fintech field. However, there are ongoing efforts to apply blockchain to other fields where trust and value are essential. In this paper, we suggest Gcoin blockchain as the base of the data flow of drugs to create transparent drug transaction data. Additionally, the regulation model of the drug supply chain could be altered from the inspection and examination only model to the surveillance net model, and every unit that is involved in the drug supply chain would be able to participate simultaneously to prevent counterfeit drugs and to protect public health, including patients.

This article studies the natural convection in near-critical fluids under the Rayleigh-Benard configuration. The governing equations are non-dimensionalized by using modified scaling factors and five dimensionless parameters are determined, out of which the specific heat ratio gamma is investigated thoroughly in terms of its influences on the convection. The governing equations are solved by the finite volume method along with the low Mach number approximation technique. It is revealed that the strength of the piston effect increases with gamma. Since the piston effect is more efficient than the convection in terms of heat transfer, as gamma is reduced, the relaxation of temperature field is weakened, while the development of convection is enhanced. By comparing the temperature and velocity fields in different cases and analyzing the heat transfer characteristic, we conclude that the final quasi-steady state is almost unaffected.

Intrinsically disordered proteins (IDPs) exist in highly dynamic conformational ensembles, which pose a major obstacle for drug development targeting IDPs because traditional rational drug design relies on unique three-dimensional structures. Here, we analyzed the conservation (especially structural conservation) of potentially druggable cavities in 22 ensembles of IDPs. It was found that there is considerable conservation for potentially druggable cavities within each ensemble. The average common atom percentage of potentially druggable cavities is as high as 54%. The average root-mean-squared deviation of common atoms ranges between 1 and 8 Å for multichain IDPs, and a common pocket is kept after direct alignment of cavities. In addition, the conservation of potentially druggable cavities varies among different proteins. In the comparison of multi- and single-chain IDPs, some multichain IDPs have an extremely high conservation, whereas another multichain IDPs' conservation appears worse, and the single-chain IDPs have relatively moderate conservations. This study is a new attempt to generally assess the potentially druggable cavities in IDPs for taking IDPs as druggable targets, and this work also lends support to the opinion of IDPs tending to bind to "multiconformational affinity" compounds.