Science & Mathematics

The use of generative AI in game development is growing rapidly and is the subject of intense debate in which player reactions play a central role. This paper contributes to this discussion by analyzing Steam reviews to examine whether games with AI disclosure are associated with different review patterns. Using a dataset across multiple disclosure groups, we apply linguistic analysis including keyword-based, word frequency, and characteristic word analysis to investigate the AI-related discussion on Steam. Our results show that AI-disclosed games are associated with higher levels of AI-related criticism in negative reviews. Our findings provide a first exploratory step towards understanding how players perceive AI disclosure.

While Multimodal Large Language Models (MLLMs) have revolutionized general visual perception, existing methods suffer from severe hallucination in the aerial domain, where extreme scale variations and visual ambiguity abound. Existing Supervised Fine-Tuning (SFT) paradigms force models to mimic ground truth without modeling uncertainty, often resulting in blind guessing on tiny or unrecognizable objects in aerial scenes. In this paper, we present Aerial-R1, the first aerial MLLM framework to reinforce visual reasoning capabilities for aerial reasoning and segmentation. Drawing inspiration from the reasoning-incentivizing mechanisms of DeepSeek-R1, we find that trustworthy perception stems from consistency across views, establishing a visual parallel to the Chain-of-Thought in logic. To effectively adapt policy optimization for non-autoregressive mask generation, we introduce Hybrid-View Group Advantage-Weighted Regression (H-GAWR). This strategy constructs dynamic groups of high-resolution and low-resolution views, seamlessly bridging group-relative advantage estimation with advantage-weighted policy extraction. By enforcing consensus on clear inputs and strictly penalizing hallucinations on ambiguous ones, Aerial-R1 develops an explicitly incentivized refusal mechanism. Extensive experiments on VSAI-Ref, the first aerial MLLM benchmark dedicated to jointly evaluating dense reasoning and hallucination robustness, demonstrate that Aerial-R1 significantly outperforms SFT baselines in segmentation precision while substantially mitigating fatal hallucinations in low-information scenarios, paving the way for more reliable and trustworthy aerial agents. Codes are available on https://github.com/phy709/Aerial-R1.

Reflective text-space optimizers such as GEPA and SkillOpt have reported large gains on frontier and mid-size models. We test replication across two reasoning-native target-serving configurations: DeepSeek V4 Flash through DwarfStar and Qwen3.8-27B through LM Studio. For each target, we evaluate the same held-out tasks and four conditions - minimal baseline, GEPA prompt, SkillOpt skill, and manually engineered prompt - on HotpotQA (n=500 per condition) and MATH-500 (n=250). In the primary Qwen v2 rerun under the same frozen protocol, GEPA selects the HotpotQA seed and exactly reproduces baseline behavior, whereas SkillOpt's validation gain reverses into held-out degradation. On MATH-500, GEPA raises exact match from 0.828 to 0.904 while reducing latency but increasing token consumption, so lower response time does not imply lower token cost. DeepSeek likewise shows its strongest GEPA effect on MATH-500 (0.852 to 0.908). Across both targets, GEPA's strongest gains occur on MATH-500, while neither optimizer improves the HotpotQA baseline. Optimization is therefore task- and method-dependent, and validation gains may reverse on held-out evaluation. An earlier Qwen run is retained as serving-sensitivity evidence. These results suggest that textual headroom for a model-task pair and the observed effect of a text-space optimizer are more informative when decomposed into content/reasoning, answer-realization/evaluator alignment, and serving-completion/reliability. The cross-benchmark pattern further suggests that optimization may benefit from low-ambiguity, actionable evaluation signals.

We introduce and study the adaptive covering complexity C adapt S (s), defined as the minimum number of contiguous factor blocks required to partition a string s when each block is either a factor of a fixed external dictionary S or a factor of the target's already decoded prefix. This hybrid admissibility model formalizes the structural interaction between static dictionary matching and dynamic nonoverlapping self-referencing. Our main theoretical result is the exact logarithmic law which is tight and proves that dynamic self-reference can reduce phrase count by at most an exponential factor over static dictionary covering alone. We establish structural equivalences showing that the measure bridges word-segmented Relative Lempel-Ziv and non-overlapping LZ77, and we prove exact preset-history LZ77 sandwich bounds. Furthermore, we establish the optimality of the longest-match greedy parse via an exchange lemma and present an optimal implementation based on dual suffix automata running in O(M + n) time and space for a constant alphabet. These analytical results are complemented by counting bounds, stationary-ergodic asymptotic invariance, a dictionaryselection NP-hardness result, and a concise empirical illustration.

Reinforcement learning with verifiable rewards, particularly Group Relative Policy Optimization (GRPO), has improved video reasoning in multimodal large language models. However, standard video GRPO varies only linguistic trajectories while conditioning every rollout in a group on the same sampled frames, which can produce homogeneous rewards and weaken group-relative learning when the shared visual evidence is insufficient or distracting. We propose \textbf{FrameGRPO}, a simple and effective reinforcement-learning framework that requires neither auxiliary networks nor changes to the model architecture. Specifically, FrameGRPO combines two lightweight components: (1) \textbf{Multi-Frame Rollouts}, which use only uniform temporal subsampling of a single maximal-frame decode to construct aligned views that diversify visual evidence and reasoning trajectories; and (2) \textbf{Gated Temporal Efficiency Credit}, which selectively reinforces a correct lower-ratio trajectory when all rollouts in a fixed full-ratio reference subset fail, without imposing an unconditional preference for fewer frames. Extensive experiments across diverse video understanding and reasoning benchmarks demonstrate robust overall improvements over standard GRPO, highlighting temporal-view diversification as an effective mechanism for exploration and credit assignment in video MLLMs.

LLM-based agents increasingly rely on reusable skills. However, scaling skill libraries makes efficient and accurate skill use difficult. Full skill loading incurs high token and runtime overhead. In contrast, flat semantic retrieval treats skills as independent text entries, ignoring dependencies, execution order, state transitions, and data flow constraints. We propose SPEM, a lightweight structure-aware procedural memory framework for efficient skill use in LLM-based agents. SPEM maintains Atomic Skill Memory for individual skills and Skill-DAG Memory for reusable compositions derived from successful executions. For a new task, SPEM reuses high-confidence Skill-DAG templates when available. Otherwise, it retrieves atomic skills and invokes Skill-GNN for state- and structure-aware skill selection. Successful executions are written back as new Skill-DAGs, allowing procedural memory to accumulate over time. Experiments on ALFWorld and SkillsBench show that SPEM achieves the best overall effectiveness--efficiency trade-off, reaching 87.1% ALFWorld success and 38.38% SkillsBench reward while reducing token and runtime costs against strong baselines.

Recommender systems often suffer from over-concentrated results when optimized for accuracy, limiting recommendation diversity. While diversity has become an important objective, existing methods still face fundamental limitations from both semantic and structural perspectives. On the one hand, multimodal recommender systems provide rich semantic signals, yet are often treated as auxiliary to ID embeddings, leading to ineffective semantic utilization for diversity. On the other hand, the accuracy--diversity trade-off persists during training, where accuracy gains often come at the cost of diversity. We show that this phenomenon is fundamentally governed by both the structural properties of learned representations and the dynamics of optimization. To address these challenges, we propose MOSS, a unified framework for multimodal decoupling through semantic and structural shaping. Specifically, we first introduce an MLLM-based Decoupled Feature Extraction module to disentangle multimodal content into informative representations. To enhance semantic utilization, we design a Category-aware Embedding Selection module that extracts compact, category-aware semantic features. To address structural limitations, we propose two complementary modules: Alternating Dual-Graph Propagation, which alleviates representation collapse by integrating collaborative and multimodal propagation, and Entropy-based Diversity Regularization, which mitigates optimization bias by promoting balanced category exposure. Extensive experiments on real-world datasets demonstrate that MOSS consistently achieves a superior accuracy--diversity trade-off over strong baselines.

Diffusion Transformers (DiTs) deliver high-fidelity image and video generation, but inference requires many sequential model evaluations. Two acceleration strategies are increasingly important: model compression makes each evaluation cheaper, while adaptive inference avoids selected teacher computation through reuse, forecasting, or speculative execution. Their combination, however, is brittle: directly applying the evaluated adaptive methods to a compressed student can degrade quality, an observed failure mode we call the \textbf{Combination Gap}. We link this gap to a mismatch between pointwise distillation and short-horizon execution: a student trained at teacher states need not remain reliable when its own predictions are chained. We propose \textbf{TamedFlow}, which trains a compressed DiT for this deployed role through cross-architecture feature alignment, a near-identity Velocity Alignment Head, and randomized rollout supervision. At inference, a lightweight Learnable Trajectory Estimator routes completed drafts: accepted drafts bypass the teacher, while rejected drafts roll back and receive one teacher rescue step. Distilling DiT-XL/2 into DiT-B/2, TamedFlow reaches FID $2.58$ at $2.25\times$ measured speedup over DDIM-50 on ImageNet $256\times256$. Ablations support trajectory supervision, and routing controls favor LTE over random decisions at nearly matched latency. Evaluations on ImageNet $512\times512$, MS-COCO, and Latte/FaceForensics extend the empirical scope.

Once historical clients become unavailable, a federation must serve newcomers through the artifact it leaves behind. How much of its knowledge remains useful in an architecture chosen afterward? We introduce FEDNBO, a seven-dataset benchmark that separates knowledge supply, policy identification, and private-model compilation under disjoint data roles. Controlled studies expose inflated oracle gains, architecture-dependent transfer reversals, and a gap between teacher accuracy and private-model utility. Ordinary public-response distillation provides a competitive reference without public task labels or newcomer samples: gains over strengthened local training are 23.52 and 16.62 percentage points on 300 CelebA identities and 300 FedEMNIST writers, respectively. In a complete-pipeline FedEMNIST comparison, this reference reaches 97.13% accuracy with 12.4 seconds of compilation per architecture, versus 97.03% and 67.5 seconds for a HeteroFedTTA reconstruction at the same epoch budget. Equal-count public-input interventions extend support sensitivity to trainable visual backbones. Together, these results establish an auditable benchmark and a strong reference for evaluating the utility, cost and reliability of post-federation knowledge transfer.

Robotic systems with revolute joints evolve on circular state spaces, yet most control architectures represent these states using Euclidean variables that ignore this underlying geometry. We propose a fully spiking neural control architecture in which ring attractors encode joint states, enforcing representations consistent with the topology of revolute systems. Homeostatic regulation computes goal-state mismatch directly over these distributed representations, and opponent gain-modulated populations convert this mismatch into controlled phase shifts of the activity bump, implementing set-point regulation as motion on the ring. We show that the bump dynamics follow an SO(2) geometry, and that gain asymmetries induce a signed phase velocity on the ring. To interface distributed activity with scalar regulation signals, we derive an analytic Fourier phase decoder and implement its signed components using push-pull spiking populations. For multi-joint systems, we demonstrate that local ring-based Fourier features compose into product representations that linearly decode SO(3) orientations through forward kinematics. Experiments in NEST validate the full SO(2) regulation mechanism and the spiking product-feature decoder for lift, pitch, and yaw. Robot experiments on a Tiago arm show that decoded ring dynamics generate joint velocity commands, producing directed motion with joint-dependent accuracy. Together, these results establish a neuromorphic control framework in which representation, regulation, and action are unified within a ring-attractor substrate aligned with the geometry of revolute systems.

We study regret minimization in finite-horizon episodic Markov Decision Processes (MDPs). While minimax-optimal algorithms are known, tractable approaches to instance-dependent optimality are still lacking. Motivated by this gap, we introduce $π_0$-PSRL, a variant of PSRL that uses posterior samples to decide when exploration is needed, while following a fixed reference policy $π_0$ during exploration episodes. This decouples the test for exploration, triggered when the sampled and empirical MDPs have different optimal policies, from the choice of the policy used to gather information. The resulting design addresses a limitation of standard PSRL, where the sampled optimal policy may not be the optimal choice for exploration. We prove instance-dependent regret bounds for $π_0$-PSRL, identifying the logarithmic exploration cost induced by $π_0$ and taking a step toward matching asymptotic lower bounds for episodic RL. Our proof techniques showcase a novel proof structure to derive problem-dependent regret bounds in episodic MDPs, and concentration results for Dirichlet random variables, that may be of independent interest.

Looped transformers reuse a shared block of layers to increase computational depth without increasing the number of trainable parameters. They are trained with backpropagation through time (BPTT). To save training compute, BPTT is often truncated to the last few loops, but truncation yields biased gradients that may weaken the benefit of looping. In this work, we study the value of truncation in looped transformer training in a systematic way: we jointly vary model size, training tokens, and forward and backward loop counts across nearly 600 training runs and fit a unified scaling law that captures the effects of looping and truncation through effective parameters. Despite the training-compute savings of truncation, our fitted law favors full BPTT when forward and backward loop counts are jointly optimized. In matched-compute runs, full BPTT through fewer loops consistently outperforms truncated settings of the same compute. The law further shows that at fixed model size, the compute-optimal loop count and training-token budget both grow with compute. When model size is also free, the law identifies two fully trained loops as the best looped setting. Experimentally, this setting achieves lower validation loss than the compute-optimal non-looped model with about 27% fewer non-embedding parameters and only 8% more training compute.

The Te(1-x)(GeSe0.5)Yx (x = 0.05, 0.10, 0.15) alloys were prepared by a melt-quenching method and characterized using dielectric spectroscopy. The dielectric parameters and electrical conductivity were evaluated to assess their potential for phase-change memory and other electronic applications. The temperature and frequency dependence of the dielectric constants, ac conductivity, and dielectric loss of prepared chalcogenide glassy alloys were investigated over the frequency range 1 Hz - 1 MHz, and the temperature range 302 - 422 K, below the glass transition temperature. Both dielectric loss and dielectric constant were found to decrease with increasing frequency and decreasing temperature. The observed dielectric loss change was interpreted using the Guintini model of dielectric dispersion, correlated barrier hopping (CBH) model, and orientation polarization. The Cole-Cole analysis confirmed a distribution of relaxation times and yielded activation energies consistent with the CBH conduction mechanism. The frequency dependence of the ac conductivity follows the power law (Aωs), where the frequency exponent (s < 1) decreases with increasing temperature, further supporting the CBH conduction mechanism. The incorporation of Y increases the density of localized tail states, resulting in enhanced ac conductivity. These findings suggest that the investigated materials are promising candidates for future phase-change memory and other electronic device applications.

Satellite-based quantum links are a promising enabler of long-distance entanglement distribution, overcoming the exponential losses of optical fibers. In uplink architectures, entangled photons are generated at ground stations and transmitted to orbiting satellites, reducing space-segment complexity but introducing significant atmospheric and geometric losses. Single-satellite dual-uplink configurations are additionally limited by the requirement of simultaneous visibility between both ground stations and the satellite. In this work, we investigate a dual-uplink dual-satellite network architecture and compare its performance with the single-satellite scheme under identical physical assumptions. The proposed topology introduces an inter-satellite link (ISL) that decouples the ground-station separation from the atmospheric uplink geometry. A realistic channel model is employed, incorporating geometric attenuation, atmospheric turbulence, background noise, and Bell-state measurement efficiency. The overall success probability and practical fidelity are evaluated as functions of satellite altitude and ground-station separation. Our results show that while the single-satellite configuration is suitable for regional-scale networks, it becomes infeasible for large separations due to geometric constraints. The proposed dual-satellite architecture mitigates this limitation and significantly extends the achievable communication distance, representing a key step toward scalable quantum internet infrastructures.

Referring segmentation in low-altitude drone imagery requires distinguishing small, visually similar instances under complex linguistic and spatial constraints. Global inference may yield multiple plausible masks or leave the intended referent insufficiently represented, motivating selective local re-inference. We propose DRA-SAM, a disagreement-driven framework that allocates an additional inference pass according to competing referential hypotheses and an unresolved state. A lightweight EvidenceScorer estimates beliefs over existing candidates and the unresolved state. A training-free Referential Disagreement LOOK policy balances belief-weighted candidate competition and unresolved-region coverage to select a crop for local re-inference. The same benchmark-adapted SAM3-I backbone processes the selected region, and the resulting evidence updates existing hypotheses and introduces new candidates. Among the compared methods, DRA-SAM achieves state-of-the-art performance on six of seven test metrics on both RIS-LAD and UAVid-RIS, including mIoU and [email protected]@0.9, and on all seven RIS-LAD validation metrics. On RIS-LAD, DRA-SAM improves the matched global SAM3-I baseline from 49.55 to 52.34 mIoU and from 46.80 to 58.05 [email protected]. On UAVid-RIS, the independently adapted DRA-SAM achieves 73.26 mIoU and 86.86 [email protected]. Controlled crop-policy comparisons demonstrate the effectiveness of disagreement-guided crop selection, while target-size analysis reveals particularly pronounced gains for aerial targets occupying less than 1% of an image. These results establish referential disagreement as an effective criterion for allocating local inference. Code is available at https://github.com/phy709/DRA-SAM.