arXiv:2606.27378v1 Announce Type: new Abstract: We introduce an axiomatic evaluation framework for latent thought representations in LLMs, comprising metrics that are independent of downstream benchmark scores and reveal representational failures that benchmark accuracy masks. Existing evaluations conflate representation quality with model capacity. Therefore, failures cannot be attributed to the representation rather than to the model that processes it. We formalize four functional axioms (Causality, Minimality, Separability, and Stability) and define a quantitative measure for each, computed directly on the representation independently of downstream accuracy. We audit open-weight LLMs across 23 reasoning tasks (e.g., Spatial Reasoning, Factual QA). We find that no candidate satisfies all four axioms simultaneously, that the representations distinguish task type reliably but cannot distinguish between two questions within the same task, and that the representations encode little information beyond what is already present in the input embedding. The failure is consistent across dense, reasoning-distilled, and RL-trained model families, indicating that the gap is structural rather than a property of model size or training procedure.
Science Journals
arXiv:2606.27398v1 Announce Type: new Abstract: Context: Generative AI adoption in software engineering education raises opportunities for learning support alongside concerns about superficial learning and academic integrity. Objective: This study investigates how explicit instructional guidance aligned with Bloom's taxonomy supports responsible GenAI use in SE/CS education, exploring student and instructor perceptions of GenAI-supported learning. Method: We designed a Bloom-aligned GenAI framework that articulated appropriate GenAI roles at different cognitive levels. The framework was embedded in course instructions, labs, and assessment across multiple SE/CS courses at Queen's University Belfast and Azerbaijan Technical University. Data were collected via anonymous questionnaires and learning artifacts, analyzed using thematic analysis with Bloom's taxonomy as an analytic lens. Results: Students perceived GenAI as most valuable for higher-order cognitive activities (analysis, evaluation, reflection) and less suitable for foundational learning. Explicit Bloom-level guidance influenced students to use GenAI reflectively, with delayed or intentional non-use when independent thinking was prioritized. Both students and instructors reported pedagogical benefits alongside challenges in cognitive effort and instructional design workload. Conclusion: GenAI's educational value lies in intentional alignment between cognitive learning goals, instructional guidance, and learner self-regulation. Bloom's taxonomy provides a scalable, pedagogy-driven framework for responsible GenAI use in SE/CS education, offering a practical alternative to enforcement-focused responses.
arXiv:2606.27460v1 Announce Type: new Abstract: In this study, we use a developmental approach to investigate the statistical learning and mental representation of neural language models (NLM). A series of Generative Transformer models are trained on a synthetic grammar. The model states are saved at multiple stages in the course of training. Through analyzing how the internal representations of these models change in the developmental path, we found that NLMs acquire the most abstract global statistical knowledge at the beginning of learning and later acquire the relatively local statistical dependencies. This learning path contains many over-generalizations from the very beginning and these over-generalizations are gradually constrained in the later stage of learning. Based on this observation, we propose a new framework to explain the statistical learning and language cognition of NLMs.
arXiv:2606.27467v1 Announce Type: new Abstract: Molecule generation methods that leverage generative models have been successfully applied to drug discovery. However, they often require extensive pre-training, suffer statistical biases in the training data, and might suffer from limited interpretability of generated chemical structures. In this work, we introduce SpectralMol, an algorithm based on evolutionary computation that processes chemical structures as a compact matrix of Fourier coefficients, projected onto a fixed basis to generate position-wise latent vectors for SELFIES decoding. The NSGA-II algorithm enforces diversity and enable separate objective functions rather than collapsed objectives into a scalar reward. The quality of the algorithm was tested against standardized benchmarks. The results show comparable aggregate benchmark performance with a task-dependent profile: SpectralMol is strongest on several multi-parameter optimization tasks. The same benchmark was used to perform an ablation study to demonstrate the advantages of a structured latent matrix. Finally, method was tested on a realistic ClpP-targeted drug-discovery benchmark, comparing it with the reinforcement-learning-based model under a fixed oracle-call budget. SpectralMol generates more docking hits and more diverse scaffolds while maintaining competitive physicochemical properties. The representation adopted in this work can cleanly separates scaffold-level modifications from localized substructure variations, as the former occur with perturbations of low-frequency Fourier modes and the latter with perturbations of high-frequency Fourier modes. The results support the evidence that frequency-controlled evolutionary dynamics provide an interpretable, efficient, and training-free route to multi-objective molecular design.
arXiv:2606.27547v1 Announce Type: new Abstract: Human motion generation models are fundamentally constrained by the limited diversity of motion capture datasets, which predominantly contain common, repetitive actions and fail to cover the long tail of complex human movements, resulting in a restricted motion vocabulary in learned latent representations and poor generalization to rare, compositional, and highly dynamic motions. In this work, we propose a framework for expanding the motion representation space by leveraging large-scale synthetic human motion, introducing a data generation pipeline that produces diverse, physically plausible motion sequences beyond the distribution of existing datasets and integrating it with a redesigned VQ-VAE tokenizer that adapts to this expanded motion space. Unlike conventional tokenizers trained on narrow data distributions, our approach jointly scales both the training distribution and the discrete codebook, enabling the model to capture a significantly richer set of motion primitives. We demonstrate that training with synthetic motion substantially improves the coverage and compositionality of the learned motion vocabulary, leading to consistent gains across motion generation tasks such as text-to-motion and motion continuation, while remaining fully compatible with existing frameworks including MotionGPT. Our results suggest that the primary bottleneck lies in the limited support of the learned motion representation, rather than model architecture alone. Scaling synthetic motion in tandem with representation learning offers a principled path toward more expressive, controllable, and generalizable human motion synthesis.
arXiv:2606.27550v1 Announce Type: new Abstract: Multi-token prediction has been shown to increase data density during training, improve downstream text-generation quality, and serves as the defacto approach for self-speculative decoding. Existing foundation and open source models that use MTP heads commit to a static tree-based attention topology throughout the entire generation sequence, meaning the speculation depth, and thus the compute required during verification, stays constant regardless of the context. This is fundamentally misaligned with the entropy patterns of natural language where low-entropy regions often support reliable multi-step drafting, while high-entropy regions require more conservative speculation. To address this, we propose Entropy-guided Multi-Token Prediction (EntMTP), a training-free scheduler that toggles between tree-based attention topologies from a set of task-specific pareto-optimal trees conditioned on a running estimate of local generation entropy. By matching speculation depth to context predictability, EntMTP maximizes expected accepted-token throughput across the full distribution of generated text without sacrificing generation quality. When evaluated across Humaneval, ShareGPT, GSM8k, and Litbench benchmarks, EntMTP consistently achieves a 1.15x speedup against Hydra and peak speedup of 1.36x against Medusa baselines respectively.
arXiv:2606.27596v1 Announce Type: new Abstract: Large Vision-Language Models (LVLMs) exhibit sophisticated reasoning but remain susceptible to object hallucination. Deviating from the prevailing attention intensity assumption, we reveal a deeper dynamic structural misalignment: hallucination is triggered at decision-critical steps where specific attention heads, acting as risky mediators, decouple from visual evidence to lock onto language priors. This establishes a pathological shortcut that bypasses visual grounding. To dismantle this, we propose Fox (Faithfulness and Observational-flow via eXpression-rectification), a training-free inference-time framework. Fox diagnoses structural misalignment using a visual attention entropy probe to localize risky mediators unsupervisedly. We then execute a targeted causal intervention via numerical logit saturation to physically sever the shortcut path. Finally, a conflict-gated cooperative decoding strategy reconciles interventional faithfulness with observational fluency. Extensive experiments demonstrate that Fox achieves SOTA performance, outperforming SID by 29.1% while preserving linguistic richness. Code is available at https://github.com/Cc2021start/Fox.
arXiv:2606.27598v1 Announce Type: new Abstract: Ultra-fine entity typing (UFET) assigns highly specific types to entity mentions, but current approaches struggle with types in the long tail. We hypothesize that a key limitation is the reliance on sentence-level context, since disambiguating evidence is often spread across multiple sentences. Testing this has been difficult because all existing UFET resources are sentence-level. We present Narrative-UFET, a controlled extension of UFET in which each entity mention is paired with an automatically generated short, coherent narrative. Synthesizing narratives lets us isolate the effect of specific discourse properties. We experiment with two paired variants: one in which the entity's type is held constant across the narrative (Maintain) and one in which it shifts (Change). We show that narrative context yields consistent improvements on long-tail types over sentence-level baselines, with the Change variant providing the stronger signal. A comparison against naturally occurring contexts shows that synthetic narratives yield stronger gains, indicating that controlled discourse construction can surface signals that real text leaves implicit. Substantial room for improvement remains, suggesting open directions in both discourse modeling and narrative construction.
arXiv:2606.27620v1 Announce Type: new Abstract: In space plasma studies, spacecraft measurements often provide time histories of the solar-wind plasma. However, many heliospheric plasma processes are organized over spatial scales that cannot be directly resolved by limited local sampling. This creates a persistent challenge: how to use limited probe measurements to recover the spatial plasma distributions needed to interpret evolving solar-wind structures. In this work, we present a recurrent reduced-order learning framework to address this challenge. The method is demonstrated using WSA-ENLIL solar-wind simulation data to reconstruct two-dimensional meridional and equatorial fields from a small number of virtual probes, with radial velocity and plasma density considered as target quantities on both planes. From sparse temporal probe signals as inputs, the model recovers the dominant radial and latitudinal variations in the meridional plane and the spiral-shaped organization of the equatorial solar wind. It is also able to reconstruct spatial distributions of dynamically coupled plasma fields not directly sensed. Sensitivity studies are performed to assess the dependence of reconstruction accuracy on key parameters of the machine-learning framework: modal rank, number of probes, and input-history length. The outcomes underline the methodology's promise as a practical route for extracting spatial plasma-state information from spacecraft measurements in support of studies on the underlying physics of space and solar-wind plasmas.
arXiv:2606.27622v1 Announce Type: new Abstract: Byzantine-robust federated learning seeks to protect distributed model training from malicious or corrupted clients without requiring access to their private data. FLTrust addresses this challenge by introducing a trusted server-side root dataset that assigns trust scores to client updates for more robust aggregation. In this work, we propose FOGGYTRUST, a hierarchical extension of FLTrust that localizes trust computation to fog nodes, allowing the framework to better handle globally heterogeneous data while preserving robustness within locally homogeneous client groups. We further show that this two-level architecture can simultaneously address distribution mismatch in trust estimation and client drift across groups by combining local trust-based aggregation with heterogeneity-aware global optimizers such as FedAdam and SCAFFOLD. Across benchmark datasets, FOGGYTRUST achieves its strongest gains on more challenging heterogeneous settings, particularly on CIFAR-10 under Krum and Trim attacks, where it achieves an over 50% improvement over FLTrust. We also test FOGGYTRUST in a real-world safari dataset to show the promise of hierarchical trust networks for robust federated learning in socially impactful, safety-critical settings such as distributed wildlife monitoring.
arXiv:2606.27542v1 Announce Type: new Abstract: In this paper, we present an experimental evaluation study of the Alternating Direction Method of Multipliers (ADMM), which is a widely used technique in the distributed optimization of power distribution networks. The focus of this study is on how real 5G communication performance affects ADMM in a fully experimental platform that features commercial 5G connectivity and real-time control. The ADMM-based Distributed Optimal Power Flow (DOPF) problem is solved using the IEEE 123-bus unbalanced distribution feeder subdivided into five areas, each managed by a local controller implemented on a Raspberry Pi. To mitigate the impact of the communication network variability, we propose a delay threshold-based mechanism that yields a 7.75% reduction in convergence time compared to a no-threshold baseline. We also devised a policy to dynamically update the threshold value based on communication and computation conditions, achieving a 26.42% reduction in the convergence time compared with the static optimal threshold. These results demonstrate the potential of adaptive, communication-aware control strategies for real-world Smart Grid (SG) deployments.
arXiv:2606.27627v1 Announce Type: new Abstract: Discrete audio representations have become increasingly popular for building multimodal text-audio systems and integrating audio capabilities into Large Language Models (LLMs). However, numerous studies report performance degradation on various downstream tasks due to information loss during discretization. To address this, we propose a novel approach combining temporally compressed discrete tokens with dimensionality-reduced continuous residuals. Our framework consists of a hybridized discrete-continuous focal modulation codec and a hybrid Transformer. This architecture performs autoregressive inference in the discrete domain, coupled with non-autoregressive prediction and continuous residual upsampling. Experimental results show that our approach significantly improves the retention of speaker characteristics compared to discrete-only methods, while simultaneously reducing the number of required autoregressive steps.
arXiv:2606.27709v1 Announce Type: new Abstract: Recent work has shown that fine-tuning large language models (LLMs) for social warmth degrades factual reliability and increases sycophancy. We investigate a related but distinct failure mode: warmth fine-tuning also weakens adversarial safety, making models more susceptible to jailbreaks and harmful output generation. We examine whether this reflects an inherent consequence of empathetic adaptation or an artifact of data construction. To address this, we introduce a persona-driven rewriting pipeline that conditions user turns on low agreeableness and pairs this with warm, de-escalating assistant responses. Across three experiments on four models, our approach reduces jailbreak susceptibility and harmful output rates relative to generic warmth fine-tuning baselines, while preserving conversational warmth. Representational probing provides suggestive evidence that this conditioning reduces the geometric alignment between warmth and compliance directions in latent space. These results show that safer empathetic fine-tuning is achievable through data design alone, without safety labels, harm detectors, or changes to the training objective.
arXiv:2606.27721v1 Announce Type: new Abstract: Compositional generalization, the ability to solve complex problems by combining solutions to simpler sub-problems, is a fundamental capability of both natural and artificial intelligence, and a key mechanism underlying chain-of-thought reasoning. However, the theoretical underpinnings of compositional generalization remain poorly understood: when and why does decomposing a problem into parts yield more efficient learning than solving it directly? We study this question through the canonical problem of learning to simulate semiautomata (predicting the outcome of $T$ steps of sequential computation), a model that captures state tracking, regular language recognition, and modular arithmetic. We show that an autocurriculum-based approach building on Part I of this series, recursively decomposing longer sequences into shorter sub-problems, learning to solve them, and composing the solutions, achieves dramatically better statistical complexity than direct methods. (i) For a setting inspired by supervised fine-tuning (SFT) where the learner receives interactive feedback on intermediate states of the computation, curriculum facilitates learning from only $2^{\mathcal{O}(\sqrt{\log T})}$ tokens of supervision; i.e., subpolynomial in the sequence length $T$, overcoming the $\Omega(T)$ token barrier required by direct simulation. (ii) For a setting inspired by reinforcement learning with verifiable rewards (RLVR), where the learner improves a pre-trained reference model using an outcome verifier, we show that curriculum reduces the requirement on the reference model from coverage at the full sequence length $T$ to coverage at a shorter block length $B \ll T$, an exponentially weaker condition.
arXiv:2606.27629v1 Announce Type: new Abstract: Cross-platform deployment of offensive comment detection for Chinese social media suffers performance degradation. The paper proposes a dual-threshold hard mining method to address this. First, the clean-Chinese-base RoBERTa is finetuned on COLD to establish a binary baseline for fair comparison. Second, a three-class fine-labeled test set covering Weibo, Xiaohongshu, Tieba, and Zhihu is constructed, domain distances from the source are quantified using Jaccard and Proxy-A Distance, as well as the degradation bottleneck of the baseline under domain shift is systematically revealed. Herein, a dual threshold hard example mining strategy is proposed. High- and low-confidence error-prone samples are filtered from unlabeled corpora by prediction confidence. The model is secondarily finetuned under implicit contexts with merely a small set of manually labeled hard examples, realizing low-cost cross-platform domain adaptation. Experiments reveal significant performance gains of the optimized model across four platforms.
arXiv:2606.27794v1 Announce Type: new Abstract: Language-guided Medical Image Segmentation (LMIS) has shown great potential to improve the delineation of anatomical structures and lesions by integrating clinical textual information. Existing methods generally rely on either implicit interaction between textual and visual features or auxiliary coarse-grained supervision for cross-modal alignment. However, these methods lack explicit and fine-grained constraints to ensure semantic consistency, causing a mismatch between language and the segmentation outputs. To address this issue, we propose Text-as-Illumination Retinex Network (TIRNet), a novel Retinex-inspired framework that treats text embeddings as semantic illumination for feature modulation, thereby improving semantic consistency in LMIS. TIRNet introduces two key blocks integrated at each decoder stage: (1) the Retinex-inspired Text Modulation Block (RTMB), which employs positive and negative illumination maps to enhance text-relevant foreground features and suppress background interference; and (2) the Consistent Detail Compensation Block (CDCB), which selectively recovers high-frequency details via a consistency-gated mechanism conditioned on illumination reliability. Furthermore, we propose a Multi-Scale Illumination Supervision Loss (MSIS-Loss), comprising a Region-Grounded Contrastive Loss (RGC-Loss) that enforces cross-modal similarity to be concentrated in text-relevant foreground regions and suppressed in background regions, and a Background Suppression Loss (BS-Loss) that provides pixel-level supervision for negative illumination maps, jointly ensuring a precise cross-modal alignment at each decoder stage. Extensive experiments on the MosMedData+ and QaTa-COV19 datasets demonstrate that TIRNet achieves state-of-the-art performance in LMIS. The code is available at: https://github.com/anaanaa/TIRNet.
arXiv:2606.27407v1 Announce Type: new Abstract: We introduce the \emph{observational hierarchy}, a new axis of classification for formal languages, orthogonal to the Chomsky hierarchy. An observer is a function $O : \Sigma^* \to S$ that determines which information about the input is accessible to a computational system. The order-blind automaton, which perceives the input as a multiset of symbols rather than a sequence, constitutes the paradigmatic case. We prove that the class of languages recognisable by any machine equipped with such an observer coincides exactly with the permutation-closed languages. We then define a partial order on observers that induces a hierarchy of language classes parametrised not by the computational power of the machine, but by the structure of the observer. We prove that this hierarchy has the structure of a partial order with a diamond-shaped profile sub-lattice, comprising the length branch $O_\bot \prec O_{\mathrm{len}} \prec O_{\mathrm{prof}} \prec O_\top$ and the parity branch $O_\bot \prec O_{\mathrm{par}} \prec O_{\mathrm{prof}} \prec O_\top$, with $O_{\mathrm{len}}$ and $O_{\mathrm{par}}$ incomparable, and an infinite subsequence branch $O_\bot \prec O_1 \prec O_2 \prec \cdots \prec O_\top$, both converging to the complete observer. We prove that the observational hierarchy is strictly incomparable with the Chomsky hierarchy, and introduce the notion of \emph{observational complexity} of a language. We further define observer-parametrised complexity classes $\mathbf{P}_O$ and $\mathbf{NP}_O$, and show that computational hardness and structural blindness are two independent phenomena. In particular, $\mathbf{P}_{O_{\mathrm{prof}}} = \mathbf{NP}_{O_{\mathrm{prof}}}$ holds as a structural collapse strictly inside $\mathbf{P}$.
arXiv:2606.27445v1 Announce Type: new Abstract: We present a study on the time-domain propagation of electromagnetic waves in dielectric materials modeled by a nonlinear Debye medium with random perturbations. Polynomial Chaos Expansions are employed to transform the random nonlinear Debye polarization model into a deterministic framework. We extend the Yee discretization to the resulting coupled system, establish second order accuracy, and verify convergence numerically. We investigate the sensitivity of nonlinear properties to uncertainty, particularly when the amplitude of the input signal is large. Given the challenges in manufacturing where uncertainties can cause optimal parameters to vary and potentially disrupt nonlinear effects, our approach incorporates these uncertainties within the simulation. This can enable the model-based design identification of realizable materials that maintain their desired effects despite variations. The findings from this study contribute to a deeper understanding of wave propagation in complex media, with potential implications for applications in optical communications, material science, and electromagnetic wave control.
arXiv:2606.27448v1 Announce Type: new Abstract: This paper studies the problem of regret minimization in Markovian bandits with \emph{non-observable states} and possibly \emph{constrained} decision epochs. The focus is restricted to a ``pure'' regret benchmark, that compares the performance of the learning algorithm to the best \emph{pure policy} which -- akin to optimal policies of stochastic bandits -- picks the optimal arm from start to finish without ever switching. We introduce a generalization of rested Markovian bandits, \emph{self-degrading Markovian bandits}, for which pure policies are always asymptotically optimal.We show that without prior knowledge on the underlying bandit, the regret of algorithms that switch arms rarely necessarily scales super-logarithmically for every bandit, i.e., as $\omega(\log(T))$, where $T$ is the learning horizon. Despite the unreachability of the logarithmic regime, we design UCB-NOM, an optimistic algorithm inspired by UCB, of which the regret is nearly logarithmic. Lastly, we show that given prior knowledge on the Markovian bandit in the form of a bound on the bias functions of its arm, a proper instantiation of UCB-NOM achieves $O(\log(T))$ regret. We further show that this prior knowledge allows for a $O(\sqrt{T \log(T)})$ worst-case regret bound for UCB-NOM. Notably, our regret bounds do not depend on the number of states of the underlying Markov chains. Our findings suggest that the non-observability of states is a mild inconvenience in self-degrading Markovian bandits.
arXiv:2606.27517v1 Announce Type: new Abstract: We prove two Korovkin-type approximation theorems for sequences of positive linear operators acting on continuous functions on $[0,\infty)$. Under the assumption of pointwise convergence on suitable test functions, we establish pointwise convergence for all functions with polynomial or exponential growth. As direct applications, we obtain convergence results for the classical Baskakov and Sz\'asz--Mirakjan operators. The proposed method offers an elementary framework that can be applied to a broad class of positive linear operators.
arXiv:2606.27540v1 Announce Type: new Abstract: The intermittency of the rate of turbulent energy dissipation ${\epsilon}$ is investigated experimentally, with special emphasis on its scale-similar facets. This is done using a general formulation in terms of multifractals, and by interpreting measurements in that light. The concept of multiplicative processes in turbulence is (heuristically) shown to lead to multifractal distributions, whose formalism is described in some detail. To prepare proper ground for the interpretation of experimental results, a variety of cascade models is reviewed and their physical contents are analysed qualitatively. Point-probe measurements of ${\epsilon}$ are made in several laboratory flows and in the atmospheric surface layer, using Taylor's frozen-flow hypothesis. The multifractal spectrum $f({\alpha})$ of ${\epsilon}$ is measured using different averaging techniques, and the results are shown to be in essential agreement among themselves and with our earlier ones. Also, long data sets obtained in two laboratory flows are used to obtain the latent part of the $f({\alpha})$ curve, confirming Mandelbrot's idea that it can in principle be obtained from linear cuts through a three-dimensional distribution. The tails of distributions of box-averaged dissipation are found to be of the square-root exponential type, and the implications of this finding for the $f({\alpha})$ distribution are discussed. A comparison of the results to a variety of cascade models shows that binomial models give the simplest possible mechanism that reproduces most of the observations. Generalizations to multinomial models are discussed.
arXiv:2606.27562v1 Announce Type: new Abstract: We introduce a two-dimensional non-hydrostatic model for shallow water wave dispersion. The model is based on a locally adapted application of a non-hydrostatic correction to the hydrostatic shallow water equations (SWE) in a predictor-corrector scheme. Applying the non-hydrostatic correction uniformly to the entire domain demands a high computational cost, since an elliptic system of equations needs to be solved for the correction terms. We demonstrate that by determining the area where the non-hydrostatic effects are significant, and applying the correction only locally, the computational effort can be reduced by approximately 40\% without sacrificing accuracy in tsunami-like scenarios. As indicators for the non-hydrostatic effect, we use the ratio between total water depth and surface elevation, as well as horizontal velocity norms. Results are shown for several well-known test cases, including wave trains over a semi-circular shoal, static, and moving bottom tsunami-like wave propagation.
arXiv:2606.27579v1 Announce Type: new Abstract: Accurate assessment of tumor proportion score (TPS) in non-small cell lung cancer (NSCLC) is critical for treatment planning and prognosis. Key challenges include the tedious manual work required to annotate each slide, combined with the limited number of experts certified for this task. Multiple instance learning (MIL) has proven to be an effective approach for predicting TPS scores at the slide level; however, existing methods struggle with non-expressive (zero class) images. Our approach involves two models: (1) an embedding-extraction and multiclass-classification network that captures the histopathological features of individual patches, and (2) a MIL model that aggregates these embeddings to predict zero-inflated beta (ZIBeta) parameters representing the overall TPS probability distribution for the entire slide. Using only slide-level TPS scores as labels, we demonstrate how this end-to-end framework can leverage a novel distribution-based architecture to improve prediction accuracy and explainability. ZIBeta modeling significantly outperforms baseline linear and ridge regression while capturing expected accuracy through distribution concentration.
arXiv:2606.27651v1 Announce Type: new Abstract: In recent years, with the emergence of Temporal Knowledge Graphs (TKGs), research on learning entity and relation representations in TKGs has attracted increasing attention, giving rise to a large number of TKG embedding methods. TeRo is a simple and efficient temporal knowledge graph embedding approach. However, TeRo does not do well in modeling the mapping properties of various relations, such as one-to-many, many-to-one, and many-to-many. Meanwhile, it also has limitations in the expression of temporal information. To address these issues, we propose a novel TKG embedding method named TeRoR. This method divides the temporal evolution of entity embeddings, and conducts independent rotation transformations on head and tail entities in the complex vector space to strengthen temporal information modeling capacity. In terms of relational characteristics, we train a radius to constrain the rotated and translated head entities within a circular region centered on the tail entity, which effectively captures the diverse mapping properties of relations. Experimental results demonstrate that TeRoR achieves competitive performance against state-of-the-art models on four distinct TKG datasets.
arXiv:2606.27883v1 Announce Type: new Abstract: In stealth-constrained swarm robotics, visual communication provides a critical alternative to active radio transmissions, which might be jammed. This research investigates motion-based communication for non-active information exchange, utilizing modular, dynamically feasible planar trajectories as visual cues. On the receiver drone end, a pose estimator tracks the transmitting drone's pose, feeding it into our custom 3DTrajDecoder. The decoder is designed to classify and segment the spatiotemporal sequence while simultaneously regressing its size and normal vector. To robustly train the decoder on both communicative and non-communicative trajectories, we developed a configurable online procedural generation pipeline. We validate our system through real-world testing and simulation to define its operating domain, supported by an extensive ablation study detailing our architectural choices and system limitations.