Functional Analysis
Some applications of binary numeration systems with nonzero redundancy to the theory of locally complex functions
Svitlana Vaskevych, Yuliia Vovk and Oleksandr Pratsiovytyi
In this paper, we study a numeration system with a non–integer base $a>1$ over the binary alphabet $A=\{0,1\}$: $$\left[0;\frac{r}{a-1}\right]\ni x=\sum\limits_{n=1}^{\infty}\frac{\alpha_n}{a^n} \equiv \Delta^{r_a}_{\alpha_1\alpha_2\ldots\alpha_n\ldots}, \quad \alpha_n\in A.$$ We investigate the geometry of $r_a$–representation of numbers, including the
geometric interpretation of digits, the structure of cylinder overlaps, and the associated metric properties. The main object of our study is a nowhere monotone function of unbounded variation defined by $$f\left(x=\sum\limits_{n=1}^{\infty}\frac{\alpha_n}{2^n}\right) =\sum\limits_{n=1}^{\infty}\frac{\alpha_n}{a^n},$$ where the classical binary representation of the argument is assumed not to end with the infinite period $(1)$. We derive a system of two functional equations satisfied by $f$. We prove that the function has fractal level sets and a self–affine graph. We also evaluate the integral of $f$ over the interval $[0,1]$.
Read the full abstract ▾
Machine Learning
No Pain, More Gain: Iterative Merging for Effective Multi-Teacher On-Policy Distillation
Seonghyeon Kim, Chaeyun Jang, Noah Lee et al. (+2 more)
Multi-teacher on-policy
distillation (MOPD) combines independently developed domain teachers into a single student by distilling their predictions on student-generated samples. We study a setting where teachers share a reference model but undergo different post-training procedures, and find that MOPD can struggle to recover some teacher capabilities. Because
distillation occurs on student-generated prefixes, the student initialization can strongly affect subsequent recovery. However, initial
benchmark performance is not a reliable predictor of a good MOPD initialization. For example, merge initialization can start below SFT warm-up yet finish higher after MOPD. We further find that effective merging depends on both the relative teacher contributions and the overall merge scale, with some strong configurations lying outside the simplex of convex parameter averaging. Thus, selecting a good merge initialization requires evaluating not only its immediate performance but also the learning it enables under MOPD, making one-shot coefficient search difficult. We propose Iterative Merging for MOPD (IM-MOPD), which starts from a uniform merge and progressively adds task-vector increments for under-recovered domains during
distillation. In a 5-domain setting, IM-MOPD achieves higher average normalized recovery than MOPD with either uniform merge initialization or SFT warm-up, showing that effective teacher contributions can be determined progressively during training.
Read the full abstract ▾
All authors Seonghyeon Kim, Chaeyun Jang, Noah Lee, Boseop Kim, Juho Lee
Cryptography and Security
Optimizing and Securing the Modern Watermarking Channel for Images
Enoal Gesny and Eva Giboulot
To comply with recent regulations requiring traceable generated content, modern watermarking has adopted multi-bit post-hoc watermarking schemes. These modern designs rest on an encoder-decoder pair implemented as deep neural networks. These models are usually treated as pure black-boxes trained end-to-end, with the noise of the watermarking channel modeled through a fixed set of
geometric and valuemetric transforms applied to watermarked images. We argue that this purely empirical approach leads to unquestioned design flaws and a lack of theoretical performance guarantees. This work proposes a general theoretical model of modern post-hoc watermarking schemes grounded in a statistical analysis of the outputs of the encoder/decoder pair. We show that these deep neural networks implicitly define a watermarking channel modeled as parallel AWGN channels, with messages transmitted using BPSK modulation. This imposes a binary alphabet, greatly limiting the capacity of these watermarking systems. Another fatal flaw is their lack of a secret key, making them intrinsically insecure. We make this notion of watermarking security precise for post-hoc schemes by linking it to the possibility of estimating the secret key under a given statistical model of the decoder's output. By putting together the results from this theoretical analysis, we introduce SNW: a novel post-hoc watermarking system that significantly outperforms existing state-of-the-art baselines in terms of capacity while also providing strong security guarantees. Notably, it does not depend on a fixed codebook or binary alphabet, allowing it to reach a rate close to Shannon capacity through the use of capacity-achieving error-correcting codes.
Read the full abstract ▾
Robotics
Simulation for Planetary Robotic Perception and Autonomy: A Concise Survey of Recent Capabilities and Gaps
Hoyun Kim and Giseop Kim
Planetary robotics is an important enabler of scientific exploration in environments where direct human-in-the-loop operation is costly, hazardous, or infeasible. However, developing and validating planetary robotic systems remains difficult because representative field testing is expensive, limited, and often unrepeatable under mission-relevant conditions. In this setting, simulation serves as a central tool for perception and autonomy research, synthetic data generation, system integration, and pre-deployment evaluation. Despite its importance, the literature on planetary robotics simulation remains dispersed across different simulation engines, implementations, and application settings. This paper surveys simulation works for planetary robotic perception and autonomy across four practical axes: Openness and Availability, Scenario and Platform Coverage, Sensor and Perception Support, and Environmental and Operational Realism. The surveyed simulation works report visual or physical fidelity and support perception-oriented workflows. They also indicate uneven public availability, rover-centered coverage, partial support for specialized sensing modalities, and uneven reporting of operational constraints such as onboard computation, energy, and communication restrictions.
Read the full abstract ▾
Computer Vision
Do Emotion Concepts Generalize Across Sources, Modalities, and Architectures in Vision-Language Models?
Bohao Xing, Xin Liu, Kaishen Yuan et al. (+5 more)
Recent studies suggest that large
language models encode
emotion concepts as structured internal representations, but most existing work focuses on text and a single architecture. Therefore, we ask, do
emotion concepts generalize across sources, modalities, and architectures in
vision–
language models (VLMs)? To address this, we construct CMES (Cross-Modal
Emotion Stimuli), a multi-source collection of
emotion-conditioned stories, real facial expressions, synthetic portraits, and synthetic
emotion-evoking scenes. For each stimulus source, we extract a separate set of six Ekman
emotion vectors from each of three VLMs. We report four main findings as follows: 1) Image-derived
emotion vectors form a low-dimensional geometry similar to that of text-derived vectors. Valence is relatively stable across sources, while arousal varies more. 2) Text- and image-derived
emotion vectors have modest cosine similarity but still show held-out cross-modal correspondence. Text-derived vectors can also steer image interpretation. 3) Cross-architecture correspondence remains even when native cosine is near zero. Transformations estimated from generic ImageNet activations recover both correspondence and causal transfer without using the six
emotion vectors or their labels. 4) After aligning representations across architectures, we construct a shared
emotion subspace that preserves affective geometry and selective steering effects. The corresponding consensus
emotion vectors also generalize to a held-out fourth architecture at two model sizes. These results suggest that
emotion representations can share relational structure and causal effects across sources, modalities, and architectures, even when individual vector directions differ.
Read the full abstract ▾
All authors Bohao Xing, Xin Liu, Kaishen Yuan, Deng Li, Rong Gao, Guoying Zhao, Xiaolan Fu, Heikki Kälviäinen
Astrophysics of Galaxies
Before fragmentation: from uniform clouds to stellar clusters
A. Giannetti, T. Pillai, J. Kauffmann et al. (+4 more)
Our study of the early stages of fragmentation in massive clumps has revealed a striking anomaly: $\sim15\%$ of the CoCoA sample exhibit no fragmentation at all. These sources could represent the initial conditions before core formation. We investigate whether these sources are indeed at the onset of core formation, and whether
gravity or
turbulence dominates their early fragmentation. We study the transition from uniform gas to active cluster formation, and compare contrasting predictions of
turbulence- and
gravity-dominated models of clump fragmentation through ALMA observations of three massive clumps spanning this phase. The fragmented source exhibits higher average $n$(H$_2$) and a core-to-envelope density contrast 5-6 times larger than the minimally fragmented sources. All sources display pc-scale velocity differences comparable to their linewidths and gravitational velocities. This suggests that the linewidth is dominated by the large-scale motions, and that the latter are driven by
gravity. The high CH$_3$OH abundances and the mass build-up timescales align with astrochemical models where complex molecules form at intermediate densities before the appearance of cores. All independent diagnostics in our analysis are consistent with the minimally fragmented clumps being an early evolutionary phase, potentially marking the onset of core fragmentation. If
turbulence were the primary driver of fragmentation, sources with comparable Mach numbers should exhibit similar structural properties. Instead, we observe significant variations in density contrast and fragmentation. These findings challenge
turbulence-supported models and lend support to hierarchical,
gravity-driven models of clump fragmentation, where large-scale ordered motions account for a significant fraction of the observed linewidths and drive the
evolution of the clump toward increasingly dense and fragmented states.
Read the full abstract ▾
All authors A. Giannetti, T. Pillai, J. Kauffmann, A. Palau, E. Vazquez-Semadeni, K. Morii, P. Sanhueza
Machine Learning
Predictive Dual Smoothing for Column Generation
Senne Berden, Noah Schutte, Andrea Lodi et al. (+1 more)
Solving large-scale linear programs efficiently is an important challenge in many optimization settings. A key technique is column generation, which alternates between solving the master problem over a restricted subset of the variables, and using a pricing subproblem to identify new variables to add. The pricing subproblem is guided by the dual solution of the current restricted master problem, but oscillations in these dual solutions can substantially slow convergence. Dual stabilization methods address this issue. Dual smoothing is a common stabilization method, which guides the pricing subproblem using a combination of the current dual solution and duals from previous iterations. However, while past dual solutions can stabilize the dual trajectory, they do not necessarily guide pricing towards useful new variables. We therefore introduce predictive dual smoothing, which instead combines the current dual solution with a learned prediction of future duals to steer pricing towards variables that are more useful in subsequent iterations. The predictor is trained offline using supervision extracted from standard column generation trajectories and is used only to modify the pricing subproblem's objective function, while exact reduced-cost checks and fallback pricing with the unsmoothed duals preserve correctness. Experiments on cutting stock and generalized assignment problems show that predictive dual smoothing substantially reduces generated columns and wall-clock time relative to standard column generation and existing classical and learned stabilization methods. These gains extend to out-of-distribution instance sizes, and predictive smoothing provides further improvements when combined with strong classical stabilization.
Read the full abstract ▾
All authors Senne Berden, Noah Schutte, Andrea Lodi, Tias Guns
Optics
Surface Acoustic Wave-Mediated Brillouin Temperature Sensor on Thin-Film Lithium Niobate
Varun Marathan Kovil, Justus Nölkensmeier, Gladys Jara-Schulz et al. (+8 more)
The emergence of thin-film lithium niobate (TFLN) has paved the way for realizing strong Brillouin interactions in lithium niobate platforms, enabling new opportunities in integrated photonics. However, fully harnessing stimulated Brillouin scattering (SBS) in TFLN for the plethora of Brillouin applications remains an exciting area of research, with significant opportunities for further exploration and technological advancement. Here, we present the first comprehensive experimental and numerical study of the surface acoustic wave (SAW)-mediated Brillouin temperature sensor on a TFLN-on-
insulator chip. The proposed technique directly senses the surrounding environment via SAWs, providing an optical-power-efficient platform compatible with existing integrated photonic architectures. Exploiting the strong intermodal Brillouin interactions in x-cut TFLN waveguides, we report a high Brillouin gain of 24.8$ \pm $1.3\, m$^{-1}$W$^{-1}$ and a temperature coefficient of $-0.73$ $\pm$ 0.02\,MHz/$^\circ$C. Considering the anisotropic nature of lithium niobate, we investigate the temperature dependence of the Brillouin frequency for the propagation directions 0$^\circ$, 10$^\circ$, 20$^\circ$, 30$^\circ$ relative to the crystallographic y-axis.
Read the full abstract ▾
All authors Varun Marathan Kovil, Justus Nölkensmeier, Gladys Jara-Schulz, Pratham Kulkarni, Silia Babel, Christian Golla, Rajkumar Jadhav, Christof Eigner, Laura Padberg, Christine Silberhorn, Birgit Stiller
Computation and Language
Beyond Token Alignment: Event Completion for Cross-Tokenizer On-Policy Distillation
Jiacheng Liu, Jingwei Song, Qituan Zhang et al. (+2 more)
On-policy
distillation (OPD) transfers knowledge between
language models through teacher supervision on student-generated trajectories. With different tokenizers, a single teacher
token may require multiple student
tokens to generate, creating intermediate states where the event is entered but not yet completed. Existing cross-tokenizer methods align
tokens or text spans to construct comparable prediction targets. We study a complementary problem after partial generation: once the student produces a prefix of a teacher
token, multiple next
tokens may complete the same remaining bytes, but the teacher only specifies the required completion rather than how probability should be divided among these valid continuations. We introduce Event-Set Completion
Distillation (ESCD), which complements cross-tokenizer probability
alignment with completion-set supervision. ESCD aggregates prefix-related teacher events and supervises the total probability of byte-compatible one-step student completions, avoiding tokenizer-dependent probability splits among individual
tokens. The method reuses student trajectories and predictions, requiring neither additional rollouts nor changes to the student vocabulary. Experiments demonstrate consistent gains in mathematics, code, and scientific
reasoning across model families and tokenizers, extending to large-scale MoE
distillation from a 1T teacher to a 35B student. Local analyses show that retaining completion sets better matches the reference supervision, while one-step completion covers over 99% of observed compatible teacher mass after partial event entry in the studied tokenizer pairs. These findings support event entry and event completion as complementary supervision targets for cross-tokenizer knowledge transfer. Code will be released on GitHub.
Read the full abstract ▾
All authors Jiacheng Liu, Jingwei Song, Qituan Zhang, Siheng Chen, Linfeng Zhang
Signal Processing
Movable-Antenna-Empowered ISAC: 3D Ambiguity Function Analysis, Waveform Design and Antenna Position Optimization
Tianqi Mao, Yuanshuo Gang, Guangyao Liu et al. (+5 more)
Multiple-input multiple-output orthogonal frequency division multiplexing (
MIMO-OFDM) provides abundant spatial and time-frequency degrees of freedom for integrated sensing and communication (
ISAC), while movable-
antennas (MAs) further introduce reconfigurable spatial freedom through array geometry adjustment. However, how the array geometry and the
MIMO-OFDM waveform design jointly influence the sensing performance, i.e., the three dimensional (3D) ambiguity function property, remains insufficiently understood. Against this background, we investigate MA-empowered
ISAC design under the
MIMO-OFDM architecture, where the
radio-frequency (RF)
antenna positions and the baseband waveform are co-designed with symbol-level precoding (SLP) philosophy. The discrete periodic angle-range-Doppler ambiguity function is firstly derived for the MA-enabled
ISAC framework, revealing its interesting features with respect to the waveform
matrix rank. Specifically, it is proved that a rank-one waveform yields a 3D ambiguity function that is separable into angular and range-Doppler components. On the other hand, a higher-rank waveform combines multiple spatial components with angle-dependent weights, allowing the range-Doppler ambiguity function to vary across candidate angles and enabling the MA positions to further shape the 3D ambiguity function. Based on the aforementioned analysis, a 3D integrated sidelobe level ratio (ISLR) is defined, a joint optimization problem of the MA positions and the SLP waveform
matrix is formulated, and a penalty dual decomposition based alternating optimization
algorithm is developed. For the radar-only case, the existence of a rank-one globally optimal waveform is established, which reduces the original high-dimensional waveform optimization to a low-dimensional joint design of the spatial beam and MA positions.
Read the full abstract ▾
All authors Tianqi Mao, Yuanshuo Gang, Guangyao Liu, Shiqi Cui, Ying Sun, Yifeng Xiong, Christos Masouros, Ana García Armada
Artificial Intelligence
SeLMRoute: Probabilistic Semantic Evidence for Large Language Model Routing
Vasilis Perifanis, Nikolaos Pavlidis and Symeon Symeonidis
Large
language model (LLM) routing aims to select the most suitable model for each incoming query. Most existing routers learn this decision directly from query
embeddings, model representations, preference data, or clusters of similar examples. Such approaches can be effective, yet the representation used for routing rarely states what a query actually requires. We introduce SeLMRoute, a routing framework that separates the extraction of candidate-independent semantic evidence from the learning of candidate performance and the application of deployment objectives. A decision model first evaluates a set of interpretable questions about the query, such as its
reasoning requirements and use of external knowledge, with each judgment retained as a probability distribution. The resulting probabilistic semantic state is used by a lightweight supervised router to estimate candidate model performance. Routing objectives are applied after performance estimation, which allows the same semantic state to support performance-oriented and cost-aware decisions. On the LLMRouterBench (15
datasets, 20 candidate models, 11,481 queries), SeLMRoute achieves an average accuracy of $72.08\% \pm 0.45$, while grouped five-fold out-of-fold evaluation reaches $72.64\%$, compared with $69.23\%$ for the strongest fixed candidate. The representation achieves the highest mean performance among the evaluated semantic, dense, lexical, and domain-level representations. In a separate 13-model performance-cost setting, SeLMRoute improves performance in all five grouped splits, with a mean PerfGain of $2.66\%$. Our code is available at https://github.com/Indigma-Innovations/SeLMRoute.
Read the full abstract ▾
Artificial Intelligence
From Human Narrative to Harmonic Structure: A Human-Centered Investigation of Algorithmic Music Generation through the Chord Wheel Diagram
Josef Pavlíček, Petra Pavlíčková and Irena Štrausová
Contemporary AI-based music generation can produce compositions that satisfy formal requirements of tonality and musical coherence. However, whether musical expression can be described by mathematical properties alone remains a fundamental question. Human composers operate within personal and cultural contexts that influence harmonic decisions and deliberate departures from established patterns. This study investigates six narrative-driven popular songs by Bob Dylan, Johnny Cash, and Ritchie Valens. Original human harmonies are compared with outputs of an explainable computational harmonizer operating on the same melodies without access to the original chord progressions. We examine harmonic vocabulary, functional persistence, repetition, non-diatonic events, and tension-resolution patterns using Chord Wheel Diagrams and BPMN-based representations. Results show that high melody-chord compatibility does not necessarily imply preservation of the original human harmonic decision pattern. Some generated harmonizations retain the economical structure of the reference, while others alter harmonic diversity or suppress distinctive events while remaining compatible with the melody. Rather than quantifying artistic quality, the study introduces narrative-conditioned harmonic structure as a complementary perspective for computational music analysis. The findings suggest that generative systems may benefit from modeling not only harmonic correctness, but also structural identity, context, and human compositional intention.
Read the full abstract ▾
High Energy Physics (Phenomenology)
Termination of the Radial Ladder in Mixed-Flavor Heavy Mesons
Wen-Xuan Zhang and Wen-Nian Liu
A fundamental open question in
hadron spectroscopy is whether the radial excitation ladder of heavy
mesons terminates at a finite level once confinement is screened by light-
quark pair creation. We address this question for the mixed-
flavor mesons $D_s$, $B_s$, and $B_c$ by solving a screened Godfrey–Isgur
Hamiltonian with the Gaussian expansion method. The screening length is fixed by threshold-dependent string breaking analysis, and is matched to the charmonium and bottomonium anchors. The calculated spectra saturate at $3.36$, $6.78$, and $8.19~{\rm GeV}$ in $D_s$, $B_s$, and $B_c$ families, while the RMS radii grow by several fm and mass gaps fall below $10~{\rm MeV}$. Mass-radius decoupling places the end of the safe radial ladder at $n_{\mathrm{safe}}\approx 4$–$6$ for $D_s$, $5$–$7$ for $B_s$, and $9$–$11$ for $B_c$, so the ladder terminates first in $D_s$, then in $B_s$, and last in $B_c$. Above the limiting-mass windows, at $3.49$, $6.86$, and $8.24~{\rm GeV}$, states of the radial ladder cannot be discovered. This termination is directly testable with forthcoming high-statistics data from BESIII, Belle II, and LHCb.
Read the full abstract ▾
Computer Vision
SurgGMF: Fully Causal Gaussian Motion Forecasting for Anticipatory Surgical Scene Rendering
Jingqian Sun and Yichao Tang
Dynamic
surgical scene modeling is essential for robotic perception, simulation, and decision support. Although existing neural rendering methods enable efficient reconstruction and rendering of deformable
surgical scenes, they remain primarily focused on observed-frame reconstruction rather than
forecasting future scene states. To this end, we present SurgGMF, a fully causal Gaussian motion
forecasting framework for anticipatory
surgical scene rendering. Rather than predicting future RGB images directly, SurgGMF forecasts future Gaussian motion states represented by position, scale, and rotation residuals (X/S/R) from historical Gaussian motion fields. To prevent target leakage, we introduce a full-causal-last rendering protocol, where future Gaussian states are rendered without accessing target-frame Gaussian attributes while preserving causal appearance propagation. We evaluate SurgGMF on 12 EndoNeRF and StereoMIS
video slices using neural temporal learners and classical dynamics baselines under a unified
forecasting protocol. Learned Gaussian motion
forecasting consistently outperforms classical dynamics baselines in render space, demonstrating gains beyond hand-crafted state extrapolation. Latency analysis further reveals an accuracy–efficiency trade-off: under the current implementations, TKAN achieves the highest accuracy, whereas GRU and LSTM provide more favorable module-level latency profiles. These results establish SurgGMF as a reproducible framework for causal Gaussian motion
forecasting and advance
surgical Gaussian representations from retrospective reconstruction toward predictive scene modeling.
Read the full abstract ▾
Computer Vision
What Visual Generators Need from Teachers: Rethinking Representation Alignment
Yongcong Wang, Hingchin Chen, Mingyu Fan et al. (+6 more)
Representation
alignment speeds up
diffusion transformer training by pulling an intermediate block of the model (student) toward features of a frozen pretrained encoder (teacher). Which teacher layer to align, and for how long, is still set by convention, and each alternative costs a training run. We find that
alignment helps where the student cannot linearly recover the teacher's features, not where it already resembles them. Since a deep teacher layer is largely predictable from the one below, we isolate what each layer adds, its increment, and measure how much of it an unaligned student recovers. The student fills the teacher's hierarchy from the bottom up and stalls near the top, which we call hierarchy filling: even after 400K steps it recovers almost none of the deepest. The recoverability gap is the unrecovered share of an increment, read from one unaligned checkpoint. In short runs that each align one teacher layer at one block, the gap nearly reproduces their ranking by FID improvement, and CKA, a measure of feature similarity, largely reverses it. Representation
Alignment and Recoverability Estimation (RARE) picks the teacher layer with the largest gap before training. During training, it tracks each
token's remaining distance to that layer, the online counterpart of the gap, weights
tokens by it, and phases out the loss once the average distance stops falling. With SiT-B/2 on ImageNet $256\times256$, RARE reaches an FID of 18.02 without guidance and 4.46 with it, ahead of seven
alignment baselines including REPA, iREPA and HASTE. It also trains in 14% fewer GPU-hours than iREPA. Its FID stays below iREPA's across model scales, teachers,
datasets and backbones.
Read the full abstract ▾
All authors Yongcong Wang, Hingchin Chen, Mingyu Fan, Shuo Jiang, Teer Zhang, Yucong Sun, Zijia Wang, Yiming Lu, Chengchao Shen
Machine Learning (Statistics)
Information-Theoretic Analysis of Next-Token Prediction under Markovian Data
Masoud Kavian, Abdellatif Zaidi and Milad Sefidgaran
We develop an information-theoretic framework for generalization in next-
token prediction under temporally dependent data. We consider independent trajectories generated by finite-
memory Markov processes and distinguish algorithmic dependence, quantified by mutual information, from temporal dependence, characterized by mixing. For cross-
entropy loss, we derive an expected generalization bound using the Donsker–Varadhan variational representation and a McDiarmid-type concentration inequality for Markov chains. A refinement captures the joint effect of context length and temporal mixing through the mixing properties of the history-state process. We then extend the bound through a rate–distortion formulation, replacing mutual information with the minimum information rate required to represent the learned model within a prescribed distortion in the generalization gap, yielding informative guarantees for deterministic
algorithms over continuous hypothesis spaces. For margin-based prediction, we derive explicit bounds for linear and self-
attention next-
token predictors via noisy low-dimensional compression, revealing the roles of context length, model complexity, sample size, margin, and temporal mixing. Experiments on TinyStories and ETTh2 show that longer contexts can reduce both training and test losses, but typically reduce training loss more, enlarging the generalization gap. A complementary ETTh2 analysis identifies an effective predictive-
memory scale near 24 hours, with no statistically supported improvement beyond this scale, offering a plausible explanation for test-performance saturation at larger contexts.
Read the full abstract ▾
Robotics
Action Sequence Transfer via LLMs for Heterogeneous Environments
Choongho Chung, DongHwan Shin and Sung-Hee Lee
We present an action sequence transfer system that adaptively transfers user action sequences across different target spaces. Given an input action sequence from a source space and scene graph representations of both the source and target environments, our system predicts a corresponding action sequence in the target space by adapting to the spatial and object constraints of the new environment. To achieve this, we leverage multi-level representations of user activity to generalize actions at varying levels of abstraction. To demonstrate our system, we collect a new scene graph-based
dataset derived from the Ego4D GoalStep
dataset for evaluation. Results indicate that our system can generate valid action sequences even between spaces with drastically different object configurations.
Read the full abstract ▾
Machine Learning
Edge-Level Automorphism in GNNs: A Quantitative Framework and Effective Designs For Link Prediction
Chen Shao, Donald Loveland, Tobias Käfer et al. (+1 more)
Graph Neural Networks (GNNs) are effective for learning node and link
embeddings through permutation-equivariant aggregation. However, standard GNNs collapse automorphic nodes, i.e., those with identical structural roles (or orbits) into indistinguishable representations, leading to the node
automorphism problem. This collapse limits their expressive power and degrades link prediction performance. Existing approaches to characterize GNN expressiveness rely primarily on Weisfeiler-Lehman (WL) analyses, but these methods are typically qualitative and often misaligned with empirical results. To address this gap, we begin by introducing a novel quantitative framework to assess GNN expressiveness for link prediction. We first formalize edge-level
automorphism through edge orbits, which capture the set of structural role pairs for nodes that share a link. Then, we introduce the edge
automorphism ratio (EAR), a scalar metric that quantifies a GNN's ability to distinguish links in a given graph. We empirically demonstrate that EAR correlates strongly with performance, validating its practical benefit. Building on this insight, we design EDGE-ORBIT EQUIVARIANT GRAPH NEURAL NETWORK (EO-GNN), a GNN architecture that addresses
automorphism collapse while preserving equivariance and incurring minimal computational overhead. EO-GNN accomplishes this through two core designs combined with WL-based node hashes: (i)
automorphism-aware dropouts and (ii) subgraph orbit-biased aggregation. Empirical evaluations on synthetic and real graphs show improvements of up to 42.36% and 28.44%, respectively, in predicting links in scenarios with high
automorphism.
Read the full abstract ▾
All authors Chen Shao, Donald Loveland, Tobias Käfer, Danai Koutra
Information Theory
Robust Resource Management for SAGIN using DNN-Driven Channel Uncertainty Learning
Haonan Zhang, Weihua Wu, Qi Zhang et al. (+3 more)
This paper focuses on the joint robust beamforming and resource allocation for space-air-ground integrated networks (SAGIN) under uncertain channel state information (CSI). In SAGIN, uncertain CSI undermines the precise adjustment of beamforming and resource allocation, posing a major challenge to meeting heterogeneous users' strict quality of service (QoS) requirements. To address this challenge, we first formulate a chance-constrained optimization problem to minimize the total transmit power while satisfying QoS requirements under a predefined outage probability. By leveraging semidefinite relaxation (SDR), the objective function is transformed into a linear function of the traces of the beamforming
matrices. Then, we propose a deep neural network (DNN)-driven channel uncertainty learning to dynamically learn and model the uncertain CSI as an asymmetric uncertainty set. Under the constructed CSI uncertainty set, a robust counterpart method based on pre-trained network parameters is developed. It characterizes the uncertainty set as a finite union of convex sets, thereby providing a tractable approximation for the original chance constraints. Finally, we design an adaptive iterative
algorithm to jointly optimize the resource allocation and beamforming vectors. Simulation results show that our proposed DNN-driven method outperforms traditional robust and non-robust methods, achieving a superior energy efficiency and robust reliability in SAGIN.
Read the full abstract ▾
All authors Haonan Zhang, Weihua Wu, Qi Zhang, Runzi Liu, Zewei Jing, Weijia Han
Operating Systems
Dynamic Flow, Static Graph: KV Cache Reuse for Efficient LLM Serving on Mobile NPUs
Zhengxiang Huang, Shengheng Chen, Chaoyue Niu et al. (+6 more)
On-device large
language model (LLM) serving is a cornerstone of local-first personal intelligence, offering users data sovereignty, strong
privacy guarantees, and freedom from cloud API latency and cost. Although KV caching is widely used to reduce latency in long-context inference, existing designs were primarily optimized for cloud GPUs with dynamic execution environments and abundant
memory bandwidth. These architectural assumptions do not hold on mobile NPUs, where computation graphs must be statically compiled and both
memory capacity and I/O bandwidth are severely constrained. In this work, we present a compute-storage co-design for mobile-centric prefix and non-prefix KV reuse. We first propose an intra-graph mechanism that maps selective KV recomputation onto static NPU graphs, reconciling algorithmic dynamicity with NPU staticity. We further develop an inter-graph scheduler to optimize chunk merging and minimize padding with dynamic programming. To address mobile bandwidth limitations, we introduce a hierarchical KV manager featuring a tree-hash-semantic hybrid structure, along with cost-aware prefetching and eviction policies. We also build a two-dimensional pipeline that overlaps KV loading, rerotation, and storage with NPU execution, hiding data-movement latency. Experiments across representative on-device workloads and LLMs show that our design reduces time-to-first-
token (TTFT) by $40-60\%$ compared with no reuse and prefix-only caching.
Read the full abstract ▾
All authors Zhengxiang Huang, Shengheng Chen, Chaoyue Niu, Yujie Sun, Zhaode Wang, Zeyu Zhao, Chengfei Lv, Fan Wu, Guihai Chen
Solar and Stellar Astrophysics
Limits on eccentricity excitation by tidally enhanced mass loss in Wolf-Rayet binaries
Viktória Fröhlich and Zsolt Regály
Wolf–Rayet (WR) binaries frequently display high orbital eccentricities, and their eccentricity distributions are inconsistent with that of high-mass
Galactic field binaries at high significance. This raises the question of whether these eccentricities are
primordial or generated during the WR phase. We test whether a
tidal enhancement of the mass-loss rate can pump the eccentricity of a WR binary on a sample of 14 WR systems. We model each system as a variable-mass two-body problem and integrate the orbital
evolution over the WR lifetime under three prescriptions: i) adiabatic isotropic mass loss, ii)
equilibrium-tide circularization, and iii) a tidally enhanced, pericenter-focused mass loss in which the companion-facing reduction of the effective surface
gravity raises the local Eddington parameter. Adiabatic mass loss can not change eccentricity and affects only the orbital separation, which increases by up to a factor of ~20.
Tidal circularization is effective only in the tightest orbits, which are compatible with the observed systems only for high mass-loss rates, a regime in which any
primordial eccentricity is erased. The tidally enhanced mass loss channel pumps the eccentricity by only $\leq10^{-4}$, three to four orders of magnitude below the observed values, and therefore cannot be the responsible mechanism. We conclude that the high eccentricities of WR binaries are predominantly
primordial rather than generated by mass loss during the WR phase.
Read the full abstract ▾
Image and Video Processing
Gen2-VC: Unlocking Generative Priors for Video Compression
Yinhuan Huang, Jingkai Ying, Pu Chen et al. (+1 more)
Under stringent bitrate constraints, existing
video codecs struggle to balance source fidelity and perceptual realism. Distortion-oriented codecs often oversmooth details, while generative codecs risk introducing content and structural deviations and rely on codec-specific designs. This motivates a question: Can existing codecs achieve a better distortion–perception trade-off through simple, reusable adaptation? Our insight is that native codec reconstructions provide a shared interface through which generative refinement is anchored to source content while remaining decoupled from codec-specific representations. We therefore propose Gen2-VC, a generative
video compression framework that enhances the outputs of learned and conventional codecs with a pretrained
video prior, leaving their bitstreams and reference update processes unchanged. With the codec, VAE, and generative backbone frozen, lightweight LoRA adapters refine codec reconstruction through single-frame spatial adaptation followed by multi-frame temporal adaptation with the
video prior. Using Wan2.1-T2V-1.3B, Gen2-VC-DCVC-UF outperforms previous leading codecs in LPIPS/DISTS and, to our knowledge, is the first generative
video codec to surpass VTM-23.0 in both PSNR and MS-SSIM at low bitrates, based on BD-rates averaged over six
datasets. Compared to VTM-23.0, it reduces bitrate by an average of 86.65% and 94.24% at matched LPIPS and DISTS, respectively. Adapters trained only on DCVC-UF improve perceptual quality on DCVC-RT, ECM, VTM, and HM without retraining.
Read the full abstract ▾
All authors Yinhuan Huang, Jingkai Ying, Pu Chen, Zhijin Qin
Robotics
DexWeave: Learning Dexterous Humanoid Loco-Manipulation from Human Demonstrations
Naichuan Sun, Haotian Shen, Yizhang Zhang et al. (+5 more)
Learning dexterous
humanoid loco-manipulation from human demonstrations requires transferring not only human motion, but also the coordinated interaction structure underlying the demonstrated behavior. This is challenging because embodiment differences distort the coupling among body motion, wrist placement, finger articulation, and object interaction, while kinematically accurate references may still be difficult to realize under
robot dynamics. We present DexWeave, a unified framework that connects interaction-consistent motion retargeting with anatomy-aware whole-body policy learning. DexWeave first employs a two-stage retargeting procedure that initializes body and hand motions with specialized solvers and subsequently performs coupled refinement over the upper-body interaction chain while preserving lower-body support. The resulting references are tracked by an anatomy-aware
Transformer policy that represents anatomical regions as structured
tokens and uses directed masked
attention to model their dependencies, with object information selectively conditioning the upper-body pathway for dexterous interaction. The policy jointly outputs body and dexterous-hand actions and is trained directly with reinforcement learning, without pretrained tracking policies, teacher-student
distillation, or subsequent residual refinement. DexWeave improves retargeting fidelity and interaction consistency while achieving higher manipulation performance and faster policy convergence than MLP baselines. We further deploy the learned policies on a physical Unitree G1
humanoid equipped with Inspire dexterous hands, demonstrating dexterous whole-body loco-manipulation in the real world. See our project page (https://dexweave.github.io) for
videos.
Read the full abstract ▾
All authors Naichuan Sun, Haotian Shen, Yizhang Zhang, Luying Feng, Haoze Wang, Yuanbo Xiangli, Yaochu Jin, Peidong Liu
Soft Condensed Matter
Effects of chain stiffness on the breakdown of the Cox-Merz rule in linear and ring polymer melts
Keishin Tsujino, Yuhi Sakamaki, Shota Goto et al. (+2 more)
The Cox-Merz rule is an empirical relation between the magnitude of the complex viscosity, $\lvert\eta^*(\omega)\rvert$, obtained from linear viscoelasticity, and the steady-shear viscosity, $\eta(\dot{\gamma})$. In this study, we performed molecular dynamics simulations of coarse-grained
polymer melts using the Kremer-Grest model and systematically examined the validity of the Cox-Merz rule as a function of chain length and chain stiffness for linear and ring
polymers. For short chains and, more generally, for flexible chains, $\lvert\eta^*(\omega)\rvert$ and $\eta(\dot{\gamma})$ show good agreement at corresponding values of angular frequency $\omega$ and shear rate $\dot{\gamma}$. For linear
polymers, however, the discrepancy between the two viscosities becomes increasingly pronounced with increasing chain length and chain stiffness, with $\lvert\eta^*(\omega)\rvert$ exceeding $\eta(\dot{\gamma})$ at high corresponding values of $\omega$ and $\dot{\gamma}$. This deviation is associated with the pronounced nonlinear response under strong steady shear, where flow-induced chain extension and
alignment can modify the
entanglement constraints that govern relaxation near
equilibrium. In contrast, ring
polymers retain substantially better agreement between $\lvert\eta^*(\omega)\rvert$ and $\eta(\dot{\gamma})$ with increasing chain length and stiffness within the range examined here, indicating a weaker breakdown of the Cox-Merz rule. These results demonstrate that the applicability of the Cox-Merz rule depends strongly on chain length, chain stiffness, and molecular architecture. When the molecular response under steady shear deviates substantially from the relaxation behavior characterized by linear viscoelasticity near
equilibrium, the complex viscosity can no longer accurately predict the steady-shear viscosity, resulting in a pronounced breakdown of the Cox-Merz rule.
Read the full abstract ▾
All authors Keishin Tsujino, Yuhi Sakamaki, Shota Goto, Kang Kim, Nobuyuki Matubayasi
Computer Vision
Geometry as Address: Routing Attention to Visual Memory for Long-Horizon Camera-Controlled Video Generation
Zesong Yang, Weikai Chen, Liyuan Cui et al. (+10 more)
Long-horizon camera-controlled
video generation requires recovering previously observed content from an ever-growing visual history. Existing approaches either search historical context implicitly or reconstruct it into persistent 3D
memory, facing inefficient
memory access or accumulated
geometric errors. Our key insight is that geometry need not explain the scene–it only needs to determine where visual
memory should be read from, while
attention decides what should be recovered. Based on this insight, we introduce GEAR, a Geometry-Enabled
Attention Routing framework that uses geometry as an explicit
token-level address for visual
memory. Rather than fusing historical observations into a persistent global 3D representation, GEAR retains them as frame latents and uses per-frame geometry only to establish
token-level correspondences with target views, thereby avoiding persistent error accumulation from global fusion. Guided by these correspondences,
Geometric Correspondence
Attention (GCA) selectively injects geometrically matched historical features into noisy target patches during denoising. We further introduce an Invisible Octree to accumulate visibility evidence and reject geometrically plausible but occluded correspondences. Extensive experiments demonstrate that GEAR achieves state-of-the-art visual quality, precise camera control, and revisit consistency, enabling minute-long
video generation along challenging trajectories.
Read the full abstract ▾
All authors Zesong Yang, Weikai Chen, Liyuan Cui, Lutao Jiang, Runze Zhang, Yingda Yin, Xiaoyang Huang, Kai Yan, Keyang Luo, Wangguandong Zheng, Xin Wang, Hujun Bao, Zhaopeng Cui
Mathematical Physics
Scattering Measures and Information Geometry
Tatsuro Oda
We formulate scattering experiments in terms of parameterized finite measures on outcome spaces and develop the associated information geometry defined by the Fisher metric. The decomposition of a scattering measure into total mass and normalized shape separates Fisher information into contributions from changes in overall scattering strength and redistribution among outcomes. Measurement processes are represented by probability or sub-probability kernels acting on the scattering measure. Such kernels contract Fisher information, placing a broad class of experimental operations within a common information-theoretic framework. This formulation connects scattering experiments to the theory of optimal experimental design through the Fisher geometry of scattering measures. The same finite-measure structure arises naturally in
quantum measurements with incomplete detection, clarifying its place in the hierarchy from
quantum measurement to classical outcome statistics.
Read the full abstract ▾
Computer Vision
DBCF: Dual-Branch Complementary Fusion of Foundation Models for Generalized Deepfake Detection
Fengming Gu, Mingjie He, Zonghui Guo et al. (+2 more)
As image generation and editing technologies have progressed substantially, facial forgeries pose significant challenges to
privacy and public safety. Due to limited ability to capture forgery cues, existing small-scale forgery detection models often struggle to generalize across various domains and unseen manipulations. To address this limitation, researchers have turned to large-scale foundation models, which can provide richer representations and better generalization. Nevertheless, relying on a single foundation model alone remains insufficient for effective forgery detection. While models like CLIP offer robust global semantic cues, they lack the capacity to capture detailed local facial features. In contrast, DINO excels at capturing local structural features of faces, but provides weaker global semantic context. To fully utilize the synergies among multiple foundation models, we propose a hierarchical multi-granular framework that integrates complementary pretrained representations. Specifically, a Global Context Branch (GCB) based on CLIP captures holistic semantic cues, while a Fine-grained Cue Branch (FCB) built on DINOv3 captures localized structural irregularities. In addition, we design a feature fusion module that enables parameter-efficient adaptation of the frozen foundation backbones by adaptively extracting and integrating complementary features from the two models. By jointly leveraging global context and fine-grained cues, our method learns more comprehensive forgery representations and achieves strong cross-manipulation performance. Extensive experiments on multiple
benchmarks demonstrate the benefit of the proposed design, particularly under cross-
dataset and cross-manipulation settings.
Read the full abstract ▾
All authors Fengming Gu, Mingjie He, Zonghui Guo, Jie Zhangb, Shiguang Shan
Robotics
ExcavaTwin: Training-Free Geometry-Guided Semantic Elevation Mapping for Autonomous Excavation
Yu Deng, Lingshan Zeng, Tong Hu et al. (+1 more)
Autonomous excavation requires a spatial representation that jointly captures terrain geometry and task-relevant semantics. Existing excavation mapping is largely elevation-centric, while generic semantic models remain unstable in unstructured outdoor scenes. We present ExcavaTwin, a pure-
vision geometry-guided semantic elevation mapping framework without excavation-specific training. Given multi-view RGB images, the framework: 1) reconstructs scene geometry and semantic observations using frozen
vision models; 2) derives terrain and non-terrain
geometric support; 3) performs geometry-constrained multi-view semantic fusion to suppress implausible predictions and recover incomplete observations; and 4) projects the fused state into a task-oriented semantic elevation map. Experiments on public
datasets and real excavation scenes demonstrate reliable
geometric and semantic perception. In real excavation, the system achieved an average update interval of approximately 1.4 s and a mean elevation error of 12.74cm in dynamically modified regions. Larger errors mainly occur during rapid terrain changes and transient visual disturbances caused by machine motion.
Read the full abstract ▾
All authors Yu Deng, Lingshan Zeng, Tong Hu, Rushi Dai
Machine Learning
Quality Determines Direction, Length Shapes Magnitude: Length Control for Open-Ended Reinforcement Learning
Zijun Weng, Zhongan Bi, Xuanang Gao et al. (+5 more)
Reinforcement learning (RL) changes not only what
language models say, but also how much they say, often increasing response length at the cost of
token efficiency. Controlling this length growth is particularly challenging in open-ended RL because (i) response length is entangled with quality, (ii) open-ended tasks lack a natural success boundary for deciding when efficiency should be prioritized, and (iii) dense, graded rewards often yield small within-group quality margins, making quality-induced advantages especially sensitive to reward-level length shaping, which can perturb their magnitudes and even reverse their signs. We therefore adopt an asymmetric principle: quality should determine the direction of reinforcement, while length should only shape its magnitude. We instantiate this principle with Quality-Gated Length Advantage Shaping (QGLAS), which first computes advantages from quality rewards alone, then adds bounded bonuses only to shorter positive-advantage responses, leaving all other advantages unchanged. The bonus strength is further adapted to within-group quality separation, allowing conciseness to matter more when quality-favored responses are similar and less when their quality differences are clear. Across different model families, open-ended
benchmarks, and reward sources, QGLAS consistently achieves a stronger quality–length trade-off than representative baselines. At approximately 30% compression, QGLAS retains 98.4–102.0% of the macro-average quality gains achieved by quality-only RL over the base model, compared with 68.3–75.5% for these baselines at comparable compression.
Read the full abstract ▾
All authors Zijun Weng, Zhongan Bi, Xuanang Gao, Xiaohui Hu, Shuangyong Song, Yongxiang Li, Kaidong Yu, Xuanjing Huang
Computation and Language
ReMCTS: Reflection-Enhanced Monte Carlo Tree Search for Code Generation
Huifei Wang, Xinying Huang, Yiheng Sun et al. (+1 more)
Open-weight large
language models (LLMs) can generate function-level programs from natural-
language prompts, but plausible candidates still fail on hidden semantics and repeat mistakes across repair attempts. We present ReMCTS, an execution-grounded,
memory-augmented, LLM-guided MCTS-style search framework. It organizes program candidates as tree states, retains branch-local debugging context, retrieves failure experience across branches, and distinguishes failed checks from unavailable evidence. On HumanEval and MBPP-Sanitized, visible-test ReMCTS improves over direct generation in 8 of 10 model-
dataset pairs under held-out evaluation, whereas proxy-only search is less stable. Controlled tree-search, sampling, repair, and
memory ablations characterize the source and limits of these gains. A 30-task HumanEval-X C++ pilot further demonstrates compatibility with compiler-backed execution, but does not constitute a broad
multilingual evaluation.
Read the full abstract ▾
All authors Huifei Wang, Xinying Huang, Yiheng Sun, Yifan Yuan